DISASTER RISK MANAGEMENT

National Systems for the

Comprehensive Management of Disaster Risk
and Financial Strategies
for Natural Disaster Reconstruction

Paul K. Freeman
Leslie A. Martin
Joanne Linnerooth-Bayer
Reinhard Mechler
Georg Pflug
Koko Warner

Inter-American Development Bank

Sustainable Development Department
Environment Division
Integration and Regional Programs Department
Regional Policy Dialogue

Washington, D.C.
Cataloging-in-Publication data provided by the
Inter-American Development Bank
Felipe Herrera Library

Disaster risk management: national systems for the comprehensive

management of disaster risk and financial strategies for natural disaster recon-
struction / Paul K. Freeman … [et al.].

p. cm.
“Prepared for the Natural Disaster Network”
Includes bibliographical references.

1. Emergency management--Latin America. 2. Disaster relief--Latin

America. 3. Natural disasters--Latin America. I. Inter-American Develop-
ment Bank.

363.348 D774—dc21

The views and opinions expressed herein are those of the authors and do
not necessarily reflect the official position of the Inter-American Develop-
ment Bank.

August 2003

Environment Division
Sustainable Development Department
Inter-American Development Bank
Stop W-0500
1300 New York Avenue, N.W.
Washington, D.C. 20577

E-mail: [email protected]
Fax: 202-623-1786
Web site: http://www.iadb.org/sds/en
 Acknowledgments

The authors wish to thank the following persons for contributing to the success of the
studies. Howard Kunreuther of the Wharton School, University of Pennsylvania, contrib-
uted especially to the document on national systems for risk management. Sergio
Saldana of the International Institute of Applied Systems Analysis participated actively
in the preparation of the chapter on financial strategies for reconstruction. Omar Dario
Cardona of the Universidad de los Andes, Colombia, provided substantial comments for
both chapters. Allan Lavell of La Red, headquartered in Costa Rica, reviewed the initial
drafts of the text on national systems. Paul K. Freeman was the principal researcher in

Martin, Joanne Linnerooth Bayer and Koko Warner contributed to both chapters of
this project and Leslie Martin was the project coordinator. Paul K. Freeman, Leslie A.

this publication. The contributions of Reinhard Mechler and Georg Pflug focused on
chapter 2.

The two chapters were presented at the Regional Policy Dialogue on Natural Disasters
during meetings at the IDB in Washington, D.C., in November 2001 and May 2002.
They were reviewed by 32 government officials and members of the Regional Disaster
Dialogue Network from 15 countries in Latin America and the Caribbean.

Kari Keipi of the IDB’s Sustainable Development Department coordinated the IDB pro-
ject team for the studies. The other members included Caroline Clarke and Sergio Mora
of the Regional Operational Department 2, Javier Cuervo and Kim Staking of the Re-
gional Operational Department 3, Guillermo Collich of the Regional Operational Depart-
ment 1, Edgardo Demaestri and Justin Tyson of the Sustainable Development Depart-
ment. Mr Tyson is currently with the Department of Treasury of the United Kingdom.

Special thanks belong to the Publications Section of the External Relations Office of the
IDB for facilitating the final editing and publication of this document.

Financing for the studies was provided by the Integration and Regional Programs De-
partment. The Sustainable Development Department was responsible for the technical
supervision of the work.
 Foreword

This report was commissioned as part of the Regional Policy Dialogue on Natural Disas-
ters of the Inter-American Development Bank (IDB) to examine national systems and
institutional mechanisms for the comprehensive management of natural disaster risk.
Latin America and the Caribbean are only too familiar with the devastating impact of
hurricanes, floods, earthquakes, landslides, volcanic eruptions, and other natural disas-
ters. With an average of 40 significant disasters a year, Latin America is second only to
Asia in frequency of disasters affecting the region.

Natural hazard policies in much of Latin America and the Caribbean have traditionally
focused on establishing efficient disaster response. However, modernization of the sys-
tems calls for a more comprehensive vision of disaster risk management that includes
an emphasis on prevention and mitigation and strives to involve citizens and the private
commercial sector. In this respect, the Regional Policy Dialogue on Natural Disaster
commissioned a two-stage study focused on understanding national, integrated disaster
risk management systems and the related financing, a report which is based upon lit-
erature reviews, case studies, and consultation with experts regarding the existing good
practices of natural disaster risk management programs worldwide.

While the first phase of the study discusses the components of a national system, the
second focuses on instruments for financing reconstruction after a disaster. The re-
search compares centralized, government-directed management systems with those
that are localized and decentralized, and also analyzes the factors affecting the financial
and political stability of alternative approaches. As natural disasters may result in major
resource gaps for governments facing the task of financing reconstruction, the report
presents case studies of four countries—Bolivia, Colombia, the Dominican Republic, and
El Salvador—to highlight the various policy options. Alternative sources of ex ante fund-
ing are identified, including reserve funds, contingent credit, and insurance. These in-
novative methods of funding are compared with ex post funding possibilities through
international aid, loan diversions and increased external debt, budget reallocations, and
tax increases.

In dealing with the management of natural disasters, Latin America and the Caribbean
have made a gradual shift from an ex post response approach toward a system con-
cerned with investments in prevention and mitigation. The main challenge facing gov-
ernments is to incorporate these preventative investments with planning for possible
reconstruction as part of the overall strategy for disaster risk management. In this re-
spect, the authors aim to demonstrate the wide range of policy options available, which
will prove to be contingent upon the circumstances of each country.

Carlos M. Jarque Nohra Rey de Marulanda

Manager Manager
Sustainable Development Department Integration and Regional Programs Department
 About the Authors

Paul Freeman is an adjunct professor at the Graduate School of International Studies

at the University of Denver. His research and publications focus on strategies for devel-
oping countries to cope with the costs of natural disasters. He has worked extensively
with the Inter-American Development Bank, the World Bank, and the International
Monetary Fund. He is a graduate of Harvard Law School and completed a PhD in eco-
nomics at the University of Vienna.

Joanne Linnerooth-Bayer is a research associate at the International Institute for

Applied Systems Analysis in Laxenburg, Austria, where she leads the project Risk, Mod-
eling, and Society. She completed her PhD in economics at the University of Maryland.
She has published widely on topics related to environmental and technological risks, is
an associate editor of the Journal for Risk Research, and serves on the editorial boards
of Risk Analysis and Risk Abstracts.

Leslie A. Martin is working with the World Bank’s Resident Mission in La Paz, Bolivia to
design a systems approach to analyzing key production chains and setting targets in the
social sector and at the municipal level. Previously she was a research scholar with the
Natural Catastrophes and Developing Countries Project at the International Institute for
Applied Systems Analysis. Ms. Martin completed a bachelor’s degree in mathematics
with a minor in economics at the Massachusetts Institute of Technology.

Reinhard Mechler is an economist and research scholar at the International Institute

for Applied Systems Analysis in Laxenburg, Austria. His research focuses on the impacts
and costs of natural disasters in developing countries and strategies to reduce those
costs, in particular, risk financing. Dr. Mechler has been a consultant at the Inter-
American Development Bank, the World Bank, and the Gesellschaft für Technische
Zusammenarbeit. He holds a diploma in economics from Heidelberg University and a
PhD in economics from Karlsruhe University in Germany.

Georg Pflug is a professor at the University of Vienna, head of the Department of Sta-
tistics and Decision Support Systems at the University of Giessen, and a research
scholar with the International Institute of Applied Systems Analysis. He is an associate
editor for several research journals on statistics and probability, has written or edited
several books, and has written more than 60 articles published in refereed journals.

Koko Warner is an economist and research scholar at the International Institute for
Applied Systems Analysis. For the past five years, she has worked on the economic and
social impacts of natural catastrophes in developing countries. Her current research in-
vestigates catastrophe risk management and weather-related catastrophe peril in de-
veloping countries. Dr. Warner completed her PhD in economics and is an assistant pro-
fessor at the University of Richmond’s Emergency Service Management graduate pro-
gram.
 Contents

Executive Summary i

Chapter 1
National Systems for the Comprehensive Management of Disaster Risk 1
Elements of Comprehensive Disaster Management
Country Experiences with Disaster Management Systems
Designing Effective National Disaster Management Systems
Sustainability of National Systems
Conclusion

Chapter 2
Financial Strategies for Natural Disaster Reconstruction 35
Lessons Learned
Measuring the Resource Gap
Financing Reconstruction
Policy Recommendations

Bibliography 75
 Executive Summary

National Systems for Disaster Risk Management

This publication presents the results of two interrelated projects completed for the Re-
gional Policy Dialogue on Natural Disasters for Latin America and the Caribbean fi-
nanced by the Inter-American Development Bank. The first project addresses the sus-
tainability of national risk management systems for coping with natural disasters, and it
is primarily a survey of existing practices in Latin America, Asia, the United States and
Europe. The work explores important characteristics of national systems that make
them sustainable both politically and fiscally. The second project focuses on one charac-
teristic of national risk management systems: the financing of reconstruction on the ba-
sis of case studies for Bolivia, Colombia, El Salvador and the Dominican Republic.

Central Government or Community-driven Disaster Risk Management

A national disaster risk management system comprises the formal or informal interac-
tion between institutions, financial mechanisms, regulations and policies. It is com-
monly believed that for a national disaster system to succeed governments must be ac-
tive participants in its creation and implementation. Concern exists on focusing natural
disaster policy on existing government systems that sometimes enhance narrow power
structures and draws away from local concerns and initiatives. Those holding this con-
cern favor reducing natural hazard risk to community-driven projects and programs de-
veloped by nongovernmental organizations (NGOs). Such an approach to risk manage-
ment is not guaranteed to be comprehensive, but applies directly to identifiable needs
and the empowerment of local populations.

These two approaches to risk management need not be mutually exclusive. The task
facing policymakers is to create effective, integrated national systems that engage sen-
ior government policymakers and accommodates and supports local decision-making
and private market initiatives.

Two Phases of Disaster Management

The key elements of risk management are divided into two phases: the pre-disaster
phase and the post-disaster period. The pre-disaster phase includes risk identification,
risk mitigation, risk transfer, and preparedness; the post-disaster phase is devoted to
emergency response, rehabilitation and reconstruction. Comprehensive risk manage-
ment addresses all these seven components. Many countries have developed, or are
developing, national programs to partly or fully incorporate these elements of an inte-
grated disaster risk management involving the public sector, civil society, and private
sector commercial actors. Each national disaster system reflects the political and eco-
nomic cultures and conditions of the particular country.

Risk Assessments

Risk assessments are an essential part of the process of integrating natural disaster
programs with overall development objectives. These assessments identify sources of
risk, vulnerable groups, and potential interventions. In the first stages risk maps can be

i
integrated with poverty valuations to identify communities most in need of disaster-
related mitigation and preparedness projects. This identification process may serve as a
cornerstone for all of the initial risk management activities. Risk assessment allows poli-
cymakers to specifically define the objectives of the risk management programs and to
establish vulnerability reduction targets.

Evolution of the National Systems in the Region

Most national disaster risk management systems in Latin America and the Caribbean
consisted initially of government bodies dedicated to emergency response. Some coun-
tries still continue to rely almost primarily on civil defense. In other countries, large
natural catastrophes over the past decades have highlighted the need for more com-
prehensive systems that would include prevention, mitigation, preparedness, and provi-
sions for reconstruction and rehabilitation activities with the participation of civil society
and market actors. As a result, several countries, such as Argentina, Brazil, Chile, Co-
lombia, the Dominican Republic, El Salvador, Nicaragua, Guatemala, Honduras, and
Mexico, have begun to transform their approaches to coping with natural disaster risk.
In Central America and the Caribbean, subregional disaster reduction organizations
such as the Centro de Coordinación para la Prevención de los Desastres Naturales en
América Central (CEPREDENAC) and the Caribbean Disaster Emergency Response
Agency (CDERA) have fostered this transition.

Characteristics of Effective National Systems

Drawing from country experiences and recommendations from policymakers involved in

national systems, some key guidelines for effective disaster management emerge. First,
a national system should rely on an explicit disaster strategy. An appropriate national
disaster strategy should be closely integrated with national policies for development and
environmental protection. Second, successful national systems should also incorporate
key players in the disaster management process. Such players include, among others,
the finance ministry, local community leaders, NGOs, and private market actors. Third,
successful national systems should have provisions to ensure sufficient resources for
key players to carry out their responsibilities.

Financing Reconstruction

One consistent shortcoming in the risk management strategies of developing countries

has been the lack of planning and financial protection against disasters. Traditionally,
developing countries have relied on emergency transfers from their limited government
budgets and reallocation of existing loans, and donations from international agencies to
fund their disaster losses. This use of resources for reconstruction financing places an
increasing strain on the ability of the countries to fund longer-term economic and social
development programs.

In developed economies, considerable reliance is placed on ex ante financing tools to

provide needed post disaster reconstruction funding. Is there a meaningful role that
these tools can play in assisting developing countries to finance their reconstruction?
This issue is subject to considerable debate in the development community.

Financial Planning

By their very nature, ex ante risk management tools are complicated. The benefit of
these tools lies in an understanding of probability. These instruments require monies to

ii
be spent today to reduce the consequences of an unknown, but probably occurring, fu-
ture event. If the future event does not occur, the value of the money spent to protect
against the event looks lost. Even worse, the perceived benefit of spending the funds on
other important investments is also absent. To use ex ante risk management tools, a
policymaker must bridge the psychological gap of weighing the cost of current expendi-
ture against future unknown but predictable consequences. At the country level, the
tradeoff is usually framed as a tradeoff between growth (a result of more money being
spent now) and stability (a guarantee of funds to pay for future losses). There is a need
to establish an appropriate framework for balancing these two competing needs for de-
veloping countries with restricted resources and immediate poverty reducing needs.

Filling the Potential Resource Gap

The starting point for addressing the financial problem related to catastrophes is identi-
fying potential funding sources for reconstruction after a disaster in order to fill the “re-
source gap.” The resource gap measures the inability of a country to finance its recon-
struction obligations from traditionally available sources after a disaster. The calculation
of the resource gap requires three computations. First, the risk of the country to natural
hazard events must be estimated. Risk is a function of the hazard (or probability of
phenomena of different magnitudes impacting a country) and the vulnerability (or sus-
ceptibility of the exposed population and assets to loss). The second calculation con-
cerns financial data that the government assumes to finance the losses not only of its
own assets but also its responsibility to cover some private losses (for example, of the
poor). Primary losses from natural hazard events may be covered by various parties in
addition to the government—industry, businesses, homeowners, and individuals—but
the concern here is with government responsibility. Third, the capacity of the govern-
ment to meet its financial obligations must be calculated. To the extent that the gov-
ernment lacks the resources to fund its obligations, it has a natural disaster resource
gap. The required resources may come from the government budget or diversion of re-
sources from other programs, revenues (tax), reserves, insurance proceeds, borrowing,
or international aid. All of these alternatives have an associated cost and limitations on
availability.

Case Study Results

The analysis of the cases through the modeling exercise suggests that several countries
in the region may find it profitable to engage themselves in an in-depth analysis of al-
ternative financial protection schemes especially in preparation for potential large scale
natural disasters. Policies directed at reducing risk or guaranteeing post-disaster re-
sources are likely to pay high dividends. More detailed information should be generated
for rational risk management decision making since the conditions may vary.

Small countries with historically high incidence of natural disasters may face the possi-
bility of significant shortfalls in their ability to finance post-disaster reconstruction. This
is the situation in the Dominican Republic and El Salvador, cases analyzed in this study.
For large countries with more modest or diversified disaster risk, the study suggests a
greater ability to absorb losses from disasters. This has been the case for Bolivia, which
also has had sufficient resources to respond, thanks to traditional access to low interest
loans from multilateral institutions. On the other hand, in the similarly geographically
diverse but more populous and relatively high per capita income country of Colombia,
the government has been able to expand tax revenues to cover disaster losses. But for
any country, changes in their vulnerability (increasing urbanization in disaster prone

iii
areas, for example) or economic situation should compel a reexamination of past finan-
cial solutions to finance potential future disaster losses.

Recommendations

In examining policies, institutional framework and financial tools available in Latin

American and Caribbean countries, the following general conclusions and recommenda-
tions can be made:

• Integrated national systems for risk management. The countries in the region have
been creating national systems for the comprehensive management of disaster risk.
These should implement prevention, mitigation and emergency management as well
as reconstruction after a disaster has occurred in order to facilitate sustainable de-

•
velopment in the region.
Risk analysis. Governments should analyze the risk of natural hazard events. The
techniques for evaluating risk exist and most countries have the necessary data to
assess hazard exposure and vulnerability. What is lacking is the time and resources
to integrate the known information, thus limiting the ability of the government to
plan for disasters, instead of only responding to them. The evaluation should be
done at the national, regional, and municipal levels especially for all essential infra-
structure and buildings. Schools, hospitals, bridges, and roads are all examples of

•
assets for which models can be developed.
Government risk. Each government needs to create a clear inventory of obligations
for which it is responsible. If the government is responsible for a risk, this should be
made clear and the obligation should be budgeted. If the government does not as-
sume responsibility for some private sector risk, it should examine strategies to as-

•
sist the private sector to assume that risk on its own behalf.
Prevention and mitigation. Countries should invest in prevention and mitigation to
avoid rebuilding exposure after a disaster occurs. Land-use planning, building codes
and proper reconstruction standards should be developed before a disaster occurs.
If not carried out before, reconstruction after a disaster should provide the opportu-
nity to implement the proper risk reduction measures for the future. Initial invest-
ments in prevention and mitigation can significantly and cost-effectively reduce vul-
nerability to natural disasters. However, the marginal net benefit of such invest-
ments diminish gradually. Countries also need to develop alternative ways of loss fi-

•
nancing from several internal and external sources.
Loss financing. Countries should evaluate the ways by which they finance losses, be
it through reserve funds, calamity funds, contingent credit, insurance or through ex-
ternal credit. The level of use of each instrument should be a result of an integrated
risk management strategy. If conversions of existing loans are considered, the crite-
ria for their use should be openly discussed in advance, and not be left to a poten-
tially hasty decision-making process during an emergency. It is noted that in many
countries in the region insurance is included as an alternative tool applied by their
governments. However, considerable savings or substantial increases in insurance
protection could be accomplished if the countries systematically reviewed its insur-
ance purchasing opportunities.

iv
 Chapter 1

National Systems for the Comprehensive

Management of Disaster Risk

Although disaster response is important, it fails to address the causes of disaster losses.
Those causes are rooted in the complex interaction of human settlement and the natu-
ral environment. Recurring natural events become disasters because populations exist
in harm’s way in structures inadequately prepared to withstand anticipated natural haz-
ard events. To protect people and their assets, natural disaster policies must deal with a
broad set of issues. In developing countries, those issues are tied to the network of
policies addressing economic development. The best protection from natural disasters is
an economically viable country with strong democratic institutions. Just as the reduction
of poverty requires a comprehensive mix of policies that involve many components of
society and government, reducing the toll of natural disasters requires a comprehensive
approach that accounts for the causes of a society’s vulnerability to disaster. Not only
must a comprehensive strategy be articulated, the political and economic will must be
created to sustain the new policies.

A national disaster system is composed of the interaction of the institutions, financial

mechanisms, regulations, and policies that constitute a country’s approach to disaster
risk management. This interaction can be formal or informal. It is commonly believed
that for a national disaster system to succeed in being comprehensive, national gov-
ernments must be active participants in the creation and implementation of a formal
system. This view is well expressed by the Asian Development Bank in its Disaster
Manager’s Handbook (Carter, 1992).

There is, however, disagreement in the literature regarding the advisability of depend-
ing on national governments as the appropriate foundation for a comprehensive pro-
gram. As described in the classic analysis of the political economy of large disasters by
Albala-Betrand (1993), focusing natural disaster policy through existing government
systems enhances narrow power structures and draws away from local concerns and
initiatives. Those holding this view favor reducing natural hazard risk through commu-
nity-driven projects and programs developed by nongovernmental organizations. Such
an approach to risk management is not guaranteed to be comprehensive, but applies
directly to identifiable needs and the empowerment of local populations.

These two approaches to risk management need not be mutually exclusive. The task
facing policymakers is to create an effective national system with a comprehensive vi-
sion that engages senior government policymakers and accommodates and supports
local decision-making and private market initiatives.

While it is helpful to discuss specific programs in terms of both promising and

problematic practices, these must be understood in the context of the specific country
or region being discussed and the perspective of the policymaker interpreting the
practice. For example, a practice that permits the integration of probabilistic risk from
catastrophe events into macroeconomic planning models would be a promising practice
for those interested in engaging development planners in the dialogue. In contrast,

1
those interested in creating decentralized projects would consider it problematic to have
a complicated macroeconomic modeling process that diverts significant institutional
energy and financial resources from locally directed initiatives.

Elements of Comprehensive Disaster Management

In the risk management literature, the key elements of risk management are divided
into two phases: the pre-disaster phase and the post-disaster phase. Table 1.1 divides
the key components of disaster risk management into two phases: actions required in
the pre-disaster phase and actions needed in the post-disaster period. The pre-disaster
phase includes risk identification, risk mitigation, risk transfer, and preparedness; the
post-disaster phase is devoted to emergency response and rehabilitation and recon-
struction. A comprehensive risk management program addresses all these components.

Risk Identification

Risk identification includes hazard assessment, vulnerability studies, and risk analysis.
Hazard assessment identifies the probable location and severity of dangerous natural
phenomena and the likelihood of their occurring within a specific time period in a given
area. These studies rely heavily on available scientific information, including geologic,
geomorphic, and soil maps; climate and hydrological data; and topographic maps, ae-
rial photographs, and satellite imagery. Historical information, in the form of written
reports and oral accounts from long-term residents, also helps characterize potential
hazardous events. To be most successful, hazard assessment requires data and scien-
tific teams trained to evaluate the data. In some countries, the lack of extensive histori-
cal data on catastrophic events makes hazard assessment difficult. In the case of floods
and landslides, human factors can drastically impact the environment, and historical
data may be of little value. For earthquakes and tropical cyclones, the international re-
search community has collaborated significantly to pool resources and scientific knowl-
edge to develop global and regional hazard maps. Much work remains to be done on
flood and landslide mapping.

Vulnerability studies estimate the physical, social, and economic consequences that re-
sult from the occurrence of a natural phenomenon of given severity. Physical vulnerabil-
ity studies analyze impacts on buildings, infrastructure, and agriculture. The Applied
Technology Council, for example, publishes detailed vulnerability curves for the resis-
tance of 50 different types of structural facilities to earthquake hazards (ATC, 1985).
Social vulnerability studies estimate the impacts of especially vulnerable groups, such
as the poor, single parent families, pregnant or lactating women, the mentally or physi-
cally handicapped, children, and the elderly. Social vulnerability studies take into ac-
count the public awareness of risk, the ability of groups to self-cope with catastrophes,
and the institutional structures in place to help them cope (Coburn, Spence, and
Pomonis, 1991).

Economic vulnerability studies estimate the potential impacts of hazards on economic

assets and processes. These studies include indirect losses (such as business interrup-
tion) and secondary effects (such as accentuated poverty, higher unemployment, or in-
creases in levels of external debt). The United Nations’ Economic Commission for Latin
America and the Caribbean (ECLAC) has contributed significantly to this effort by pub-
lishing reports since 1972 on the economic impacts of catastrophes in Latin America
and the Caribbean (Caballeros and Zapata Marti, 2000).

2
The risk analysis stage of risk identification integrates information from the hazard as-
sessment and the vulnerability studies in the form of an estimate of the probabilities of
expected loss for a given hazardous event. Formal risk analyses are time-consuming
and costly, but shortcut methods are available that give adequate results for project
evaluation (Bender, 1991). In the United States and Europe, a large part of the funding
for risk modeling comes from the private sector; major reinsurance companies commis-
sion projects from private modeling firms such as EQECAT (www.eqecat.com) and RMS
(www.rms.com). However, these private sector initiatives require a guarantee that in-
vestment in risk identification will lead to the development of insurance markets.

Table 1.1. Key Elements of Risk Management

Pre-disaster phase Post-disaster phase

Risk Risk Emergency Rehabilitation

identification Mitigation transfer Preparedness response and
reconstruction
Hazard as- Physi- Insurance Early warning Humanitar- Rehabilitation
sessment (fre- cal/structural and re- systems and ian assis- and reconstruc-
quency, magni- mitigation insurance communication tance tion of damaged
tude, and loca- works of public systems critical infra-
tion) infrastruc- structure
ture and
private
assets
Vulnerability Land-use Financial Contingency Clean-up, Macroeconomic
assessment planning and market planning (util- temporary and budget
(population building instru- ity companies repairs, and management
and assets ex- codes ments (ca- and public ser- restoration (stabilization and
posed) tastrophe vices) of services protection of so-
bonds and cial expendi-
weather- tures)
indexed
hedge
funds)
Risk assess- Economic Privatiza- Networks of Damage Revitalization for
ment (a func- incentives for tion of emergency assessment affected sectors
tion of hazard pro- public ser- responders (exports, tour-
and vulnerabil- mitigation vices with (local and na- ism, and agricul-
ity) behavior safety tional) ture)
regulation
(energy,
water, and
transporta-
tion)
Hazard moni- Education, Calamity Shelter facili- Mobilization Incorporation of
toring and training and Funds (na- ties and of recovery disaster mitiga-
forecasting awareness tional or evacuation resources tion components
(GIS, mapping, about risks local level) plans (public, in reconstruction
and scenario and preven- multilateral, activities
building) tion and insur-
ance)
Building and strengthening national systems for disaster prevention and response: These systems are an
integrated, cross-sector network of institutions addressing all the above phases of risk reduction and disaster
recovery. Activities that need support are policy and planning, reform of legal and regulatory frameworks,
coordination mechanisms, strengthening of participating institutions, national action plans for mitigation poli-
cies, and institutional development.
Source: IDB (2000b).

3
In Latin America and the Caribbean, several international partnerships have formed to
help provide risk assessment studies. These include a joint World Bank-Organization of
American States (OAS) project in St. Lucia, St. Kitts and Nevis, and Dominica (Vermei-
ren and Pollner, 1994) and a World Bank study on Mexico (Kreimer and others, 1999).
The Natural Catastrophes and Developing Countries Project at the International Insti-
tute for Applied Systems Analysis (IIASA) also developed a methodology for incorporat-
ing natural disasters into macroeconomic projections as a function of a country’s under-
lying social and economic vulnerability, and presented results for Argentina, Honduras,
and Nicaragua (Freeman and others, 2001). Although these projects are useful, they
are isolated examples and do not diminish the need for national strategies for risk iden-
tification.

Some excellent resources on risk identification include Smith (1996) and the brochures
produced by Swiss Re and Munich Re, available on their websites (www.swissre.com
and www.munichre.com). More details are available in the documentation on the U.S.
Federal Emergency Response Agency’s (FEMA) HAZUS model, available online at
www.fema.gov/hazus/. Resources on vulnerability assessment include Blaike and others
(1994) and Coburn, Spence, and Pomonis (1991).

Mitigation

Mitigation refers to policies and activities that reduce an area’s vulnerability to damage
from future disasters. These structural and nonstructural measures are in place before a
disaster occurs.

Structural Mitigation Measures

Structural mitigation reduces the impact of hazards on people and buildings via engi-
neering measures. Examples include designing infrastructure, such as electrical power
and transportation systems, to withstand damage. Underground transmission lines, for
example, are protected from hurricane damage. Levees, dams, and channel diversions
are all examples of structural flood mitigation.

Structural mitigation projects can be very successful from a cost/benefit perspective. In

the field of landslide mitigation, experience in the city of Los Angeles, California, indi-
cates that adequate grading and soil analysis ordinances can reduce landslide losses by
97 percent (Petak and Atkisson, 1982). Argentina’s Flood Rehabilitation Project invested
US$153 million in structural improvements that spared an estimated US$187 million (in
1993 dollars) in damages during the 1997 floods, generating a 35 percent return on in-
vestment to date (World Bank, 2000).

However, structural mitigation projects have the potential to provide short-term protec-
tion at the cost of long-term problems. In areas in Vietnam, flood control systems have
exacerbated rather than reduced the extent of flooding; sediment deposit in river chan-
nels has raised the height of river channels and strained dike systems. Now when floods
occur, they tend to be of greater depth and more damaging than in the past (Benson,
1997b).

Furthermore, structural mitigation projects have the potential to provide people with a
false sense of security. The damages from the 1993 flooding of the Mississippi river in
the United States were magnified because of misplaced confidence in structural mitiga-
tion measures that had encouraged development in high-risk areas (Mileti, 1999; Platt,
1999; Linnerooth-Bayer and others, 2000). To avoid this problem, structural mitigation

4
projects should be accompanied by appropriate land-use planning and public awareness
programs.

Nonstructural Mitigation Measures

Nonstructural mitigation measures are nonengineered activities that reduce the inten-
sity of hazards or vulnerability to hazards. Examples of nonstructural mitigation meas-
ures include land use and management, zoning ordinances and building codes, public
education and training, and reforestation in coastal, upstream, and mountain areas.
Nonstructural measures can be encouraged by government and private industry incen-
tives, such as preferential tax codes and deductibles, or adjusted insurance premiums
that reward private loss-reducing measures. Nonstructural mitigation measures can be
implemented by central authorities through legislating and enforcing building codes and
zoning requirements, by NGOs initiating neighborhood loss-prevention programs, or by
the private sector in providing incentives to take loss-reducing measures. Nonstructural
mitigation measures are particularly appropriate for developing countries because they
usually require fewer financial resources.

A drawback to such measures, however, is that even when they exist, there is a ten-
dency on the part of the private and public sectors not to enforce the regulations or
standards on the books. For example, in Florida, insured property losses from Hurricane
Andrew would have been reduced by 25 percent through building code compliance.
Studies have found that inspection personnel have insufficient knowledge of the hazard
mitigation aspects of the building codes to enforce them effectively. The problem is
compounded because of limited staffing so that even competent individuals cannot keep
up with the demand for building inspections.

Another controversial area in nonstructural mitigation is land-use planning. Unplanned

growth in major cities caused devastating earthquake losses in Turkey’s heavily urban-
ized northeastern region in 1999; in Orissa, India, where cyclones left eight million
people homeless; and in Mozambique, where flooding in 1999 devastated the capital
city of Maputo (Sanderson, 2000). According to ECLAC, 75 percent of the losses of
goods and services during Hurricane Mitch resulted from land-use issues like building
too close to rivers or constructing roads and bridges in known vulnerable areas (ECLAC,
1998). On the other hand, land-use planning requires intense political support if it af-
fects property values or involves the relocation of communities. Less costly and less
controversial land-use initiatives involve passing hazard disclosure laws for real estate
purchases and/or promoting insurance policies with premiums that scale with risk.

The best practices in nonstructural mitigation are those that directly combine with de-
velopment goals. An innovative model recently developed in the Grau region of Peru
identifies hazards, assesses regional development objectives, and integrates a non-
structural approach to disaster mitigation into the overall development program. This
“microzonation” approach focuses on land-use planning and infrastructure (Kuroiwa,
1991). The World Bank has tailored neighborhood improvement programs to the needs
of the most vulnerable by helping residents of low-income urban areas improve their
houses individually or with community help (World Bank, 2000).

A good source on mitigation measures appropriate for each hazard is Smith (1996).
FEMA also maintains a very useful website with information on both structural and non-
structural mitigation measures: http://www.fema.gov/fima. The Caribbean Disaster
Mitigation Project has published extensively on mitigation practices in that region

5
(http://www.oas.org/ en/cdmp/publist.htm). Davis and Gupta (1990) identify exem-
plary mitigation practices in Asia.

Risk Transfer

A fundamental distinction between risk management policies in the developed world

and those in developing countries is the role of risk transfer. In developed countries,
entities other than the government absorb a portion of the risk of financing reconstruc-
tion after a disaster, often an insurance company. The use of insurance, the primary
risk transfer tool, has five key advantages: it permits the spreading of risk between
parties; it reduces the variance of risk for each person; it allows the segregation of risk;
it encourages loss reduction measures; and it provides a tool to monitor and control
behavior (Freeman and Kunreuther, 1997).

Insurance is not the only option for transferring risk. In dealing with natural disasters, a
recent innovation in transferring risk of loss from catastrophes is a hedging instrument
known as catastrophe bonds. Collectively, insurance and catastrophe bonds may be de-
scribed as “catastrophe hedges.” An extensive discussion of the use of catastrophe
bonds in developed countries with some insight as to how they may work for developing
countries can be found in Andersen (2001).

Risk transfer is a critical component of a comprehensive program for most developing

countries. Japan, France, Spain, the United Kingdom, and the United States all use risk
transfer to link the various components of their natural disaster risk strategy. Insurance
is a major component of the risk management strategy of wealthier countries. In the
higher-income countries, 30 percent of the loss from natural hazards is insured. In the
poorer countries, insurance covers 1 percent of the losses from natural hazards.

Existing insurance programs have a limited range. For example, they are not used to
finance the post-disaster reconstruction of government-owned buildings. In most low-
income countries, the government relies on its power of taxation and on borrowing to
fund the reconstruction of government-owned facilities. In addition, the government
continues to fund the needs of the poor after a disaster, although the poor are not part
of formal insurance programs. In most countries in Latin America and the Caribbean,
insurance is designed to transfer the risk of property owners and businesses from the
government to the insurance program. In countries with a strong middle class and ac-
tive privately owned businesses, the use of the program can be an effective policy tool
to reduce the government’s obligation to fund post-disaster needs.

Promising and Problematic Practices of Risk Transfer Strategies

The main attractions of a national risk transfer policy are shifting the risk of post-
disaster reconstruction funding away from the government and providing incentives to
mitigate risk. There is considerable worldwide activity in promoting different schemes to
use the government as a tool to provide catastrophe risk shifting for homeowners and
others. The creation of the recent Turkish Catastrophe Insurance Pool is a good exam-
ple. All existing and future privately owned property is required to contribute to it. The
payments made will contribute to a fund that will pay homeowners up to US$28,000 in
the event that a catastrophe damages their homes (Gulkan, 2001). Proposals are being
explored in Mexico, the Caribbean, Central America, and Africa to engage the govern-
ment in providing risk transfer options for farmers, homeowners, and businesses in
case of natural catastrophe losses (World Bank, 2000). The Caribbean Disaster Mitiga-
tion Project commissioned a study to explore insurance options for small states in the

6
region (Pollner, 2000). The World Bank has proposed the creation of a new insurance
program for Honduras, and the Inter-American Development Bank, pursuant to the
Puebla-Panama Initiative, is considering regional insurance options for Central America.
The most recent World Development Report on poverty devotes considerable attention
to the role of insurance in enabling countries to better deal with risk, including the risk
from natural catastrophes (World Bank, 2000). Insurance also has two key disadvan-
tages. While there are instances where insurance has contributed to loss reduction,
there is an associated moral hazard that insured parties will actually take fewer meas-
ures to reduce risk. Furthermore, it should be kept in mind that insurance is costly and
the funds spent on insurance have an opportunity cost since they could be spent on
other social projects, including risk mitigation measures.

The reduction of risk works to the benefit of the developing countries that directly bear
the losses from catastrophes and the international aid community whose mission is to
assist the long-term development and reduction of poverty in these countries. By har-
nessing the private sector to cope with catastrophe risk, the international aid commu-
nity frees itself and its resources to implement its broader agenda of development poli-
cies.

Barriers to Supplying Catastrophe Insurance

It is no coincidence that insurance is an economic tool used by wealthy countries. It re-

quires sophisticated financial institutions to operate and a complex series of laws, regu-
lations, and administrative agencies. These include the proper financial structure of in-
surance companies to ensure their financial capacity to pay future claims, the actuarial
science (including the required information base) that underpins the setting of premi-
ums and reserves, legal knowledge about insurance contracts and the protection they
provide, the functioning of insurance distribution networks, and claims payment prac-
tices and proper legal institutions to enforce sophisticated contractual agreements. In
many developing countries, the lack of institutional regulatory structures hinders the
ability to acquire insurance.

Designing major institutional reforms to permit the proper operation of financial institu-
tions is difficult. The components needed to implement an adequate regulatory scheme
for insurance industries are already known. Guidelines for proper regulatory practices
are maintained by appropriate agencies in developing countries. The National Associa-
tion of Insurance Commissioners in the United States has detailed information on
proper regulatory practices (see www.naic.org).

In addition to the regulatory issues, there are concerns related to the fundamental
structure of the market for insurance. For example, many countries may be too small to
provide adequate risk diversification to properly support a national insurance scheme.
Proposals to create regional insurance markets hope to increase risk diversification and
potential market size, thereby making the market more attractive for the insurance in-
dustry and lowering the cost of insurance. A larger potential market subject to a uni-
form regulatory scheme may encourage the international insurance industry to help de-
velop viable markets. Regional proposals, like the World Bank’s initiative for a Central
American insurance market, are based on overcoming barriers to the supply of insur-
ance.

7
Demand for Insurance in Poorer Countries

One problem with developing risk transfer as an effective policy tool is a lack of demand
for catastrophe insurance. In poorer countries, large-scale businesses can and do buy
catastrophe insurance. In Mexico, nearly 100 percent of industrial enterprises buy in-
surance. With the region’s small middle classes and medium-sized businesses (the most
frequent purchasers of insurance in developed countries), there already exists a small
natural clientele.

Professional risk bearers, like insurance companies, are fully capable of modifying their
products to adapt to local needs. However, there will be little willingness on their part to
do so if no demand exists for the modified products. One approach to creating demand
is to make insurance mandatory. Another approach is to demonstrate the benefits of
insurance by taking out policies at the government level, for example by insuring gov-
ernment-owned buildings and infrastructure. The World Bank’s recent initiatives have
focused on insuring government assets as a way to provide protection and stimulate
interest in risk transfer (Pollner, 2000). The advantages and limitations of commercial
risk transfer are summarized in Box 1.1.

Box 1.1. Advantages and Limitations of Commercial Hazard Insurance

Advantages

• It guarantees the victim some predictable recompense after loss. Such compensation is more re-
liable than disaster relief and it also appeals to those opposed to excessive government regulation
because it depends on the private market.
• If property owners in hazard areas pay premiums that reflect their actual risk and insurance pay-
ments fully compensate the victims, then insurance provides an equitable distribution of costs and
benefits.
• Although insurance is designed to redistribute losses, it can also be used to reduce hazard impact
by encouraging the adoption of measures designed to minimize damages. If residents in hazardous
areas pay the full cost of premiums for their risk, insurance provides an economic disincentive for
locating in such areas. Once properties have been built, it is possible, in principle, for insurers to
offer lower premiums to policyholders who take measures to reduce risks to their property. Such
measures might include special construction methods and building materials. In extreme cases,
insurers could require property owners to retrofit risk reduction measures before accepting any pre-
mium.

Limitations

• In practice, property owners in hazard areas rarely pay premiums that reflect their actual risk.
One reason is that for many environmental hazards, the database is insufficient to devise a realistic
premium based on predicted average annual losses at a specific site. Unless premium rates are
scaled directly according to the risk, hazard zone occupants are not likely to bear the full cost of
their location.
• In the private residential sector, a great deal of development is undertaken by speculative build-
ers rather than by the eventual occupants of the property. Only if insurance premiums became suf-
ficiently high to make the properties initially difficult to sell would it be likely that developers would
be deterred from building on such sites in the first place.
• Private insurance may be unobtainable in very high-risk areas, although this does not necessarily
discourage development.
• Even when commercial hazard insurance is available, there is frequently a low voluntary uptake.
When insurance policies are taken out, a significant number of policyholders are underinsured and
are unlikely to be fully reimbursed by the company in the event of a loss.
• Although insurance can, in some circumstances, be employed to reduce losses, the existence of
moral hazard is thought to increase damages. Moral hazard arises from the tendency of some in-
sured persons to reduce their level of care and thus change the risk probabilities on which the pre-
miums were based. Moral hazard can be lessened by the imposition, and subsequent policing, of
local planning regulations designed to strengthen buildings against hazard impact.

Excerpted: Smith (1996).

8
Because of the relatively high administrative costs associated with insurance for small
values, insurance is not an option for the very poor. The main strategy for the poor re-
quires a poverty-sensitive policy that focuses on several key components: helping poor
households maintain their consumption; ensuring that the poor do not lose whatever
access they have to basic social services; preventing permanent reversals in the accu-
mulation of human and physical capital; and averting self-defeating behavior, such as
criminal activity, prostitution, and exploitative forms of child labor (World Bank, 2000).
Among the most effective programs are workfare programs introduced or expanded in
the disaster area in conjunction with post-disaster reconstruction.

Froot (1999) is a good collection of articles on risk transfer. Pollner (2000) is a good
resource describing risk transfer options for the Caribbean.

Preparedness

Preparedness involves building an emergency response and management capability be-

fore a disaster occurs. Key disaster preparedness activities include training programs
for response personnel, exercises and drills of emergency plans, education programs to
inform citizens, hazard detection and warning systems, identification of evacuation
routes and shelters, maintenance of emergency supplies and communications systems,
establishment of procedures for notifying and mobilizing key personnel, and individual
household measures such as clearing attic space to make room for belongings in case of
a flood.

In contrast with elements such as mitigation that are often the product of major policy
decisions at a national level, preparedness projects tend to be oriented toward the ac-
tions of individuals and individual organizations. Programs must therefore focus on the
community level and a national system should include mechanisms to coordinate with
preparedness projects.

Disaster preparedness also requires significant political will. According to Smith (1996),
“it ties up facilities and people that are apparently doing nothing, other than waiting for
an event that no one wants and many believe will never happen.” It is inherently diffi-
cult to maintain impetus for diverting resources into preparedness projects if many
years have passed since the last disaster event. Outdated plans and warning systems,
however, have the potential of being worse than no provisions at all. Continued public
awareness programs are therefore a key ingredient in increasing and maintaining disas-
ter preparedness (Foster 1980; Garb and Eng, 1969). Public awareness is increasingly
important as populations become more mobile and newcomers are less aware of local
risk conditions and traditional mitigation techniques (UNDHA, UNDP, and MWR 1994).

Many programs can be used to increase public disaster awareness. Broadcasting agen-
cies can contribute to increasing public awareness by designing announcements and
disaster-related programs. Inclusion of disaster awareness in school programs is a par-
ticularly efficient and economical strategy. Other successful practices include advertising
at popular sporting events, on shopping bags, or during community programs; hosting
workshops; and organizing national disaster preparedness days.

An excellent resource is FEMA’s website on disaster preparedness publications

(http://www.fema.gov/library/publicat.shtm), which includes pamphlets on
home/family and business/industry disaster plans, disaster supplies, and emergency
preparedness checklists.

9
Emergency Response

Emergency response refers to actions taken immediately before, during, and after the
onset of a major disaster or large-scale emergency to minimize the loss of life and harm
to people and their property and enhance the effectiveness of recovery. Examples of
emergency response activities include hazard detection and warning, evacuation of
threatened populations, shelter for victims, emergency medical care, search and rescue
operations, security and protection of property, and family assistance. Other examples
include the construction of temporary levees, closure of roads or bridges, provision of
emergency water or power supplies, and response to secondary hazards such as fire or
the release of hazardous materials. The quality and timeliness of disaster response are
typically functions of the planning and training done during pre-disaster preparedness.

From decades of experience, it is clear that the best emergency response comes imme-
diately and with sufficient resources to limit the loss of life and property. Experience in
numerous disasters reveals the need for a strong, centralized system to mobilize emer-
gency efforts and channel aid resources to victims (Red Cross, 2001).

In his seminal work, Cuny (1983) recommends that the emphasis on speed or “emer-
gency response” should shift to developing a response relevant to needs at an interme-
diate or advanced phase of recovery. Cuny summarizes other important emergency re-
sponse lessons, which include considerations of livelihood protection for the poor, edu-
cation and local participation, the appropriate actors in emergency and relief efforts,
and issues related to longer-term rehabilitation. Anderson and Woodrow (1989) provide
another excellent work with similar recommendations.

A good resource on emergency response is the website hosted by the Caribbean Disas-
ter Emergency Response Agency (CDERA) at www.cdera.org. For emergency response
related specifically to hurricanes, see www.huracan.net.

Reconstruction and Rehabilitation

Reconstruction and rehabilitation refer to programs that provide longer-term assistance

for people who have suffered injuries or incurred losses due to a major disaster. The
objective is to facilitate the return of these communities to their pre-disaster condition.
Rehabilitation encompasses repairing and reconstructing houses, commercial establish-
ments, public buildings, lifelines, and infrastructure; restoring and coordinating vital
community services; expediting permit procedures; and coordinating activities among
governments. Recovery can take a few weeks or several years, depending on the disas-
ter’s magnitude and the reconstruction resources available.

The most important recommendation for reconstruction and rehabilitation projects is

that they should proceed in ways that reduce future vulnerability and promote devel-
opment objectives. It is less costly to incorporate structural mitigation components into
new structures than it is to retrofit existing ones. Ideally, mitigation measures are un-
dertaken during reconstruction to avoid recreating prior vulnerable conditions. One
good example is the reconstruction and mitigation program undertaken in Peru by the
NGO Caritas. In consultation with affected communities, this NGO promoted the use of
local earthquake-resistant materials for housing reconstruction. To directly assist the
most needy households, Caritas used a work-for-materials program in which locals re-
ceived materials in exchange for participation in community projects. An earthquake the
following year proved the success of the project: most houses built during the Caritas

10
project withstood the earthquake measuring 6.2 on the Richter scale (Schilderman,
1993).

Successful reconstruction projects involve the cooperation and participation of the local
communities and stakeholders. The September 1985 Mexico City earthquake provides a
good example of a participatory process for effective reconstruction that reduced future
vulnerability. As a part of this approach, social teams represented and included victims
in the redesign and reconstruction of housing (Kreimer and Echeverria, 1998).

The reconstruction and rehabilitation process should not ignore the importance of pro-
viding for livelihood protection throughout the recovery process. Successful reconstruc-
tion and rehabilitation programs simultaneously address both the need to provide in-
come support and the need to reconstruct. After the floods in Gujarat, India, workfare
community reconstruction projects provided both needed work and income protection
for poor families as well as necessary reconstruction activities (Bhatt, 2001). Similarly,
after the 1985 earthquake in Mexico City, workfare programs created more than
175,000 jobs for victims of the event (Kreimer and Echeverria, 1998).

Country Experiences with Disaster Management Systems

Many countries have, or are developing, national programs to partly or fully incorporate
the six elements of an integrated disaster risk management program and to involve the
public sector, market actors, NGOs, and private sector actors. These national disaster
systems reflect the political and economic cultures and conditions of the countries. Each
one combines the public sector with private market institutions, as well as incorporates
the diverse institutional practices that fall outside either sector.

The Asian Development Bank’s Disaster Manager’s Handbook, based on the experiences
of its member countries, suggests recommendations for the design of a national disas-
ter strategy and for supporting legislation (Carter, 1992). The handbook proposes the
creation of a formal national disaster strategy and an organizational structure for inte-
grating disaster management efforts. As illustrated in figure 1.1, this organizational
structure typically includes a ministry or sub-ministry responsible for disaster affairs,
that houses a national disaster management office, some form of national disaster
council to identify priorities and channel resources, and an operations control group re-
sponsible for preparing and coordinating emergency response. This organizational strat-
egy, which is reflected in a recent program instituted in Nicaragua, is markedly hierar-
chical in nature. Although the proposed organizational strategy incorporates NGOs and
local groups, the participation of market actors (such as insurance companies and other
financial institutions) is absent.

While a government-directed program is essential for an integrated disaster manage-

ment program, it is not the only way in which societies cope with disasters. Private citi-
zens and social groups are managing risks outside of government-directed systems in
their choices of where to live, what crops to plant, how to respond to neighbors in need,
and how to help with rescue operations. The private industry is also developing and im-
plementing risk management strategies alongside governments, and insurance indus-
tries may even be taking the lead in assessing risks and vulnerability.

11
 Figure 1.1. National Disaster Organizational Structure, as suggested by the Asian
Development Bank

CABINET

National International
Disaster
Assistance
Council
National Disaster
Management Office

PREPAREDNESS / RESPONSE RECOVERY

MITIGATION

Operations Control
Group Technical Advisory
Team

National Emergency
Operations Center

Government Government Government

Departments / NGOs
Departments / NGOs Departments / NGOs

Provincial Disaster Provincial Disaster

Committees Committees Provincial
Disaster
Committees
Local Local
Government Government
Committees Committees

COMMUNITIES DISASTER AREAS RECOVERY

PROGRAMS

Source: Adjusted from Carter (1992).

In balancing commitments and values, society needs “clumsy” or responsive institutions

to maintain a set of values over time. It is important to develop integrated clumsy pro-
grams that include the active participation of all the important institutional players in
the process: ministries of finance, health, and education; military organizations (such
as civil defense); regional and local government entities; many diverse NGOs active in
the risk management system (such as the Red Cross); international aid and finance or-
ganizations; private sector actors; and local communities.

This section provides several country-specific reviews of national risk management sys-
tems, starting with Latin America and the Caribbean, and extending to Europe, Asia,
Japan, and the United States. These examples demonstrate the wide variety of compre-
hensive systems that address the full spectrum of disaster risk management alterna-
tives.

12
Disaster Management Systems in Latin America and the Caribbean

Most national disaster management systems in Latin America began as government

bodies for emergency response. Some countries, like Ecuador, Peru and Venezuela,
continue to rely almost exclusively on civil defense.1 In other countries, large natural
catastrophes over the past decades have highlighted the need for more comprehensive
systems that would include prevention, mitigation, preparedness, and provisions for re-
construction and rehabilitation activities. As a result, several countries, such as Argen-
tina, Brazil, Chile, Colombia, the Dominican Republic, El Salvador, Nicaragua, Guate-
mala, Honduras and Mexico have begun to transform their approaches to coping with
natural disaster risk.

The past decade has also seen an increasing trend toward regional efforts in disaster
management. Several entities have formed to share information and technologies
across countries. The Coordination Center for the Prevention of Natural Disasters in
Central America was established in 1988 to strengthen the capacity of that region. In
the English-speaking Caribbean, CDERA works to improve disaster response and na-
tional and regional disaster management. In South America, the Andean Development
Corporation, in response to a request from five members, is developing a Regional An-
dean Program for the Prevention and Mitigation of Risk.

The countries in Latin America and the Caribbean that have broadened the scope of
their national disaster systems to encompass preparedness, mitigation, relief and reha-
bilitation activities, and, in some cases, even attention to pre- and post-disaster financ-
ing options, have taken different routes. Figure 1.2 illustrates three broad approaches.
Most countries, like Chile and Colombia, have increased the scope of disaster manage-
ment by expanding the responsibilities of an existing institution such as civil defense.
Other countries, like El Salvador, broadened the government’s mandate for disaster risk
management by creating a parallel institution responsible for mitigation and prepared-
ness. Finally, a third approach, the one taken by Mexico, is to bring in, strengthen, and
reinforce a network of key institutions.

Figure 1.2. Approach for obtaining integrated disaster management systems in Latin
America and the Caribbean.

A new institution is created to

facilitate new activities/attention
focus, while the previous institu-
tion continues monitoring activi-
ties such as emergency response.

Tackling the problem as a network

The existing organization for disaster links several different institutions
management widens the field of action
responsible for the disaster manage-
and includes new activities/attention ment system.
focus.

1
DNDC in Ecuador, INDECI in Peru, and Protección Civil in Venezuela.

13
The strengths and weaknesses of these organizational approaches depend on the larger
context in which they operate. Whether centralized, loosely centralized, or networked,
public programs should operate in a system with sufficient input, feedback and control
from the private sector, including actors in the marketplace and civil society. The Japa-
nese government, for example, has deliberately decentralized its public program to in-
clude a network of national institutions. This diffusion of power in the public domain was
a reaction to the conditions of martial law in the postwar period. While protecting citi-
zens against their loss of liberties in the case of a national emergency, the diffused sys-
tem failed to provide a timely and effective response to the Kobe earthquake.

Throughout Latin America and the Caribbean, the key obstacles faced by most nations
lie in institutional resistance to moving beyond emergency response, the limited in-
volvement of civil society and the private market, and insufficient ex ante provisions for
reconstruction financing.

Argentina

In 1998, Article 99 of the Argentine Constitution established the Federal Emergency

System (Sistema Federal de Emergencias, SIFEM) to coordinate national efforts to miti-
gate natural disaster risk, develop a plan for effective emergency response, and aid in
efforts to rehabilitate affected regions. Formed under the Chief of the Cabinet of Minis-
ters, SIFEM is the principle federal political body responsible for coordinating the efforts
to identify and mitigate the risk of natural disasters at the national, provincial, and mu-
nicipal levels. Whereas the federal government plays an important role in the mobiliza-
tion of resources and the coordination of national and international organizations in
mitigation and emergency response efforts, provincial governments assume the respon-
sibility for assessing regional vulnerabilities and carrying out mitigation projects to pro-
tect against natural catastrophes within their respective provinces. This decentralized
approach emphasizes the proactive role of local governments and appears to be a com-
promise between control at the national level, which is essential for emergencies, tem-
pered with strong decision authority at the provincial level, which has the potential to
include local interests and participation.

In recent years, Argentina has invested significantly in mitigation. In 1998, SIFEM dedi-
cated a total of US$420 million in loans from the World Bank to be used for mitigation
projects such as zoning regulations, seismic mapping and codes, reforestation, and
clearing fire corridors.

Argentina is unique in that it created ex ante an entity at the national level that will be
in charge of allocating funds for reconstruction projects. In 1998, Argentina passed Na-
tional Resolution 496/8 creating the National Advisory Board for the Recuperation of
Regions Affected by Natural Disasters (Consejo Nacional para la Recuperación de Zonas
en Emergencia Climática, CONAREC) to oversee the post-disaster rehabilitation and re-
building of affected communities. One of CONAREC’s principal objectives is to coordi-
nate and distribute funding to provincial and municipal authorities to aid in the rebuild-
ing of infrastructure such as homes, businesses and roads in the aftermath of a disas-
ter. Formed by representatives from several provincial governments, CONAREC serves
as a medium between national and provincial efforts. As for funding, there is a limited
penetration of insurance firms for disaster risk insurance. Argentina depends heavily on
national reserves of credit and international lending organizations to provide needed
liquidity in the face of natural disasters.

14
Brazil

Disaster response in Brazil is highly decentralized and proceeds from the bottom up
with minimal coordination from the national government. In the event of a natural dis-
aster, the affected municipality handles its own response. When the scope of the disas-
ter exceeds the municipality’s capacity to respond, the regional office is called in, then
the state, then the national level. This separation of powers is attributed to the Brazilian
legal structure, which ascribes a great deal of autonomy to the state and local govern-
ments.

At the national level, the disaster management plan is known as the National Civil De-
fense System (SINDEC). Coordination of SINDEC falls to the National Secretariat of Civil
Defense (SEDEC), which is connected to a branch of the Ministry of National Integra-
tion. SEDEC bears responsibility for coordinating disaster management across all levels
in Brazil. Beneath SEDEC are multiple entities at the regional, state, municipal and
lower levels that are responsible for disaster response and coordination in their areas.

Brazil is another country that enlarged its civil defense organization to include ex ante
disaster measures after large-scale floods and landslides struck the state of Rio de Ja-
neiro in 1966. It continues to evolve at the local level through the involvement of NGOs
and multilateral organizations.

Chile

Chile is a classic example of a national system that became increasingly comprehensive

through the expansion of a single entity’s mandate. The Office of National Emergencies
(Oficina Nacional de Emerencia, ONEMI), the administrator and coordinator of the na-
tional system in Chile, was created after the 1965 earthquake in central Chile.2 In con-
trast to several other Latin American systems based under ministries of defense, ONEMI
has been under the aegis of the Ministry of the Interior since 1970. After the response
to the 1965 earthquake, the system grew to incorporate other elements of comprehen-
sive disaster management. It now emphasizes preventive and mitigation strategies,
paying attention to vulnerability as a key intervention factor in risk management.

Although it appears centralized and hierarchical, the Chilean system is composed of

committees at the community, provincial and regional levels that are responsible for
evaluating proposed actions and designing and prioritizing prevention, mitigation and
preparedness projects appropriate to each administrative level. In the case of an emer-
gency, all of the available resources in the affected community are used first. If the
magnitude of the event exceeds the local capacity, additional resources are mobilized
successively from the provincial, regional and national levels.

Colombia

The national disaster system in Colombia was created after the Nevado del Ruiz
eruption and the destruction of Armero in 1985. Colombia broadened the disaster
management paradigm beyond emergency response by creating the National System
for Risk Mitigation and Disaster Preparedness (Sistema Nacional para la Prevenvión y
Atención de Desastres, SNPAD), led by the Presidency of the Republic. SNPAD
encourages participation from a network of scientific, planning, education, and

2
Oficina Nacional de Emergencia, Ministerio del Interior, www.onemi.cl.

15
emergency response institutions, and it expands the mandate of provincial and
municipal committees. It not only coordinates emergency response, but also helps
determine policy with regard to ex ante preventive and mitigation measures. Figure 1.3
shows the interrelationships among the actors in the system.

Colombia’s model of expanding an existing institution to embrace more aspects of dis-

aster management has been promoted by many international organizations in Latin
America. However, mitigation activities tend to relate to reconstruction rather than to
risk reduction, and trying to avoid the perpetuation of vulnerability has been a chal-
lenge (Red Cross, 2001). Although not feasible at the present time, the private market
could play a role in the future to promote risk reduction activities.

Other important obstacles faced by the Colombian national system are the lack of
strategies to finance reconstruction and the historical tendency to channel reconstruc-
tion resources in ways that circumvent the national system. After the Paez earthquake
in 1999, for example, the Presidency created a new entity, Fondo para Reconstrucción y
Desarrollo Social del Eje Cafetero (FOREC), to supervise all of the reconstruction efforts.

Dominican Republic

In 2000 the IDB and the Secretary of the Presidency of the Dominican Republic began
investing close to US$12 million to develop a more comprehensive disaster manage-
ment system in the country, with a specific focus on mitigation and prevention activi-
ties. Prior to this effort, disaster management fell exclusively under the control of the
military-based civil defense organization. Proposals for the new Dominican system most
closely resemble the system currently in place in Colombia, yet they effectively involve
creating a new parallel—if not superior—entity that would act as an ex ante counterpart
to the ex post emphasis of the civil defense (Lavell, 2001b). There is considerable insti-
tutional resistance to the new proposal. It is still too early to evaluate the successes
and shortcomings of the new system that is being developed.

El Salvador

The Committee of National Emergencies (Comité de Emergencia Nacional, COEN) has

been the principal organization for disaster management, but it has focused almost
solely on emergency response, working closely with the Salvadoran Armed Forces and
other rescue organizations. Following the 2001 earthquakes, and spurred by interest in
risk reduction on the part of the government, NGOs, the United Nations, and municipal
associations, El Salvador created a new technical entity for risk management within the
Ministry of Environment and Natural Resources. This entity has a broad range of duties,
particularly related to mitigation, in coordinating comprehensive disaster management
with COEN. In addition, because of its ministerial location, the new entity will be able to
impart a broad vision of risk analysis and disaster management to important economic
sectors. The success of this vision will be critically dependant on the strength of the
ministry, and also on the institutional legal arrangements that provide needed checks
and balances on a centralized system. The government should be conscious of the im-
portance of building private sector institutions that can eventually play a parallel or
perhaps coordinated role in managing disasters.

16
Figure 1.3. Organizational Chart of Colombia’s Natural Disaster Management System

National Calamity Fund FNC

National Committee
Presidency of the Republic
Ministers: Interior, Finance, Defense, Health, Communications, Transportation, Environment*
Directors: National Department for Planning, Colombian Civil Defense, Colombian Red Cross Consulting Board of the
President’s Representatives: Camacol, Colombian Society of Construction Engineers
G l Di t f th P ti d Att ti t Di t DGPAD FNC
Minister of the Interior*
Minister of Finance*
Minister of Health*
Minister of Transportation*
Minister of Agriculture*
Bank Superintendent *
National Technical Committee DGPAD General Secretary for the Presidency of the Republic*
General Directorate for National Firemen DGPAD Director *
Disasters Prevention Board Colombian Civil Defense Director *
General Director for Prevention and Attention to Disasters
(President) CNT Colombian Red Cross Director *
Undersecretary of the Consulting Board FNC
National Delegation
Ministers: Defense, Health, Transportation, Environment
Designated Officials from:
DNP, Colombian Civil Defense, Ideam, INAT, Telecom, In-
urbe, Sena, Ingeominas, Igac, Colombian Red Cross, Inver-
Departmental National Operative Com-
Delegaction
nar, National Firemen Board Delegate, CTN Secretary: one
DGPAD official
mittee
Firemen Colombian Civil Defense Director *
G t (President)

National Advisory Committees National Services DGPAD Director *

Integrated Delegates: Minister of Health
Information System Colombian Red Cross
National Firemen Board
Watersheds and Human Set- National Seismic National
Environmental tlements and Network Communications
Cleaning Housing National Communications
System
Advisory Committees National
Education Health
Southwestern
Seismological
Regional Committee for the Search and Rescue
National Reserve
Centres Network
Observatory OSSO
Prevention and Attention to
Reservoirs and
Awareness
Raising
Vulcanological Disasters Communications
National Food
Distribution Network
Dams Observatories Gobernor (CREPAD President)
Programs
Commander: Military Unit or Brigade, National Police, CON-PNC
Development Industrial and Hydro- Firemen Protection against oil, gas CISPROQUIM
Planning and Technological metereological Alert Directors: Ministry of Health, Regional Autonomous and other Dangerous Sub- Center for Information
Environment Risks Network l b l f l b d stances Spills on Chemical Products

Seismic and Tsunamies or Tsunami or Tidal

Technical Education Operative
Vulcanic Risk Tidal Waves Wave Detection
Commission Commission Commission National Computer
Network
System Network
Geographic Information
CTN-PNC Local Committees for the Prevention and Attention to System
Pacific Coast Against Hidro- National Catalogue of Disasters PNC
Phenomena carbon and Spill Control
Dangerous Teams and Experts Municipal Mayor (CLOPAD President)
Substances PNC
Spills Commander: Military Unit or Brigade, National Police, Fire-
men, Chief of Sectional Health Service
Permanent One representative of: Colombian Civil Defense, Colombian
Advisory Red Cross
Committee on Capital City Mayor
Earthquake Two Representatives of the Mayor
Resistant CLOPAD Secretary, Chief of Planning
Construction

Technical Education Operative

Commission Commission Commission

Community
Guatemala

The National Coordinator for Disaster Reduction (Coordinadora Nacional para la Reduc-
ción de Desastres, CONRED) was created in 1996. The Higher Council of CONRED in-
volves disaster response organizations, development agencies, and (as an important
example to other systems) representatives from civil society. CONRED has a mandate
to engage in risk mitigating activities, but lacks the financial backing or human re-
sources to be as effective as other countries in the region. The approach in Guatemala
is more comprehensive and integrated than a single institution, but is not yet at the
level of a full national system (Lavell, 2001a).

Honduras

Honduras has followed an approach of widening its existing framework beyond emer-
gency response. With the support of the World Bank, the national system in Honduras,
the Permanent Commission for Contingencies (Comisión Permanente de Contingencias,
COPECO), is exploring new disaster management strategies that place greater emphasis
on prevention and mitigation. A law has been drafted to expand COPECO’s focus, and
includes responsibilities on prevention and mitigation implied by its proposed new
name, National System for the Prevention, Mitigation and Attention to Emergencies and
Disasters. An issue faced in Honduras and other countries attempting this expansion
(such as Nicaragua) is to blend the new paradigm of preparedness with the old focus on
response. While laws may change the mandate of these systems, dominant actors in
responding to disasters (such as civil defense) may make the shift in paradigm difficult
to implement. This has been the experience of Nicaragua’s civil defense and Institute of
Regional Studies, which are both highly effective in emergency response, but less re-
ceptive to the newer focus on mitigation and preparedness (Lavell, 2001a).

Mexico

Mexico increased the public sector’s role in disaster risk management through a net-
work approach. The government established the National Civil Protection System (Sis-
tema Nacional de Protección Civil, SINAPROC) in 1986 as the main mechanism for in-
teragency coordination of disaster efforts. SINAPROC is responsible for minimizing the
loss of lives and property and the interruption of essential social services caused by dis-
asters. Responsibility for the system lies with the General Coordinating Body for Civil
Protection in the Ministry of the Interior, but the system is networked in that the coor-
dinating body synchronizes the technical work of various ministries, for example, the
Ministry of Social Development for Geologic Hazards. In 1990, the National Council for
Civil Protection was added to SINAPROC. The council is an advisory, planning, and co-
ordinating committee and is headed by the president of Mexico and made up of 12 min-
isters plus the mayor of the Federal District of Mexico City. The network also includes
the National Center for Disaster Prevention (CENAPRED), a unique institution located on
the campus of the National Autonomous University of Mexico that reports directly to the
Directorate of Civil Protection of the Ministry of the Interior. CENAPRED serves as a link
between research work on natural disasters and policymakers and is involved in both
research and information dissemination. The Mexican government allocates budgetary
funds for disaster relief and reconstruction efforts by placing them in the Fund for Natu-
ral Disasters (FONDEN), which provides for the repair of uninsured infrastructure, im-
mediate assistance to restore the productivity of subsistence farmers, and relief to low-
income victims of disasters (Kreimer and others, 1999).

18
Venezuela

In the mid-1990s, Venezuela began exploring the Colombian disaster management

model, but eventually created the National System of Civil Defense instead. The system
is similar to a previous approach that focused on municipal emergency response. Mu-
nicipal fire departments have played a central role in Venezuela’s history of disaster
management, particularly in Caracas. A principal difference between the old and new
systems was a heightened role for civilians vis-à-vis military participation. Since 1999,
however, the government has begun to shift involvement back toward the military (par-
ticularly during the mudflows in the state of Vargas). After the disaster in 1999 and the
constitutional change, the government began exploring a new model with greater em-
phasis on risk mitigation.

Both the science and technology ministries have participated in disaster management,
and their target is to work more in territorial zoning and land-use policy as well as vul-
nerability reduction. The UNDP has supported these activities, as has the Andean Cor-
poration for Development.

Caribbean States

The English-speaking islands of the Caribbean established the Pan Caribbean Disaster
Preparedness Project (PCDPP) in 1981 to improve national and regional disaster man-
agement. Although it was conceived as an 18-month project, focused solely on disaster
preparedness, the PCDPP operated for almost 10 years (Poncelet, 1997). In 1989, when
the project extended its work to the prevention of disasters, its acronym was changed
to the Pan Caribbean Disaster Preparedness and Prevention Project (PCDPPP).

Disaster preparedness offices have been created in several locations, such as the Cen-
tral Emergency Relief Organization in Barbados, the National Emergency Management
Agency in Trinidad and Tobago, and the Office of Disaster Preparedness and Emergency
Management in Jamaica. In 1991 the Caribbean Community (CARICOM) approved the
creation of CDERA to replace the PCDPPP, providing a new agency funded by member
states and donor agencies and responsible for mobilizing resources among CARICOM
countries (www.cdera.org). CDERA commands a stronger institutional position than the
PCDPPP, including the right to mobilize the military (such as the CARICOM Disaster Re-
sponse Unit). CDERA was created to improve disaster response and national and re-
gional disaster management. Its main focus is disaster preparedness, but it also pro-
motes risk mitigation activities.

Obstacles in Latin America and the Caribbean

Three important issues that affect the success of comprehensive disaster management
in Latin America and the Caribbean emerge from this brief discussion. These include a
continued concentration on emergency response with institutional obstacles that slow
the shift toward a more integrated system, a limited involvement of private market ac-
tors and civil society, and limited provisions ex ante for financing reconstruction.

Institutional resistance to moving beyond emergency response. In Latin America and

the Caribbean, the primary focus of disaster management—apparent in legislation, pro-
gram structures, and policy—has been emergency response (Maskrey, 1993; Blaike and
others, 1994). It appears difficult to incorporate activities related to prevention and
mitigation in the legislation, planning and institutions that deal with territorial and sec-
tor issues related to disasters (Lavell, 2001b). The challenge in the region is to find

19
ways to promote and concentrate on risk mitigation in an environment dominated by
institutions created specifically to respond to emergencies rather than to reduce risks.
To ensure the success of the transition, new integrated approaches face the challenge
of involving rather than antagonizing traditional emergency response actors. The case
of Colombia illustrates that institutional structural changes within the government are
not, by themselves, sufficient. In addition, there must be political will to carry out the
intent of the institutional restructuring, as well as checks and balances originating from
outside the government system. Actors in the private sector and NGOs can serve this
function; however, they have largely been absent from the restructuring plans in Latin
America.

Limited involvement of civil society and the private sector. For most countries in the re-
gion, disaster management remains dominated by central government institutions and
lacks the opportunities provided by—and constraints imposed by—nongovernmental ac-
tors in civil society and the market. Insurance still plays a limited role in Latin America
and the Caribbean. Although inroads in regional reinsurance and insurance projects
have been undertaken, insurance premiums—according to Vatsa and Krimgold, 2000—
are still beyond the disposable income of most of the population. These authors report
that most homeowners (excluding those in Barbados), as well as small- and medium-
sized businesses, do not carry insurance except when required to do so by lending insti-
tutions. The supply of insurance is also a problem. The Caribbean Disaster Management
Project carried out by the U.S. Agency for International Development and the Organiza-
tion of American States (OAS) showed that an issue in improving the insurability of as-
sets is that local insurance companies and agencies in the region retain little of the risk
they are insuring (Vermeiren, 2000). The OAS has led an effort to improve underwriting
practices in the region, which may improve the participation of market mechanisms in
disaster management in the coming years.

Insufficient ex ante provisions for reconstruction financing. Many countries have reserve
funds for emergency operations, such as Fundação Cearense de Apoio ao Desen-
volvimiento Científico Tecnológico (FUNCAP) in Brazil, or the National Calamity Fund in
Colombia, but few have designated entities responsible for carrying out reconstruction
and relief, let alone for providing funds to do so. Several problems arise from this situa-
tion. Some countries may find themselves unable to fully reconstruct their key infra-
structure or provide for the very poor after a disaster. This lack of funds for reconstruc-
tion and rehabilitation could have a ripple effect both on the national economy and in
allowing the poor to drop even further below the poverty line.

Furthermore, mitigation only occurs when the interests of the ultimate risk bearer are
aligned with the party incurring the cost of mitigation. If reconstruction financing is left
out of the equation, policies directed at reducing risk through a culture of prevention
lack the attention they require.

Finally, if there is no ex ante plan for channeling reconstruction and relief funds, their
distribution can easily become a highly politicized task appointed to organizations out-
side the national system, thus undermining the credibility of the system and hindering
its ability to later engage in ex ante risk preventing and mitigating activities.

Summary of Latin American and Caribbean Experiences

There are three general approaches to implementing more comprehensive national dis-
aster management programs at the government level: some countries expand the
mandate of existing entities, others create parallel institutions, and others strengthen

20
the network between existing and new institutions. The strengths and weaknesses of
these organizational approaches depend on the larger context in which they operate.
Whether centralized, loosely centralized, or networked, the success of public programs
depends on the input, feedback and control of the private sector, including actors in the
marketplace and civil society.

Market-like tools, such as reinsured catastrophe funds (Mexico) and mitigation-focused

insurance schemes (Barbados), have been implemented in a few countries. There is an
increasing awareness of the importance of including civil society (particularly munici-
palities, NGOs, and other stakeholders) and private markets in disaster management
solutions, and international financial organizations have provided support for pilot pro-
grams in the region (Vermieren, 2000). However, governments remain the most impor-
tant driving force for developing comprehensive disaster management. Difficulties faced
by the region include institutional reluctance to engage significantly in pre-disaster
mitigation and financing arrangements and organizational obstacles to involving market
institutions and civil society.

Disaster Management Systems in Europe

As countries in Latin America and the Caribbean face the challenges of changing their
focus from emergency response to broader and more comprehensive disaster manage-
ment systems, it is instructive to examine the experiences of developed and transition
countries in Europe, in this case Hungary, France, and the United Kingdom.3

Hungary

Hungary has a system of disaster security for all, which is funded by taxpayers and has
only recently involved private insurers. The government has traditionally and obligato-
rily compensated victims of flooding for up to 100 percent of their losses. Indeed, until
the transition from communist governance in 1989, central government control and
planning dominated the political landscape in Hungary. A cursory look at the country’s
recent past raises an important warning against the dominance of the public sector in
disaster risk management.

In Hungary, flood mitigation and defense have been the responsibility of the National
Water Authority (which today is part of the Ministry for Transport and Water Manage-
ment) and 12 regional water authorities. Until recently, this centralized state system
dominated all activities in mitigating, preparing for, and responding to floods. During
the state socialist period, the water management authorities were a large and powerful
bureaucracy, with a staff numbering more than 30,000. Not surprisingly, this unchecked
authority expanded its resource base by advocating and carrying out extensive and ex-
pensive levee-building programs throughout the country. To date, more than 4,000 km
of levees protect 97 percent of Hungary’s flood risk areas. The overriding management
philosophy was to protect the Hungarian territory rather than to institute land-use con-
trols or less costly, nonstructural measures. This goal has motivated the governments
of Hungary and many developed countries to invest heavily in structural mitigation
measures, especially to reduce losses from flood hazard. With hindsight, many of these
measures inadvertently increased flood losses, damaged ecological systems, and led to
the loss of credibility of the responsible government authorities. Hungary’s experience

3
The Hungarian case is based on Horváth and others (2001) and Vári, Ferencz, and Linerooth-Bayer (2001);
the French case is based on Gilber and Gouy (1998), Michel-Kerjan (2001), and Linnerooth-Bayer and others
(2000); and the United Kingdom discussion is based on Linnerooth-Bayer and others (2000).

21
highlights the dangers of a policy process that excludes conflicting values and critical
views.

The recovery process in Hungary, which has also been dominated by the central gov-
ernment, is placing more responsibility on private insurance systems. However, the pri-
vate insurance option is unpopular among many people who prefer the solidarity of na-
tional compensation and are concerned about the effects of privatization on the poor. A
clumsy policy approach is developing, which combines the solidarity of government in-
volvement, the personal responsibility established by insurance, and the meaningful
participation of NGOs and other actors in civil society.

France

The case of France offers a different perspective. Since 1982 private insurers in France
have been required to offer catastrophic natural disaster insurance bundled with prop-
erty insurance, and to charge a fixed rate set by the French treasury. Since more than
90 percent of all businesses and homeowners carry property insurance, the mandatory
bundling of catastrophe insurance guarantees wide distribution. In fact, the insurance
operates as a tax on property to fund the French national fund. Because rates are not
differentiated by risk level, there are cross subsidies from persons in low-risk areas to
persons in high-risk areas. Private insurer risks, in turn, are partly ceded to the French
national fund, the Central Reinsurance Fund, to which the state gives its guarantee.

It is widely recognized that the French system provides disincentives for individuals and
local communities to take risk reduction measures. A recent and imaginative decree to
counter this problem sets a deductible that increases with the number of disasters in
the same area. This means that the compensation a household or business receives will
continually decrease in high-risk areas, leading to incentives to relocate or take other
mitigating measures.

A problem with moral hazard remains from the lack of market-style incentives that ac-
company risk-based premiums, and the French have dealt with this in a clumsy fashion
by relegating mitigation to the government. The government sets land-use restrictions
and other mitigation measures. Since 1982, the government has carried out a survey of
areas susceptible to natural disasters, and has instituted construction controls in these
areas. Still, the required accompanying risk prevention plans have never been success-
fully implemented. To date, there are only about 5,000 such plans compared with
36,000 French municipalities, many of which are at risk from flooding, earthquakes,
subsidence, or avalanches. Two reasons have been given for this: first, the cost is high;
second, communities resist risk estimates because they can affect property values.

United Kingdom

The French notion of solidarity contrasts markedly with the disaster management phi-
losophy of the United Kingdom. Without any anticipation of public relief, there is an un-
usually high penetration of natural hazard insurance in the United Kingdom (some esti-
mate this at close to 70 percent), which is greatly facilitated by the automatic bundling
of all-perils coverage into household insurance policies. What is remarkable about this
solely private arrangement is that the insurance companies have an unwritten agree-
ment to avoid risk-based premiums in favor of a standard premium for disaster cover-
age. This has resulted in substantial cross subsidies across regions and perils, making
insurance affordable to the poor, who mainly populate high-risk flood plains, and has
alleviated the government from political pressure to compensate poor victims after an

22
event. As efforts intensify within the industry to estimate the risks of disasters, insurers
are moving toward risk-based premiums that will inevitably lead to reduced coverage
for low-income households and ultimately to demands for more government compensa-
tion in the wake of a major disaster.

Disaster Management Systems in Japan and the United States

Like most Latin American countries, Japan and the United States are vulnerable to
many types of catastrophic natural disasters, especially earthquakes, windstorms, and
flooding.4 While the countries are culturally diverse, the similarities in the evolution of
their disaster management systems are striking. Both countries have comprehensive
programs at the national level to manage disaster risks. FEMA is renowned for its ef-
forts at centralizing and coordinating disaster management components at the national
level. In Japan, the 1978 Large-Scale Earthquake Countermeasures Act created a na-
tional program and also set the institutional conditions for increased private market in-
volvement. The program that emerged in Japan from this law was centralized and bu-
reaucratic. The coordination was set out under the auspices of several national govern-
ment ministries, and as the Kobe earthquake highlighted, depended on overly diffused
responsibility between the authorities involved. It was three days before national civil
defense forces reached the site of the earthquake, mainly because authority for sending
civil defense troops rested with the provincial governors and not the central authorities.
Ironically, Japan has not given the kind of power to its central government for disaster
response that the United States has.

Another parallel between Japan and the United States, and perhaps the most innova-
tive, is the creation of public/private insurance systems to further recovery. Both coun-
tries have pioneered loss-sharing programs that involve government and private market
institutions. In Japan, earthquake risk insurance is offered by private insurers as a part
of fire insurance policies; in the United States, a similar but importantly different pub-
lic/private partnership exists to cover flood losses. The National Flood Insurance Pro-
gram (NFIP) is unique in that policies are offered by the private sector, but the national
government assumes the risks and automatically plays the role of reinsurer. Moreover,
the NFIP puts far greater emphasis on deductibles as a way of encouraging policyhold-
ers to take loss-reducing measures. Thus, a notable difference in the public/private in-
surance partnerships of these two countries is the greater emphasis on incentives for
individual responsibility found in the United States, as shown in box 1.2.

Government agencies in the United States and Japan also interact with market actors in
other important ways. With research support from the government, the private market
in both countries is taking initiatives for prevention. For instance, Japan Railway has
pioneered UrEDAS (Urgent Earthquake Detection and Alarm System), an information
system that detects the arrival of P-waves near the source and estimates the location
and magnitude of the earthquake. Similarly, in the United States, FEMA has developed
HAZUS, a multihazard tool with models for estimating potential losses from earthquake,
wind, and flood hazards, that is an effort to place multihazard risk models in the public
domain. The various stakeholders concerned with managing disaster risk in the United
States are depicted in figure 1.4, although the basic structure is likely to be the same in
most developed countries.

4
This Japan discussion is based on Elahi (2001) and EQE (1995).

23
Box 1.2. Examples of Public/Private Insurance in the United States

Floods
In 1973 the U.S. Congress passed the Flood Disaster Protection Act, which gave flood-
prone communities the choice of participating in the National Flood Insurance Program
(NFIP) or forfeiting federally subsidized insurance and all but emergency forms of disaster
relief. Once a community agrees to participate in the program, homes and businesses lo-
cated in the 100-year flood plain are required to purchase flood insurance as a condition
for a federally insured mortgage on their property. This increased the demand for flood
coverage considerably.
The NFIP has a combination of requirements (for example, land-use regulations
and building codes) for communities participating in the program. By restricting the loca-
tion and design of buildings in relation to the 100-year flood plain to meet NFIP standards,
the local communities are taking positive steps to reduce future flood losses. The NFIP re-
quires the cooperation of the federal, state, and local governments with the private prop-
erty insurance industry. It is the clearest example in the United States of a public/private
partnership for dealing with natural disasters.

Earthquakes
Although there has been a series of damaging earthquakes in California since the 1971 San
Fernando quake, none of them compare to the January 17, 1994 Northridge earthquake.
After Hurricane Andrew, it caused the largest insured damage of any disaster in the United
States, with total insured losses of more than US$12.5 billion. The insured damage from
Northridge led insurance companies to question whether earthquakes were an insurable
event. This concern was heightened by the large increase in demand for earthquake cover-
age following this disaster. As a result, the state-run California Earthquake Authority (CEA)
was established in 1996, whereby private insurers and reinsurers have a maximum loss of
US$8 billion with the CEA setting rates, marketing policies, and settling claims.
Elsewhere in the United States, earthquake insurance in all states (except Califor-
nia) is offered as a separate endorsement to an insurance policy. For commercial struc-
tures, earthquake protection for property damage coverage is often included as part of a
multiperil policy.

Figure 1.4. Major Actors to Manage Disaster Risk in the United States

F ed eral
R isk tran sfer G o v ern m en t
In stru m en ts A g en cies
C ap ital M ark ets (e.g ., F E M A )

C atastro p h ic C atastro p h ic
L o ss L o ss
P ro tectio n P ro tectio n In su ran ce an d
P rim ary In su ran ce M itig atio n
C o m p an ies R eq u irem en ts

In su ran ce H azard In su ran ce

F in an cial P ro tectio n
R eq u irem en ts
In stitu tio n s

P ro p erty at
R esid en tial
R isk S ecto r

E n fo rcin g m itig atio n m e asu res

B u ild in g In fo rm atio n
S ecto r In sp ectio n s S tate an d L o cal
C ertificatio n G o v ern m en ts

24
Disasters Risk Management in Other Countries in Asia

Fiji

Fiji has extensive natural catastrophe exposure from cyclones, floods, droughts, earth-
quakes, and tsunamis.5 Fiji’s national disaster management program began as an ad
hoc government committee for emergency response, but by 1990, the national program
was restructured to make it more comprehensive. It now covers prevention, mitigation,
preparedness and rehabilitation activities in addition to emergency response. In 1995
the government published the National Disaster Management Plan (Government of Fiji,
1995), which laid out a comprehensive policy and detailed the supporting roles of NGOs
in all the functions of disaster management. However, the equally important roles of
tourism, industry and commerce did not receive recognition. Fiji has a thriving tourism
industry, and it is not surprising that private insurance has a high uptake in the busi-
ness sector, whereas there is less but still significant insurance coverage for private ur-
ban dwellings.

Insurers in Fiji also take a proactive role in mitigation and prevention. After particularly
severe cyclones in 1984, the Commissioner of Insurance established the Fiji Building
Standards Committee, made up mainly of private insurers. This committee has the re-
sponsibility to oversee the preparation of a National Building Code that would set mini-
mum standards to reduce disaster-related losses and help achieve a stable or reduced
hurricane insurance premium (Government of Fiji, 1995). Of particular interest is that
upgraded homes are inspected by a structural engineer and issued a certificate, which
is required to obtain cyclone insurance coverage and mortgages. Most urban areas have
adopted the building code (Rokovada and Vrolijks, 1993).

India

By contrast to Fiji, in India market actors are not very involved. Private disaster insur-
ance exists, but there is little reliance on the private market for financing relief
(Hoogeveen, 2000). The authorities at the state level take the main responsibility for
disaster relief with financial assistance from the central government. A small Calamity
Relief Fund (CRF) has been constituted with contributions from both the state and cen-
tral governments. If a disaster overwhelms the capacity of the state government to re-
spond, the central government will provide financial and other assistance. If such a ma-
jor disaster occurs, the central government commits itself to pre-fixed reimbursement
sums for loss of life, limb, and partial and total loss of housing and productive assets.

Interestingly, India is developing a more loosely networked system with little attention
to mitigation at the government level. NGOs play an active role in risk reduction
programs in the region. A new innovation in India is the so-called “knowledge network”
that involves civil society, the scientific community, and, to a minor extent, the market.
The National Natural Disaster Knowledge Network has been designed to facilitate an
interactive, simultaneous dialogue with all the players dealing with natural disasters.
Indian NGOs, such as the Disaster Mitigation Institute, are also working with the
government, as well as the Grameen Bank, in designing tools to address disaster loss
and poverty. In addition, India appears to have a great deal of innovation from the
private sector. Micro-insurance mechanisms are being designed to reach the poorest
groups, build institutional capacity, and form the capital necessary for disaster
management targeted toward the poor. Finally, in Gujarat, workfare programs and

5
The Fiji case is based on Benson (1997a) and Carter, Chung, and Gupta (1991).

25
community reconstruction projects have provided needed work and income protection
for poor families as well as necessary reconstruction activities (Bhatt, 2001).

Lessons Learned from Disaster Management Systems Worldwide

In Latin America, governments are taking important steps to expand their reach from
disaster response to include other risk management functions, especially disaster pre-
vention and mitigation activities. However, countries throughout the region are encoun-
tering institutional resistance to the implementation and allocation of resources to these
broader government programs. Moreover, while the formal systems include NGOs,
there is little involvement by market actors such as insurance companies. This is not
surprising given the lack of legal institutions in this region to support insurance and
other financial services.

The Hungarian case shows that government programs, while essential for coordinating
the functions of disaster management, must not be so powerful as to exclude other ac-
tors holding conflicting values and views. This example serves as an important warning
to Latin American and Caribbean countries as they expand their bureaucracies to en-
compass more comprehensive disaster risk management functions, and especially as
they pursue the laudable goal of shifting resources from post-disaster functions to the
prevention of disaster losses. Today, countries like the United States, Japan, and Hun-
gary are taking active steps to incorporate public dialogue and participation in mitiga-
tion policies and, more generally, in their disaster management programs.

Alternatively, the United Kingdom illustrates the drawbacks of relying almost exclusively
on the private market for transferring risks through its fully private insurance system.
As this system evolves toward risk-based premiums, the poor living in high-risk areas
will be excluded from social protection. Insurance at the formal level (although it exists
informally through family support systems) is currently not realizable in most Latin
American and Caribbean countries due to the lack of market structures and institutions
in which private financial institutions can operate.

The United States and Japan have perhaps the most structured programs at the na-
tional level to manage disaster risks, but they have pioneered loss-sharing programs
that involve government and private market institutions. These programs underline the
importance of integrating hierarchical and individualistic forms of social organization in
a national system, but in ways that reflect national cultures.

Finally, the developing countries in Asia offer additional lessons and insights that are
valuable for the design of comprehensive disaster management systems. In India, de-
spite the recently announced formation of a national disaster management authority,
the government is active mainly in aiding states in their response to catastrophes. Fiji
has moved remarkably toward an integrated, clumsy system for disaster risk manage-
ment that includes a strong national program and plan, involvement by private insur-
ers, and a keen awareness that NGOs and local efforts are an integral part of the sys-
tem. Both India and Fiji have moved toward a balance of social cultures in the man-
agement of disaster risk.

Designing Effective National Disaster Management Systems

Drawing from country experiences and the recommendations of policymakers involved

in national disaster systems, some key guidelines for effective disaster management
systems emerge. First, successful national systems have an explicit and appropriate

26
national disaster strategy. Appropriate national disaster strategies are integrated with
national policy on development and environmental protection and are based on
vulnerability assessments. Second, successful national systems integrate key players in
the disaster management process. Such players include, among others, the finance
ministry, local community leaders, NGOs, and private market actors. Third, successful
national systems have provisions to ensure sufficient resources for key players to carry
out their responsibilities.

Explicit and Appropriate National Disaster Strategy

Involvement of a wide variety of stakeholders and policy tools is needed for developing
any disaster management program. Interested parties are brought together and policy
tools are combined depending on the nature of the institutional arrangements in the
country as well as the types of disasters the country faces. For example, in Turkey it
would be difficult to institute a partnership between insurers and financial institutions in
promoting mitigation measures because mortgages on property do not exist. Each
country must design a national system appropriate to its own circumstances.

Integrating a Strategy with a National Development Policy

Economic development, environmental protection, and disaster management are all in-
trinsically linked. A large part of the damage caused by Hurricane Mitch in Honduras
and Nicaragua can be traced to poor land-use practices and uncontrolled human settle-
ment (Bate, 1999; Olson and others, 2000). The deforestation and rural-urban migra-
tion that created such high vulnerability to Mitch were largely due to the extensive pov-
erty in the area. The aftermath of the hurricane further set back the poor. Successful
national disaster strategies recognize the linkages between the poor, the environment,
and natural disasters. Raising standards of living means not only guaranteeing access
to basic needs, but also reducing the risk to natural hazards in people’s lives.

Vulnerability Assessment

Vulnerability assessments are an essential part of the process of integrating natural dis-
aster strategy with overall development objectives. These assessments identify sources
of risk, vulnerable groups, and potential interventions. In the first stages of establishing
the new national system in Nicaragua, the Nicaraguan government and the World Bank
commissioned a study from a consulting company to integrate existing hazard maps
with poverty maps to identify communities most in need of disaster-related mitigation
and preparedness projects. This identification process served as a cornerstone for all of
the initial risk management activities. Vulnerability assessment allows policymakers to
specifically define the objectives of the risk management programs and to establish
vulnerability reduction targets.

Incorporating Key Players

Successful national systems integrate key players into the disaster management proc-
ess. The most successful systems take advantage of the existing government structure
and involve national, provincial, local, and community government as well as ministries
and other institutions. Essential institutional players are ministries (such as the minis-
tries of finance, health, and education), organizations (such as military units and civil
defense), regional and local government entities, NGOs (such as the Red Cross), inter-
national aid and finance organizations, private sector actors, and local communities.

27
Key to the success of these systems is the interaction between the coordinating bodies
and institutional players.

It is essential to understand the values, goals, and objectives of the relevant stake-
holders in a national system and recognize that they may conflict with each other. The
challenge is to construct a program that is viewed as more desirable than the status
quo by these key interested parties. There also needs to be recognition that programs
in place prior to a disaster may be greatly modified after a catastrophe occurs.

Finance Ministries

Finance ministries are important players in disaster management systems. The partici-
pation of finance ministries helps to ensure funding for the institutional framework, fa-
cilitates the incorporation of disaster management into development policy, and pro-
vides incentives for financing mitigation projects. Incorporating natural catastrophes
into development policies requires including the costs of disasters in macroeconomic
projections, future budgets, and the project investment process. Finance ministries are
responsible for preparing projections, allocating budgets, and approving investments;
they can and should incorporate the costs of natural hazards into each of these stages.

In many developed countries, the finance ministry is engaged in the disaster manage-
ment process through its responsibility to provide post-disaster reconstruction financ-
ing. The need to fund the repair and reconstruction of buildings, housing, and infra-
structure focuses attention on pre-disaster risk management practices. Finance minis-
ters are naturally interested in supporting ex ante activity, including mitigation and pre-
vention measures, which reduce their defined obligation to generate more funds for re-
construction. By linking risk prevention to reduced need for post-disaster funding, the
finance ministries have an economic stake in maintaining the integrated risk manage-
ment process. However, not until the finance ministry has the responsibility to manage
the government’s post-disaster reconstruction funding obligation will it have a strong
interest in participating in the risk management process.

Communities

Local communities also play a key role in successful disaster management systems.
When political impetus behind the national disaster system in Colombia flagged, local
community efforts continued. To allow for feedback from communities into the national
level disaster management decision process, there should be an avenue for active par-
ticipation of NGOs and representatives of civil society in the organizational structure of
the national system.

Providing Sufficient Resources

Obtaining sufficient resources for disaster management is hindered by three elements:

providing financial resources for the ongoing operation of a national system; providing
incentives for funding risk mitigation projects; and financing post-disaster reconstruc-
tion.

Funding Ongoing Operations

The creation of a budget for ongoing operations is a component of a national system. Of

course, the level of funding and its year-to-year sustainability are directly related to the
involvement of the economic planners in the process of creating and implementing the

28
system in the first place. A well conceived strategy will identify the necessary functions
and the resources required to perform those functions. If the finance ministers are not
actively involved, the national system will become minimized. As a result, over time
funding for the system will be reduced.

Financial Incentives for Mitigation

Mitigation occurs when the interests of the ultimate risk bearer are aligned with the
party incurring the cost of mitigation. For example, a homeowner is more likely to take
mitigation measures to reduce the exposure of his home to hurricanes if he must pay
the cost of reconstructing after the hurricane hits. If the homeowner believes that
someone else will pay the cost of reconstruction, he has no incentive to bear the cost of
mitigation. Similarly, unless the finance ministers have an economic stake in the cost of
disasters in the form of bearing responsibility for providing economic resources for post-
disaster reconstruction, it is unlikely they will fund money for mitigation from their re-
sources. Of course, this is true for all levels of government, and emphasizes the need
for finance ministries to articulate and bear responsibility for the post-disaster financing
needs of a country.

Insurers have also tried to provide economic incentives to encourage residents and
businesses to purchase coverage and adopt cost-effective risk mitigation measures, but
with limited success. One way to make a premium reduction (which is associated with
the property owner undertaking mitigation measures) financially attractive to the prop-
erty owner is for the bank to provide funds for mitigation through a home improvement
loan with a payback period identical to the life of the mortgage. If the annual premium
reduction from insurance was greater than the annual loan cost, then the insured
homeowner would have lower total payments by investing in cost-effective mitigation
than by not doing so.

Financing Losses and Risk Transfer

There is limited activity associated with financing losses and risk transfer in Latin Amer-
ica and the Caribbean. Several policy alternatives are available to encourage the use of
risk transfer in those cases where it is a desirable alternative. The most obvious first
step is institution building and developing the necessary information to support an in-
surance program. As discussed in the risk identification section, this entails catastrophe
modeling. The second step is creating demand for catastrophe insurance. This report
discusses possible policy alternatives to address both the supply and demand for risk
transfer. The adoption of these alternatives in specific countries should be a major pol-
icy interest.

Another pre-disaster alternative to developing an insurance program is the creation of a

disaster fund to meet short-term needs in the post-disaster period. This is also an addi-
tional approach used by Mexico and is proposed for Nicaragua. The catastrophe fund
avoids the problem of seeking a new budget allocation in the middle of a fiscal year to
meet anticipated annual needs. In fact, it is not a tool for transferring risk. Rather, it is
a means for guaranteeing funding to meet short-term needs without requesting new
appropriations in the middle of a budget cycle. It is an effective tool for focusing poli-
cymakers at the national level on the need for consistent funding for disasters. As is
now being explored for Mexico, it may provide a tool for exploring risk transfer as a pol-
icy option.

29
Finally, governments should actively explore risk management strategies to cope with
the post-disaster needs of the poor. Having a clear strategy on the obligations of the
government to meet the needs of the poor after a disaster, as well as a program to ad-
dress those obligations, is essential. While it is unlikely that risk transfer can play a role
in meeting this need, the interest of the government in looking at risk transfer to meet
other post-disaster obligations may free resources to help the poor.

Sustainability of National Systems

It is important for national systems not only to function well, but also to survive periods
in which relatively few catastrophic events occur and then remain viable during and af-
ter major hazard events. To be sustainable, national systems must function effectively
and have the continuous provision of political and financial resources. As policymakers
know, programs that are sustained have well defined objectives, resources to accom-
plish these objectives, and well stated goals. Systems that do not meet their objectives
will not be sustained.

Political Sustainability

Integration with Overall Development Goals

The more closely a national disaster system is integrated with overall development
goals, the easier it will be to maintain political interest in the system. Programs survive
changes in the political leadership when they are tied to long-term economic develop-
ment. Programs not essential to economic development have difficulty maintaining their
status in hard economic times. Natural disaster policy must find its place as a problem
of economic development demanding the year-in and year-out attention of those con-
cerned with a country’s economic well being. The long-term survival of a national sys-
tem therefore requires that those responsible for development planning be key partici-
pants in both the creation and ongoing operation of the system.

Legislation

Supporting a national disaster strategy with legislation increases the likelihood it will be
sustainable. Legislation provides a formal basis for counter-disaster action, allocates
major responsibilities in legal form, and provides a measure of protection for govern-
ments, organizations, and individuals by outlining the limited responsibilities of each in
the disaster management process. To the extent that legislation supporting the national
disaster strategy is designed as a consensus-building process, it will also increase the
likelihood of long-term support from the participants.

The Disaster Manager’s Handbook includes examples of legislation for the Cook Islands,
Papua New Guinea, and Queensland, Australia (Carter, 1992). The Nicaraguan govern-
ment publishes online the legislation it recently passed to establish its national system
(see www.sosnicaragua.gob.ni/Download/).

Continued Perception of Risk

A key problem facing stakeholders concerned with developing a sustainable disaster

management program is that prior to a disaster there is little interest on the part of po-
tential victims to protect themselves against the consequences of these events because
they feel “it cannot happen to me.” Political units sense this lack of concern and hence
place natural hazards low on their agenda. Following a major disaster, there will be

30
sympathy and concern for victims by the general citizenry and a desire by elected offi-
cials to offer disaster assistance to those in need. This reaction makes it less likely for
those at risk to protect themselves in the future.

At some fundamental level, individuals and governments underestimate natural hazard

risk. The failure to recognize the concept of risk means that action tends to be fueled by
events, rather than by the probability of the event occurring. For example, every major
change in the natural disaster policy of the United States followed a major disaster
(Kunreuther and Roth, 1998). Even then, the impetus for policy change tends to be of
short duration, with a focus on solving a short-term issue rather than tackling long-term
institutional failures.

As discussed in this chapter, various strategies can be used to promote continued public
awareness of disaster risk. Successful initiatives involve radio and television broadcast-
ing agencies in designing informative programs and including disaster awareness in
school programs. Public awareness of disaster risk is essential to sustaining programs,
particularly through periods of few disaster events.

Creation of a Constituency

The sustainability of a national system requires a constituency to maintain disaster

management on the political agenda and counteract the reluctance of existing power
structures, including civil defense organizations, to change policy. Constituencies also
act as a force to prevent the national system from being captured by political groups in
times of crisis. One way to develop a natural constituency for disaster management is
to frame the process as a poverty issue. Natural disasters impact the poor more than
any other group; the research on this issue is compelling. Those concerned about the
reduction of poverty need to become advocates for changing national policy toward dis-
asters. This includes the bilateral, multilateral, and NGO community. The advocacy
group for a natural disaster system needs to be more than the disaster management
professionals. Protection of the poor from the risk of natural disasters should be a
mainstream poverty issue.

Accountability

Finally, a key part of political sustainability is providing the structures to make different
organizations and individuals accountable for their disaster management responsibilities
when events strike. As discussed in the case of ministries of finance, the best way to
ensure that a ministry invests sufficient present resources in mitigation projects and
prepares a plan for obtaining reconstruction financing is to hold it accountable for the
reconstruction process. Legislation is the obvious choice for officially establishing ac-
countability.

Financial Sustainability

Although the key to the financial sustainability of a national system is to ensure the
continued political impetus behind the system, other methods include committing to
long-term financing contracts with external parties and responding to pressure from the
international finance community.

31
Long-term Contracts

Long-term contracts with reinsurance companies, investors, or international financial

institutions are one way to ensure continued financial support for the system. In addi-
tion, long-term obligations encourage these parties to invest more fully in measures
such as risk identification, which will lead, in the long-run, to lower transaction costs.

International Aid and Financial Institutions

The actions of the international donor community can play a decisive role in the sus-
tainability of national programs. For the poorest countries, the assistance of the inter-
national community is critical in their ability to deal with risk. Policies that can make or
break the establishment of national programs include funding by the international
community for post-disaster reconstruction projects that do not address the exposure of
reconstructed structures to hazards (including issues of citing and proper building stan-
dards); ignoring hazard risk as a component of developing country assistance strate-
gies; bypassing existing institutional structures in the provision of post-disaster assis-
tance; and providing post-disaster reconstruction funding without holding national gov-
ernments responsible for some portion of future risk. Much of the current problematic
policy related to natural disaster planning is a result of the types of programs the inter-
national donor community was willing to fund in the past.

There is considerable policy justification for the international finance community to

place this issue on the table. The multilateral financial institutions have long recognized
that they are the insurers for poorer countries’ losses from natural disasters. As the
costs of disasters continue to escalate, the demands on the multilateral community for
post-disaster assistance have dramatically increased. For example, in the past four
years, the IDB’s average annual lending for post-disaster assistance increased by a fac-
tor of 10 compared with the previous 15 years. This experience parallels the experience
of the World Bank. As noted in IDB (2000b), while the global risk from disasters is in-
creasing, the overall level of assistance available for emergencies has been shrinking
since 1992.

Conclusion

This chapter has analyzed the key elements of a comprehensive risk management sys-
tem. Increasingly, risk management professionals are recognizing that reducing vulner-
ability to disasters involves a wide range of policy initiatives that engage broad seg-
ments of society. The focus of this report has been on the experiences of the interna-
tional community in forging links at a national level to develop comprehensive national
systems.

Two broad approaches emerge from the literature: a centralized, government-directed

risk management system, and a more decentralized, locally directed approach. The cen-
tralized strategy relies on national legislation that creates an organizational structure to
integrate existing government and nongovernmental institutions into the policy process.
The most successful programs include reconstruction financing strategies, many of
which use insurance to protect against large losses and encourage the adoption of miti-
gation measures. In practice, the more centrally directed programs rely heavily on
structural engineering solutions to risk management and have difficulty implementing
nonstructural mitigation measures and finding and maintaining an appropriate role for
local institutions in the risk management process.

32
Decentralized models, on the other hand, rely on national governments and NGOs to
provide guidance and support for local initiatives. These locally directed programs have
proven effective in implementing nonstructural mitigation measures, but lack the com-
prehensive approach possible with centrally directed programs.

The best policy outcome is a national system that embodies a measured mix of the two
approaches. A comprehensive approach requires the commitment of the national gov-
ernment, and that commitment needs the attention of those directing development
policies. In addition, the role of risk transfer as a decentralized market initiative is im-
portant and, in the case of high exposure, will rely on a supporting role by the national
government. The task is to create an effective national system with comprehensive vi-
sion that engages senior government policymakers, accommodates and supports local
decision-making and initiatives, engages civil society, and promotes the institutional
conditions necessary for the constructive involvement of private market initiatives.

33
34
 Chapter 2

Financial Strategies for Natural Disaster

Reconstruction

Risk transfer is among the least understood components of risk management in devel-
oping countries. Three reasons dictate the need for countries to consider risk transfer as
a vital component of their risk management strategy. First, if disasters are not antici-
pated and planned for, the diversion of scarce financial resources to relief and recon-
struction efforts causes high opportunity costs as other projects contributing to eco-
nomic growth and the eradication of poverty cannot continue as planned. Second, the
continuing and significant reallocation of post-disaster resources disrupts the budgetary
planning process. The annual budget process is often a complicated and politically diffi-
cult one. Shifting resources in response to disasters upsets fragile compromises formed
in the initial budget plan. For many countries, this shift creates considerable institu-
tional friction (Lewis and Murdock, 1999). Third, poorer countries rely on international
assistance to pay for a substantial portion of their losses. The resources available to the
international development community are limited and have remained stagnant for
nearly 10 years (OECD, 2001).

At the same time, countries should continue their practice of financing post-disaster ex-
penses with traditional financing instruments. The mix of policy options needs to include
access to the least cost financing alternatives. To the extent that losses can be paid
from aid donations or low cost external borrowing, those resources should be tapped
first. As detailed in the following discussion, the tradeoff between access to adequate
resources and the cost of those resources should become a component of proper natu-
ral disaster planning. This chapter explores tools to illuminate the planning alternatives.

As the cost of disasters increases, the demand on the international financial community
to provide needed resources has also increased. For example, the Inter-American De-
velopment Bank has increased its average annual disaster-related spending by a factor
of 10 in the past five years in comparison to the previous 15 years (IDB, 2000b). In
consuming the limited funding available, natural disasters divert resources needed to
support longer-term economic and social development objectives. The Organization of
American States notes, “Funds intended for development are diverted into costly relief
efforts. These indirect but profound economic effects and their drain on the limited
funds now available for new investment compound the tragedy of a disaster in a devel-
oping country.” (Bender, 1991).

To evaluate risk transfer alternatives for financing post-disaster reconstruction, it is

necessary to understand the risk of countries to disaster loss as well as their current
practice to cope with that loss. Once exposure is understood and existing policy is high-
lighted, alternatives can be proposed. The problem is addressed here in three stages:
estimation of the impact of natural hazard events on the economy of each of four se-
lected countries (Bolivia, Colombia, the Dominican Republic, and El Salvador); review of
existing government strategies and practices for financing disaster loss; and a compari-
son of new policy alternatives for financing disaster loss.

35
The specific models developed herein for securing post-disaster reconstruction financing
are intended to show broad-based tradeoffs between policy options. Core pieces of in-
formation required to make these models effective as policy tools do not exist. Rather,
estimates based on known data were used to support key assumptions. The intent is to
demonstrate a method to show how tradeoffs in policies could be understood, not to
demonstrate the actual tradeoffs for each country.

Lessons Learned

A comparative study was undertaken in four Latin American and Caribbean countries
that had recently experienced major natural disasters: Bolivia, Colombia, El Salvador,
and the Dominican Republic. The four selected countries represent a cross section of
risk profiles. The main contrast among the four countries is the covariant nature of the
risk they face, that is, the likelihood that a natural hazard event would impact the entire
country. Both the Dominican Republic and El Salvador face risks that may destroy sub-
stantial portions of the country. In Colombia and Bolivia, the risks are localized; if dis-
asters occur, it is unlikely that there will be damage to more than one defined region.

The hazards faced by the countries represented are varied and include such natural
phenomena as hurricanes, earthquakes, floods, volcanic eruptions, landslides, droughts,
and tsunamis. The focus of the study is to understand the current and prospective
mechanisms used by these countries to finance substantial reconstruction after major
natural disasters. In each country, meetings were held with relevant government and
nongovernment agencies to understand their role in post-disaster reconstruction and
the sources of funding they use to support their activities. All of the countries rely on
funding from international financial institutions to support their post-disaster recon-
struction.

An important lesson for multilateral financial institutions and three broad policy lessons
for the countries emerged from the country visits. The former has to do with the com-
plex process of providing post-disaster funds and how it is reflected in the activities of
the countries. The three lessons applicable to the countries are the failure to account for
natural disaster risk in the national planning process, the inadvertent assumption of risk
by many governments, and the inefficient purchase of insurance in each country. This
section explores these themes in more detail.

Country Policy and Funding from the International Aid Community

A surprising development from the study concerns the way in which financing strategies
of the countries are driven by expectations of post-disaster aid from the international
finance community. In each country (with the possible exception of Colombia), govern-
ment officials expect that the multilateral finance community will provide post-disaster
financing for reconstruction. Much of the countries’ pre-disaster behavior is directed at
maximizing post-disaster aid.

A good example of this policy is the recent activity in Bolivia, which has passed new leg-
islation to revamp its national disaster system. Two changes in this legislation are di-
rectly related to guarantying access to international aid after a disaster. The first is the
establishment of a national disaster fund with the primary purpose, according to both
the housing and sustainable development ministries, being to ensure that Bolivia has
sufficient funds to meet its co-pay obligations that arise with disaster aid. In this way,
the country can maximize the amount of available post-disaster assistance. The second
change is the movement of the natural disaster agency from being under the direction

36
of civil defense, which is considered a component of the military. After the earthquakes
in 1998, the Bolivian government was denied some assistance from a number of Euro-
pean governments because the institution directing reconstruction was a military entity.
Reconstruction responsibility was reassigned to the sustainable development ministry.
In this way, it is hoped that additional post-disaster aid will be forthcoming.

In each country, officials recognized that the primary source of reconstruction funding
from international financial institutions is the diversion of existing loans. This has sig-
nificant advantage for the countries, since the administrative burden of diverting loans
is considerably less than the effort required for new credits. The immediacy of the loans
requires that compromises be made that would not be acceptable if the emergency was
not already on the country.

Two policy implications result from this limited focus of country officials on international
financial institutions for post-disaster financing. First, there is an institutional commit-
ment only to maximizing the post-disaster loan diversion capacity to the exclusion of
other ex ante policy initiatives, like insurance. Second, the willingness of the interna-
tional financial institutions to permit loan conversions significantly reduces their ability
to impose loss reduction measures as a condition of lending. In the moment of crisis, it
is difficult to impose additional conditions on already approved loans. As a result, the
usual bank conditions on reconstruction loans are not required for diverted loans. It is
not surprising that countries attempt to maximize their access to these types of credits.

There is a need for the international finance community to better understand the expec-
tation created around their willingness to divert existing loans to finance disaster recon-
struction. While the policy of many governments is being directed to access unfunded
loans in times of disaster, the international institutions have not clearly established the
conditions for loan conversions. As a result, the opportunity to advance policy objec-
tives, like prevention and mitigation, is lost in the loan diversion process.

Disaster Risk and National Planning

Despite the considerable research regarding the long-term impacts of disasters on sus-
tainable development, none of the visited countries formally incorporates natural disas-
ter risk in its national planning process. While all the countries budget for disaster relief
through civil defense, and have some planning activity related to risk mitigation, none
accounts for probabilistic losses from natural phenomena as an ongoing component of
the budgeting process. Without accounting for potential contingent losses, the countries
lack the necessary information to consider and evaluate alternatives to reduce or fi-
nance probabilistic losses. As a result, policies directed at reducing risk through a cul-
ture of prevention lack the attention they require, as Mitchell (1999) notes.

Safety (a prime consideration in hazards management) does not necessarily equal sus-
tainability, and contingencies (of which hazards and disasters are good examples) may
require different responses than enduring problems. The truth is that large and complex
cities require expansive management initiatives that can simultaneously address inc-
ommensurable goals. Mega-cities must be prepared to cope with unexpected or unfa-
miliar events as well as long-term problems; acute natural hazards as well as chronic
crises of environmental degradation. To assume that sustainable urban development
can be achieved without attention to problems of contingency—of which natural hazards
are a pre-eminent example—is to court frustration and failure.

37
The need to plan for contingencies is as true for countries as it is for mega-cities. De-
spite this apparent need, understanding and planning for the contingent nature of natu-
ral disaster risk is not a policy objective of any of the countries visited.

The failure to understand the risk from natural hazard events has wide ranging implica-
tions. The most obvious is that lack of understanding of contingent exposure to natural
hazards limits the ability of a country to evaluate the desirability of financial planning
tools to cope with risk. These tools, of which insurance is the best known, require that
risk be reasonably quantified as a precondition to use. While it is possible to make pol-
icy decisions without probabilistic estimates, the failure to quantify risk when it is possi-
ble constrains the decision-making process. The management of financial losses always
involves tradeoffs with respect to anticipated future consequences. Lacking any predic-
tive knowledge of potential future outcomes reduces the ability to evaluate alternative
options to finance risk. The governments of the four countries studied herein are cur-
rently incapable of evaluating such policy options.

The lack of probabilistic estimates for natural hazards has other serious implications.
Since the cost of reconstruction is not planned for in advance, the primary incentive to
promote mitigation and risk prevention is lost. Mitigation and risk prevention require
funding. Their purpose is to reduce future damages from expected events by making
efficient current expenditures to reduce risk. If the future damages are not considered
as a component of the current planning process, it is nearly impossible to sustain budg-
etary resources to reduce those potential damages.

This concern addresses the financial issue of the sustainability of national systems,
which depends on the participation of those responsible for national budgeting and
planning in the national system. A prerequisite for the involvement of a national plan-
ning office must be the recognition of risk from natural phenomena. The planning obli-
gation must be to understand how the risk will be handled prospectively. The absence
of comprehensive understanding of risk explains why efforts to advance risk mitigation
and prevention receive little national attention. It also helps explain the frustration in
maintaining sustainable national systems.

The only agency found in the four selected countries that considers the probabilistic
losses from natural hazard events in its budgeting process was the agency responsible
for national highways in Colombia. The agency is responsible for highway maintenance
and reconstruction after a disaster. For each segment of the highway, it has prepared a
probabilistic estimate of future losses from natural hazard events. These estimates,
created with the help of the University of the Andes, are used to set the agency’s an-
nual maintenance and reconstruction budget. For years in which the natural hazard
events are less than estimated, budgeted funds are used to reduce vulnerability
through increased maintenance. For this agency, the link between mitigation and risk
reduction is clearly defined. The agency understands its contingent exposure to loss and
can therefore plan its resources to manage that risk.

The tools to develop probabilistic estimates of future losses are available in all the stud-
ied countries. Why these tools are not being used to quantify risk and develop planning
options is unknown.

Assumption of Risk by Governments

In the post-disaster period, decisions were made about reconstruction based on

damage studies. Little attention was focused on whether the assets damaged were the

38
responsibility of the government. Rather, the decision process focused on damages,
potential government resources for reconstruction, and how best to spend those
resources in the event of an emergency. In both Colombia and Bolivia, after the most
recent earthquake disasters, the governments rebuilt homes destroyed or damaged by
the earthquake. The governments in both countries claimed that rebuilding housing had
not been a government obligation in past disasters. An expectation has been created
that the government is responsible for housing reconstruction.

In Bolivia, the government has met that expectation by rebuilding housing after recent
floods and fires. Colombia spent US$800 million to rebuild in Armenia and Pereira after
the 1999 earthquakes. An earthquake that impacted Bogota would incur considerably
higher damages. The policy issue is not whether the government should be obligated to
rebuild after a disaster, but whether obligations of this magnitude should be informally
incurred as a result of the post-disaster financing process now employed. Countries un-
der severe fiscal constraints should incur substantial future contingent obligations only
as part of a long-term planning process, not in response to unplanned disaster borrow-
ing.

Efficiency of Existing Insurance Purchases

The final general observation relates to existing practice with respect to government
purchase of insurance. In every country examined for this study, a requirement existed
for government agencies to purchase insurance to protect their buildings and their con-
tents. In each instance, the level of insurance, the premium paid, and the company
from whom the insurance was purchased was left to the discretion of the appropriate
minister or agency (whether at the national, provincial, or municipal level). In none of
the countries did anyone know the aggregate amount of insurance purchased, the pre-
miums paid, or the level of protection provided. What is clear is that the existing proc-
ess of acquiring insurance is inefficient.

The use of consolidated insurance purchased by the public sector in these countries
would likely substantially increase the amount of insurance protection available at no
additional budgetary cost. By matching the purchase of insurance with the risk to gov-
ernment as a whole, as opposed to the risk to a particular building or agency, the in-
surance purchase could be integrated into a comprehensive risk management process.
How the government handles risk in the aggregate is much different than how a single
agency manages risk. The diversification of risk at the government level, particularly for
idiosyncratic risk, would probably indicate that the purchase of insurance should be
done for the highest, but least expensive levels of risk. The government itself can gen-
erally absorb the lower levels of risk more efficiently using its own resources. A study
examining the management of disaster risk in Mexico finds a similar situation. Mexico
required that all its government agencies purchase insurance. In reviewing the pur-
chases of the different agencies, the conclusion was reached that consolidating the pur-
chase of public sector insurance combined with the employment of international brokers
should increase the efficiency of the risk transfer for public sector risks (Kreimer, Ar-
nold, and Freeman, 1999).

Measuring the Resource Gap

The resource gap is a measurement of the inability of a country to finance its recon-
struction obligations after a disaster. The calculation of the resource gap requires three
computations. First, the risk of the country to natural hazard events must be calculated.
Risk is a function of the hazard (or probability of phenomena of different magnitudes

39
impacting a country) and the vulnerability (or susceptibility of the exposed population
and assets to loss). The second calculation is the financial responsibility of the govern-
ment to finance country losses. Primary losses from natural hazard events may be the
responsibility of various parties in addition to the governmentindustry, businesses,
homeowners, and individualsbut the concern here is with government responsibility.
Third, the capacity of the government to meet its financial obligations must be calcu-
lated. To the extent that the government lacks the resources to fund its obligations, it
has a natural disaster resource gap. The required resources may come from interna-
tional aid, government revenues (taxes), reserves, insurance proceeds, borrowing, or
diversion of resources from other programs. All of these alternatives have an associated
cost and limitations on availability.

Hazard

Hazard is the probability of occurrence of natural events that can cause significant eco-
nomic damage. For example, floods, hurricanes, and earthquakes are responsible for 90
percent of the economic costs from natural hazards worldwide. Countries tend to be im-
pacted by the same types of natural hazard events. Earthquakes usually occur in well-
defined seismic zones; windstorms usually travel along identified hurricane paths; and
floods usually occur in river and coastal areas. The concern about hazards relates to the
intensity and timing of their occurrence. Different countries are impacted by different
phenomena. It is not uncommon for some countries to be exposed to more than one
natural hazard.

Bolivia is primarily exposed to small, recurring floods and landslides scattered across
the country. Bolivia, along with Colombia and El Salvador, is particularly concerned
about floods and drought due to the El Niño phenomenon. Seismic hazard is also an is-
sue, primarily in Bolivia’s central region, but also potentially in La Paz. The two most
recent events in Bolivia are the February 2002 floods in La Paz and the May 1999
earthquakes near Tarija that measured 6.3 on the Richter scale. Bolivia is relatively for-
tunate that it is mainly exposed to low hazards in limited geographic regions in the
country.

The geographic diversification of the disaster risk in Bolivia is mirrored in Colombia. In

contrast to Bolivia, however, natural hazard events in Colombia can be very severe. In
1906, for example, an earthquake of magnitude 8.9 Mw hit off the Pacific Coast of Co-
lombia and was felt in the entire country and surroundings. Colombia is affected primar-
ily by earthquakes and tsunamis. It also is exposed to floods and volcanoes with conse-
quent mud floods. The most important recent event in Colombia was the 1999 earth-
quake in Quindio, known as the “Eje Cafetero” disaster because the shocks affected a
key coffee producing area.

In contrast to the geographically constrained events that affect Bolivia and Colombia,
the Dominican Republic and El Salvador are susceptible to being affected in their en-
tirety by individual catastrophic events. The Dominican Republic has hurricane and
earthquake hazards. Hurricanes occur frequently; earthquakes occur rarely but with
larger associated potential losses. The last large earthquake in the Dominican Republic
occurred in 1946 with a magnitude of 8.1 on the Richter scale. The most important re-
cent event was Hurricane Georges in 1998, which ranked 3 out of 5 on the Saffir Simp-
son hurricane scale. Hurricane David in 1979 ranked a full 5.

The most important hazard events in El Salvador are earthquakes and volcanic erup-
tions, with resulting landslides likely and tsunamis possible. Earthquakes have struck

40
San Salvador 13 times over the past 400 years, almost destroying the city in 1854,
1873, 1917, and 1986. The largest recent events in El Salvador were the earthquakes
of January and February 2001, with Richter magnitudes of 7.6 and 6.6, respectively.
Figure 2.1 shows the seismic hazard for all of Central America.

Vulnerability

Figure 2.1. Seismic Hazard in Central America

Swiss

Source: Freeman and others (2000).

The entire Latin American and Caribbean region is experiencing increased vulnerability
to natural hazard events (see figure 2.2). Worse yet, the impacts of natural hazard
events are estimated to increase dramatically over the next 50 years. By some esti-
mates, the global cost of natural disasters is anticipated to top US$300 billion annually
by the year 2050 (Munich Reinsurance Company, 2001). The primary factor influencing

Figure 2.2. Increasing Number of Events Qualifying as Natural

Disasters in Latin America and the Caribbean, 1900-2000.

80

70

60

50

40

30

20

10

0
1900 1904 1908 1912 1916 1920 1924 1928 1932 1936 1940 1944 1948 1952 1956 1960 1964 1968 1972 1976 1980 1984 1988 1992 1996 2000

Source: Charvériat (2000).

41
the estimated increase in damages is the increasing concentration of wealth and popu-
lation in hazard-prone regions of the world, primarily urban areas. In Latin America and
the Caribbean, more than 75 percent of the population is projected to live in urban ar-
eas by 2030 (UNFPA, 1999). Mega-cities are highly vulnerable to natural hazard events.
Nearly half of the world’s largest cities are situated along major earthquake zones or
tropical cyclone tracks (Bendimerad and others, 1999).

The social and economic effects from natural phenomena are a consequence of the so-
cial fragility of low-income countries. A disaster occurs when vulnerable people are im-
pacted by a hazard event and suffer severe damage and/or disruption of their livelihood
system in such a way that recovery is unlikely without external aid (Blaike and others,
1994). Vulnerability is a function of a group’s socioeconomic condition; thus, the poor
are more vulnerable than the rich.

The Dominican Republic is one of the countries in Latin America where vulnerability to
natural phenomena has increased most over the past four decades. According to a
methodology developed at the University of Colorado, vulnerability in the Dominican
Republic has increased more than 30 times in that period. This measure of vulnerability
represents both increasing physical vulnerability of structures and the accumulation of
assets over the period. The same analysis estimates that if Hurricane David (1979)
were to hit again, it would cause not US$150 million in damages, but US$4 billion in
damages (Pielke and others, 2001). Figure 2.3 shows the hurricane hazard in the Do-
minican Republic, including wind speeds associated with a 50-year event, and the pos-
sible damage such an event could cause to the capital, Santo Domingo.
Figure 2.3. Hurricane Hazard in the Dominican Republic

72 20.17 N CDMP

68
W

Ross Wagenseil
for CDMP
January 2000

17.5 N

Wind Speeds
Storm Category 0 1 2 3 4 5 Damage (percentage of total value of structures)
Source of Wind speed
knots 25 50 75 100 125
damage 35 m/s 45 m/s 55 m/s
mph 25 50 75 100 125 150
Wind 1 3 8
kph 50 100 150 200 250 Water 3 6 10
m/s 10 20 30 40 50 60 70

Source: OAS, 1999a. Caribbean Disaster Mitigation Project, implemented by the OAS Unit of Sustainable De-
velopment and Environment for the USAID Office of Foreign Disaster Assistance and the Caribbean Regional
Program.

42
The vulnerability of housing is of particular concern in the four case study countries. Al-
though sources estimate that the vulnerability of structures in San Salvador has de-
creased since the 1986 earthquake, the four countries still have high levels of informal
construction, a small fraction of which is estimated to adhere to disaster-resistant build-
ing codes. Even in the country with the most developed disaster system, Colombia,
building codes formally did not exist until 1984. In regions where property rights are ill
defined, risk of disputes has been a deterrent to investing in disaster-resistant struc-
tures.

In Bolivia, vulnerability of crops and crop revenues is also an important issue for small
farmers who are subject to relatively frequent floods and landslides. The largest impact
on the productive sectors in Bolivia comes from the potential effects of El Niño-related
floods and droughts on agricultural exports. Table 2.1 shows assumptions of projected
losses used in the modeling exercise for the four case countries. It is expected that in
Bolivia, there would be direct losses to capital stock of at least US$200 million approxi-
mately every 20 years; more specifically, there is a 1-in-20 chance every year to have
a catastrophic event equaling or exceeding US$200 million in losses. Likewise, there is
a 1-in-50 chance, or 0.02 percent probability, every year of at least US$600 million in
direct losses. Including indirect losses in production, tourism, or other services could
double the magnitude of these figures.

Table 2.1. Projected Economic Loss Caused and Government Financing Needs in
Response to a Disaster in the Case Study Countries

Country 20-year event 50-year event 100-year event

Projected loss
Bolivia 200 600 1,000
Colombia 2,000 5,000 8,000
Dominican Republic 1,250 3,000 6,000
El Salvador 900 3,000 4,500

Financing needs
Bolivia 100 300 500
Colombia 1,000 2,500 4,000
Dominican Republic 625 1,500 3,000
El Salvador 450 1,500 2,250
Note: Figures are in millions of U.S. dollars. The projected losses are assumptions
made on the basis of available past information for the four case countries. Govern-
ment liability and financing needs that correspond to half of the losses are assumed on
the basis of the information available for El Salvador.

In the Dominican Republic, the vulnerability of the tourism sector is critical. Although
most hotels and resorts are insured against hurricanes, the indirect impacts of a hurri-
cane on the flow of tourists were identified as a serious concern. Tourism receipts cur-
rently exceed US$2 billion a year, contributing to more than 12 percent of gross domes-
tic product (GDP).

El Salvador is most concerned about the potential impact of earthquakes on housing

and the transportation sector. The quality of construction in San Salvador is estimated
to have improved since the 1986 earthquake, but the construction in the rest of the
country is said to be highly vulnerable to earthquakes. As for the transport sector, the
Pan American highway crosses the length of the country, San Salvador’s airport is one

43
of the largest in the region, and a new port is under construction that is planned to be
the largest port on the Pacific Coast for the region. If any of these were affected, the
entire region would feel the impacts of disrupted transport.
For purposes of calculating the resource gap, certain assumptions were made about the
vulnerability of each case study country. (Note that these figures are ballpark esti-
mates.) Combining the hazard and vulnerability estimates produces the estimates of
disaster loss in the four countries given in table 2.1.

Risks Governments Assume

Once probabilistic estimates of loss are determined, it is essential to know the responsi-
bility of the government for those losses. Two broad categories of government respon-
sibility can be generally defined: risk arising from national government ownership of
assets, and risk government assumes for others. In the first category is the probability
of loss to government buildings, including schools and hospitals, and infrastructure, like
roads, bridges, and airports. The second category focuses attention on risk that the
government assumes for others. This generally includes the risk to homeowners, agri-
culture, local and provincial governments, and the poor. There are different policy im-
plications in how the national government copes with risk from its ownership of assets
and the role it plays in coping with risk that is created from decisions made by others.
The policy issues associated with the latter risk are influenced by moral hazard: the
likelihood that behavior is changed as a result of the government’s assumption of risk.

Infrastructure

In all countries, the government is responsible for the reconstruction of real property
that it owns after a natural disaster. The rebuilding of roads, bridges, hospitals, schools,
airports, and office buildings must be borne by the national government. Three main
issues distinguish the level of this risk between different countries: the extent to which
the damaged assets are insured; the level of privatization of formerly government-
owned assets; and the ownership of assets by municipal and provincial governments.

In all four countries surveyed, there is a mandatory requirement to purchase insurance

for government-owned office buildings. In none of the countries was it possible to de-
termine the level of insurance actually purchased, the premiums paid, or the expected
reimbursement in times of disaster. As a general rule, the governments did not expect
that insurance proceeds would provide significant financial resources after a disaster. In
fact, no one agency in any of the countries knew what assets were actually insured. The
insurance purchase decision was left to each ministry.

There are anecdotal references to insurance proceeds making a difference for the re-
construction of specific buildings. In El Salvador, the reconstruction of the main ministry
offices is supported by insurance proceeds. There is discussion of extending the pur-
chase of insurance for key infrastructure, primarily bridges and roads. The view is that
significant benefit would have occurred if hospitals and schools were insured, but they
were not. Similarly, a proposal to insure all roads and bridges has been made for Co-
lombia, but was rejected as too costly. All national government office buildings in Bo-
gota are insured, but the level of coverage for each building is left to the discretion of
the relevant ministries. Bolivia requires all government office buildings be insured, but
whether this requirement is complied with is unknown. In the Dominican Republic, it is
also understood that all government buildings should be insured, but it is unlikely that it
actually happens.

44
In most of the developing world, the government owns 95 percent of public infrastruc-
ture. One byproduct of privatization is the potential shifting of risk of post-disaster re-
construction from the government to private entities. In El Salvador, the privatization of
major utilities made a large difference in reconstruction after the recent earthquakes.
The privatized companies in electricity transmission and generation were responsible for
their own reconstruction after the disaster. An additional component of the privatization
process in El Salvador was that funds received from the privatization of government as-
sets had been held in a reserve account of US$500 million. This account was tapped in
the immediate post-disaster period to fund recovery efforts in the country. In the Do-
minican Republic, the main port is now privatized, and any potential losses should not
be borne by the government.

In Bolivia and Colombia, the nature of privatization is more complicated. In these coun-
tries, privatization takes the form of granting concessions to private parties to operate
assets still owned by the government. It is not clear who bears the cost of risk to those
assets, but it seems that the risk remains with the government.

In all of the countries visited, there is a significant effort at decentralization. Responsi-

bility for government functions is being shifted from the national government to re-
gional and municipal governments. With the shifting of responsibility is also a move-
ment of financial resources to those government entities. As a general rule, the national
governments believe that the local government units should bear responsibility for the
reconstruction of their assets after a disaster. In practice, the national government con-
tinues to assume some responsibility for local repairs and reconstruction. This was less
the case in Colombia than the other countries. Bolivia is now trying to formalize respon-
sibility at the local level for both maintenance and repair of key locally owned infra-
structure, with a reduction in payments to the local government if it fails to meet its re-
sponsibilities. This process accompanies a detailed arrangement on revenue sharing be-
tween the central government and the units of local government. In neither El Salvador
nor the Dominican Republic were the lines of national and local responsibility clearly
drawn. It should be noted that all four countries have proposed or revised laws on natu-
ral disaster planning and risk management that may formalize the obligations between
different units of government on post-disaster reconstruction.

As a general rule, the national government in all these countries bears the risk of re-
construction of government-owned assets. There have been efforts to reduce the risk
borne by the national government through insurance, privatization, and devolution to
local government units. While anecdotal evidence exists as to the success in particular
cases of risk absorbed by these processes, no systematic understanding of the risk re-
duction consequences of these activities was determined. A major shortcoming in each
country was the failure to quantify the natural disaster risk on a probabilistic basis.
Without an appreciation of the risk at a national planning level, the effectiveness of
tools to reduce that risk is not evaluated.

Agriculture

The responsibility of government for agricultural losses from natural disasters is

perceived as being important for policymakers. Fluctuations in income, whether from
catastrophic loss or other causes, create significant welfare issues for farmers and have
important spillover effects for other rural households and businesses. As noted in
Varangis, Skees, and Barnett (2002), “destroyed crops and livestock reduce
employment opportunities, with serious implications for the landless rural poor in
developing countries.” Addressing farm risk through government intervention has a

45
long and complicated history. Across many developing countries, including those in
Latin America, new initiatives are constantly being explored. In Mexico, the first use of
government purchase of weather derivatives to absorb some risk to small farmers has
been completed. Recently, the World Bank and the IDB announced a new program to
help stabilize the incomes of coffee growers in Central America.

Of interest to this analysis is the drain on government resources to support the agricul-
tural sector post-disaster. In this regard, there were different approaches in all the
countries studied. In the Dominican Republic, government support of agriculture is done
through a government sponsored agriculture bank that provides needed credit to farm-
ers. After disasters, the government generally forgives those credits. There have been
proposals for crop insurance that requires some government subsidy, but the proposals
have not been seriously considered.

In El Salvador, the issue of loss of agricultural production from earthquakes is less criti-
cal than the collapse in coffee prices. In 2000, the coffee crop went unpicked because of
poor prices for coffee. The 2001 earthquakes did nothing to improve the price of coffee,
and it is likely that the crop for that year, even if it were not damaged, would have re-
mained unpicked.

Bolivia does not have a government-supported agricultural bank. The export farmers in
Bolivia are large, well capitalized enterprises. Other than the political pressure from the
government to ask the private banks to reconfigure farmer credits after a disaster, the
government does not seem to have a direct role assisting farmers. Bolivia is now dis-
cussing potential crop insurance programs.

Colombia is similar to Bolivia and El Salvador. A main export crop is coffee, which is ex-
periencing severe pricing issues regardless of disaster. There is no formal program to
help farmers after a disaster. Crop insurance exists in Colombia, particularly for ba-
nanas.

In all these countries, the main government concern is the spillover effect of lost em-
ployment for the poor after a disaster. This topic is discussed in the section on how the
government copes with the needs of the poor.

Housing

The most interesting outcome of this study was how governments dealt with the loss of
housing after a disaster. In none of the four countries is there a legally mandated obli-
gation to rebuild destroyed housing after a disaster. Yet, in each of the four countries
after their most recent disasters, the government undertook significant responsibility for
housing reconstruction. In all four of the countries, the cost of housing reconstruction
was among the most significant financial obligations for the government in its recon-
struction efforts.

The vexing issue is the level of government obligation to reconstruct private housing in
the event of future disasters. In Bolivia, Colombia and El Salvador, future disasters
could cause substantially greater housing losses than the last disasters. In all three
countries, major urban areas were not struck in the last incidents. If they were, housing
losses would have been staggering. The assumption of responsibility for housing
reconstruction in each case occurred after the disaster took place. Furthermore, the role
the government would play in reconstruction was newly developed in the wake of the

46
disaster. The international community approved funding for the government housing
reconstruction effort, and housing was rebuilt with borrowed funds.

The concern is the potential contingent liability of the government in future disasters for
housing reconstruction as well as the moral hazard if the population perceives that a
disaster is an opportunity to obtain new housing at government cost. This moral hazard
has already manifested itself in two examples. In the Dominican Republic, there is con-
cern that improvement of permanent housing by the poor is not undertaken because
the government will only replace destroyed homes. As a result, mitigation efforts that
may preserve a home after a disaster are not employed. It is better to have a new gov-
ernment home than to remain with a disaster-resistant existing house. In Bolivia,
homeowners exposed to forest fires were burning their houses in order to receive a
government rebuilt home.

The situation in these countries is reminiscent of the situation in Turkey prior to its dev-
astating earthquakes in 1999. Turkey had a constitutional obligation to rebuild de-
stroyed housing caused by natural causes. It was widely perceived that this obligation
by the government created a disincentive for homeowners to properly protect their
properties from potential loss. After the earthquake, the government changed the con-
stitution, created a mandatory insurance program for homeowners, limited the govern-
ment liability for reconstruction above the insured amount of each home, and created a
funding mechanism to meet its future obligations. In effect, the government defined,
limited, and funded its future obligation for housing reconstruction.

In all four countries, it is important that the responsibility for housing reconstruction be
defined. In each country, expectations are now being created about the role of the gov-
ernment in rebuilding housing. These expectations are likely to become significant li-
abilities in the face of large disasters. Since the obligations are not legally defined, gov-
ernment officials did not seem concerned about the potential losses. In truth, the risk to
the governments for housing damage may be the most significant risk they bear from
natural hazard events. As such, it should be defined, limited, and funded.

Local and Provincial Governments

The interrelationship between the national government and local and provincial gov-
ernments plays an important role in the risk assumed by the national government. In
each country, there is a major effort to allocate some portion of reconstruction to other
units of government. The requirement of the purchase of insurance also extends to
them, but their compliance with these provisions is even less well known than it is for
national government departments. At best, the role of the national government relative
to damages to municipal or provincial assets is not well understood. How the national
government will respond is a function of the size of the disaster, the political strength of
the local government, and party allegiances. In theory, the losses at the local level are
the responsibility of local governments in each country. This contrasts with the United
States, where the national government is obligated by law to fund 75 percent of the
losses in declared national emergencies of the state and local governments. In practice,
the decision seems to be made on a case-by-case basis. In any event, there is little
monitoring of the local risk at the national level.

The Poor

How each country handles the obligation of the government to provide income support
and assistance to the poor in the wake of a natural disaster is a function of its existing

47
safety net arrangements. In the post-disaster period, there is an urgency to
immediately respond to the needs of everyone. This response activity is handled by
each country’s civil defense agencies. The evaluation of that response capacity, which is
undergoing significant changes in some of these countries, is beyond the scope of this
analysis. The concern here is with the income support needed in the longer term after
an event.

After a disaster, one major concern is the rebuilding of income-producing activity in the
private sector. Depending on the size and nature of the disaster, it may take several
months or years to restore economic activity to its pre-disaster levels. For the poor, it is
mandatory that alternative sources of income be made available to them during this
interim period. The most common means to provide that income is to engage them in
paid work to support reconstruction; these are deemed public workfare programs.

In all four countries, efforts are made to provide work for the poor in the reconstruction
process. The most complicated problems are those that arise in situations like El Salva-
dor, where broader economic issues are blended with natural disasters. The destruction
of coffee plantations by earthquakes makes little difference in markets where the coffee
is not picked anyway. In this case, the obligation of the government to provide income
support is broader than waiting for production to return.

Government Liability

Absent the availability of better data on the magnitude of government responsibilities, it

has been decided, based on the experience in El Salvador after the 2001 earthquakes
and the Dominican Republic after Hurricane Georges, to assume that governments are
responsible for funding approximately half of direct losses.6 Estimated government li-
abilities facing disaster risk are shown in table 2.1. These estimates primarily represent
government liabilities for infrastructure reconstruction, but also include contributions to
the reconstruction of private housing and the provision of programs for the poor.

Table 2.2. Expected Average Annual Loss Relative to Economic Indicators in the Case
Study Countries (Percent unless otherwise noted)
Premium/
Pure premium Premium/ government Premium/IDB Premium
Country (million US$) GDP expenditures annual lending /ODA
Bolivia 10 0.12 0.4 11 2
Colombia 85 0.08 0.6 16 --7
Dominican Republic 54 0.34 1.8 54 23
El Salvador 48 0.40 2.2 34 19
Note: The expected average annual loss is a proxy for the pure premium. The pure premium figures are
based on table 2.1 and assumed annual distribution of government liabilities covering 20 to 100 year
events. The economic indicators reflect statistical data available for the case countries.

Table 2.2 presents the calculation of the pure premium for potential insurance on
assumed government liabilities covering 20- to 100-year events. The pure premiums

6
ECLAC estimates that the 2001 earthquakes in El Salvador were unique in that the public sector bore a rela-
tively small fraction of the losses (34 percent of the total US$750 million). The Central Bank of the Dominican
Republic estimated that the country was responsible for US$625 million of the US$1,250 million in direct
losses suffered during Hurricane Georges.
7
Colombia receives much less Official Development Assistance (ODA) than the other case study countries:
US$41 million in 1998 relative to US$590 million received by Bolivia, US$240 million by the Dominican
Republic and US$252 million received by El Salvador. Note that IDB loans do not count towards ODA.

48
represent the annualized expected loss due to all events that occur less frequently than
once every 20 years, but at least as frequently as once every 100 years. The table
reads, for example, that Bolivia has an annualized expected loss of US$10 million. This
means that to cover its losses, Bolivia will need to put aside on average US$10 million a
year.

As expected, although the pure premium in absolute terms is largest in Colombia, its
impact relative to economic indicators is most significant in the Dominican Republic and
El Salvador. In particular, the cost of disasters relative to government expenditures is
three times as high in the Dominican Republic and almost four times as high in El Sal-
vador as in Colombia.

For example, this premium is large relative to annual IDB lending in each country. In
the case of the Dominican Republic and El Salvador, the premium is significant com-
pared with annual receipts of Official Development Assistance (ODA).

Resource Gaps in Case Study Countries

Some governments cope relatively well with the disaster risks they assume; others ex-
perience serious difficulties in their ability to raise funds for reconstruction. Many coun-
tries in this latter category, like Honduras and Nicaragua, are still recovering from
events that occurred several years ago because of their struggle to raise and disburse
sufficient funds for reconstruction projects.

A resource gap is a shortfall in funds available for reconstruction that a government

should expect to face following a natural disaster event. Being able to quickly access
sufficient funds for reconstruction after a disaster is critical to a country’s ability to re-
cover with minimal long-term consequences. Countries with severe unmet resource
gaps can fall significantly short of their development goals. Studies conducted by the
International Institute for Applied Systems Analysis demonstrate that in Honduras and
Nicaragua, resource gaps could stagnate future growth projections (Freeman and oth-
ers, 2001).

A resource gap is calculated by comparing a government’s need for reconstruction

funds (given the risks of natural disasters) and its anticipated access to internal and ex-
ternal funds. Resource gaps are expressed as probabilities because the need for funds is
defined as a function of the country’s probabilistic disaster losses. The supply of funds is
determined by such factors as international aid, the size and flexibility of national budg-
ets, the depth of the domestic credit market, expected insurance payments, and the
government’s ability to reallocate existing loans and place new loans in the international
markets. If an event with a given probability of occurrence creates a demand for recon-
struction funds that exceeds the supply, a resource gap and its probability of occurrence
is identified. The resource gap is based only on the potential availability of funds, and it
does not take into account the cost of raising those funds. In other words, it is based
entirely on the constraints a government may face in raising financial resources after a
disaster.

In analyzing potential resource gaps, interesting variations emerged among the four
countries. According to this analysis, Bolivia can anticipate no resource gap over the
range of 20, 50, and 100-year events. Bolivia is the poorest country in South America,
but the level of hazard and therefore of risk is so low that it should have sufficient re-
sources to respond, particularly thanks to its access to subsidized loans from the multi-
lateral financial institutions. Colombia has a high level of natural disaster risk, but per

49
capita incomes are high and the risk is geographically diverse, so the government could
expand its tax revenues in response to a catastrophic event. Depending on assumptions
governing how much Colombia can raise, it could potentially have a resource gap asso-
ciated with 1-in-100 year events. Alternatively, El Salvador and the Dominican Republic
can anticipate resource gaps given their disaster risk. Both countries have a high vul-
nerability to large-scale natural events and limited financial resources. For each coun-
try, there is at least a 1-in-100 chance that they will suffer an event that outstrips their
ability to raise post-disaster reconstruction funds.

Resource gap projections depend on a set of assumptions about the magnitude of risk
in each country and the availability of funds from different sources. The following exer-
cise demonstrates the importance of further research in measuring disaster risk in the
four case study countries, the capacity of the Dominican Republic and El Salvador to
take on additional external debt, the access of the Bolivian government to additional
domestic credit, and the magnitude of potential tax increases in Colombia after a major
catastrophe.

Figure 2.4 illustrates the resource gap calculated for the Dominican Republic. The verti-
cal axis represents the maximum amount of funds available to the Dominican Republic’s
national government for reconstruction and any shortfall in funding. The resource gap is
calculated at three separate points: after a 1-in-20 year event (like Hurricane Georges),
after a 1-in-50 year event (like Hurricane David), and after a 1-in-100 year event (such
as a Richter 7 earthquake affecting the northern coast). The 20- and 50-year events do
not generate a resource gap, but the 100-year event does. The probabilities associated
with the different events are associated with probabilities of different resource gaps,
translating to a greater than 98 percent chance of no catastrophes or only small events
that do not generate a resource gap, and a 1 percent chance of events creating a gap of
more than US$1,148 million.
Figure 2.4. Resource Gap in the Dominican Republic

Note: Values represent the estimated financing available for reconstructing and the potential gap, as indicated
under the assumptions of table 2.3.

50
The magnitude of the resource gaps is based on the discrepancy between the demand
for funds and the potential supply of funds after an event. In the case of a 20-year
event in the Dominican Republic, direct damages are estimated at US$1.25 billion and
government liabilities at half that amount. A 20-year event does not create a resource
gap for the government because a potential supply of more than US$1.5 billion in funds
offsets the demand for US$625 million. These funds come from several sources: inter-
national aid, insurance payments, budget reallocations, new taxes, increases in domes-
tic credit, and additional external commercial or IDB/World Bank credit. It is important
to note that although the funds may be available, they are not free. A 20-year event
would shift US$500 million in resources away from current projects and increase debt
levels by US$83 million.

As the magnitude of the event increases, the amount of aid the country can expect to
receive will increase, and there may be more willingness to reallocate from other inter-
nal and external sources. However, based on experience in the Dominican Republic, it is
unlikely that these additional funds will be sufficient to cover the additional costs. For
less frequent but much larger magnitude events, a resource gap appears.

Table 2.3 gives the details of estimates of the resource gap in the Dominican Republic.
The first two rows represent the demand for funds from the catastrophe exposure esti-
mate and the government’s liability. The following rows represent estimates on the ex-
tent of (or constraints on) internal and external sources of ex post funding. This section
describes the rationale and assumptions underlying the estimates of available ex post
funding.

Table 2.3. Calculation of the Resource Gap for the Dominican Republic
Item 20-year 50-year 100-year

Direct Damages 1250 3,000 6,000

Government responsibility 625 1,500 3,000
Aid 11 26 52
Insurance payments 31 75 150
Budget reallocation 500 500 500
New taxes 0 0 0
Domestic credit 150 150 150
External credit IDB/WB 100 200 200
External credit market 800 800 800

Shortfall None None 1,148

Shortfall without IDB/WB None None 1,348
Additional debt 83 899 1,150
Note: Figures are in millions of U.S. dollars. The estimates for funding sources are assump-
tions based on historical data from previous disasters events. The results are used in this
study to show how tradeoffs in policies could be understood, not to demonstrate the actual
tradeoffs for the country.

International Aid

Aid inflows from abroad after a disaster include private and public donations from pri-
vate institutions, government agencies, and intergovernmental agencies in the form of

51
relief, technical assistance, grants, commodities, and money (Albala-Betrand, 1993).
The amount of aid appears to depend on the nature and extent of the event, which in-
fluences the will of the donors to grant assistance. There is considerable uncertainty as
regards the amount of aid available after a disaster.

There is little research on estimating how much aid a country can expect to receive af-
ter a natural disaster. Given the importance of aid in the recovery process, this lack of
research is surprising. To estimate expected aid inflows, data were collected on histori-
cal events since 1960 resulting in economic losses of more than US$50 million and the

Figure 2.5. International Aid as a Function of Direct Disaster

Note: The regression sample is disasters since 1960 resulting in economic losses of more than
US$50 million and the corresponding amounts of aid received for 16 countries in Latin America.

corresponding amounts of aid received. Sixteen Latin American countries were included
in the database.8 Figure 2.5 shows the results of a regression analysis to ascertain any
statistical relationship between economic losses and aid inflows, as measured as a per-
centage of GDP.
The regression indicates that on average 8.6 percent of direct disaster losses can be
expected to be covered by international assistance, with the exception of one outlier,
with a range of events from as little as 6 percent, to one event that received as much
as 25 percent.9

It is important to note that most aid received after a natural catastrophe is aid in kind,
that is, food, blankets, shelters, and the use of trucks and helicopters for emergency
response. It is estimated that on average only 5-10 percent of all international dona-
tions received come in the form of cash. In El Salvador, for example, after the recent
earthquakes, the country received US$500 million in donations from abroad, US$25

8
Loss data come from: Central American Natural Disasters Preparedness Center (CEPREDENAC); Munich Re;
and the OAS. Assistance data were obtained through the OAS. For events during the 1960s and 1970s, the
data source is Albala-Betrand (1993). For the Dominican Republic in 1998: IMF, IDB, European Union, Inter-
national Bank for Reconstruction and Development, United Nations, and the U.S. Agency for International
Development. For El Salvador in 2001: IMF and World Bank staff assessment for bonds; ECLAC-Mexico for
loans and donations.
9
The coefficient of correlation for this regression was at 0.55 after eliminating outliers.

52
million of which was in cash. Colombia and Bolivia reported similar experiences. There-
fore, the analysis assumes that only 10 percent of the total aid received comes in the
form of cash that contributes to relieving the government’s need for reconstruction
funds. For the purpose of calculating the resource gap for the case study countries, the
estimated aid received in cash is given in table 2.4.

Insurance Payments

Generally speaking, insurance plays a minor role in providing reconstruction funding in

the countries after a disaster. Only 3 percent of the losses from disasters in Latin Amer-
ica are covered by insurance, and this coverage is generally for large industrial custom-
ers (Swiss Reinsurance Company, 1997). There is, however, some anecdotal evidence
on catastrophic insurance coverage. In Colombia, reinsurers paid US$200 million in
claims after the recent earthquakes in Armenia. The hotel and industrial sectors in the
Dominican Republic received US$500 million in reinsurance payments after Hurricane
Georges. After the 2001 earthquakes, local insurance companies in El Salvador received
US$350 million in reinsurance payouts. However, most of these payments went to the
private sector.

Few government assets appear to be insured in the case study countries. In those cases
where there is a legal requirement for public insurance, there is incomplete information
concerning the extent to which each government ministry’s assets are actually insured.
For this analysis, it is somewhat arbitrarily assumed that five percent of each country’s
infrastructure is insured. These assumptions lead to the estimated insurance payments
shown in table 2.4.

Table 2.4. Estimates of Aid Received in Cash and Insurance Payments

Country 20-year 50-year 100-year

event event event

Aid in cash
Bolivia 2 5 9
Colombia 17 43 69
Dominican Republic 11 26 52
El Salvador 8 26 39

Insurance
Bolivia 5 15 25
Colombia 50 125 200
Dominican Republic 31 75 150
El Salvador 23 75 113
Note: Figures are in millions of US dollars. The aid estimates are based in historical data
while it is assumed that five percent of each country’s infrastructure is insured.

Budget Reallocation

The amount a government can divert from current budgets is constrained by the in-
flexibility of operational overhead, salaries, debt service, and pro-poor expenditures.
During the country visits, personnel in the budget and planning offices estimated that a
maximum of 5-10 percent of current expenditures could be redirected in the case of a
natural disaster.

53
Budget reallocation is facilitated by loan reallocation from the multilateral financial insti-
tutions. This reallocation is currently one of the primary sources of short-term liquidity
for Bolivia, Colombia, the Dominican Republic, and El Salvador. For example, immedi-
ately following the 2001 earthquakes, in one day El Salvador approved US$600 million
in financing from pending loans and then later obtained additional credits to replenish
the diverted funds.

One approach to estimating potential budget diversions is to identify discretionary and

nondiscretionary items in the national budget. This accounting-based methodology can
provide a rough indication of the upper limit of the fiscal resources that can be rerouted
to reconstruction spending. Table 2.5 can be used to establish upper limits on budget
reallocation by assuming that a large percentage of capital expenditures can be used for
reconstruction (assume 50 percent), along with a smaller percentage of wages and

Table 2.5. National Budgets by Category, 1999 (Percent unless otherwise noted)
Category Bolivia Colombia Dominican El Salvador
Republic

Total expenditure (million US$) 1,869 15,476 3,368 2,256

Total expenditure 100.0 100.0 100.0 100.0

Current expenditure 84.8 83.1 77.4 85.8
Of which pro-poor expenditure 37.3 40.2 28.2 33.3
Education 20.5 12.9 9.5 8.6
Health 0.0 25.7 9.5 12.9
Pensions 16.8 1.6 9.2 11.8
Wages and salaries 22.2 11.0 35.9 38.2
Debt service 25.3 32.0 13.2 14.3
Capital expenditure 15.2 16.9 22.6 14.2

10% from wages and salaries and 50%

of capital expenditures 9.8 9.5 14.9 10.9
10% from wages and salaries and 50%
of capital expenditures (million US$) 183 1,476 502 246
Source: Country interviews by the authors.

salaries (assume 10 percent). These estimates yield a reallocation of 10-15 percent of

total expenditures. Corresponding figures have been used for the four countries in this
study. In reality, the ability to access these funds will increase with the magnitude of
the event and sense of social solidarity felt after the event.

The cost of reallocating or diverting funds from other government budget items should
be measured in terms of the foregone returns/benefits of budgeted projects and ser-
vices. For the purposes of this analysis, it is assumed that government investments in
infrastructure bring an annual return of 16 percent (World Bank, 1994). There may, in
addition, be high political costs of diversion due to disruptions in government planning
and negotiations between the ministries over where to obtain the funds; such political
costs are not taken into account in this model.

54
In reallocating budgetary funds, it is important to keep the sources of those funds in
mind. Loan reallocation is currently one of the primary sources of short-term liquidity
for Bolivia, Colombia, the Dominican Republic, and El Salvador.

Increasing Budget Revenues

Imposing additional taxes is always problematic, even more so after a major catastro-
phe when the economy may be in a recession or low-growth situation. Although a new
tax has costs in foregone returns for the citizens of the country, the costs to the gov-
ernment are administrative and political. Taxes on financial transactions have severe
impacts on the financial sector in terms of repression and moving savings overseas. El
Salvador and Colombia have experience in increasing the tax burden after a natural
disaster, but this is not the case in Bolivia and the Dominican Republic. In El Salvador
after the 2001 earthquakes, the government increased its revenues by reducing the
number and extent of previous tax exemptions and improving tax collection. Recon-
struction after the 1999 earthquakes in Colombia’s coffee-growing area was primarily
financed by the creation of a new tax a few months prior (the dos por mil) that initially
taxed 0.2 percent of all financial transactions and now taxes 0.3 percent. This tax,
which was not originally intended for disaster relief and reconstruction, generated
US$900 million annually in revenues.

This analysis assumes that Bolivia and the Dominican Republic are unable to raise addi-
tional tax revenues after an event, whereas El Salvador and Colombia would be able to
increases revenues by 5, 10, and 15 percent after 20-, 50-, and 100-year events, re-
spectively.

Expanding Domestic Credit

Borrowing domestically also has associated costs and constraints. Domestic borrowing,
if at all possible, may compress domestic consumption, particularly in shallow credit
markets. This may result in a rise in the interest rate and a crowding out of domestic
investment.

Domestic credit can be obtained from commercial banks or, in all the case study coun-
tries except Colombia, from the central bank. However, in El Salvador and the Domini-
can Republic, central bank borrowing is constitutionally legal only in the case of natural
disasters. The sale of government bonds to the central bank is potentially inflationary if
money growth is not in proportion to real GDP growth. Using the foreign exchange re-
serves of the central bank carries the risk of provoking a balance-of-payments crisis
due to the lack of needed reserves for imports. The risks and costs of these options
make them problematic; the World Bank and International Monetary Fund (IMF) have
strongly recommended against their use.10

A country’s ability to increase domestic credit is constrained by both the government’s

ability to repay and the depth of the domestic capital markets. In Bolivia, domestic
credit is scarce. Yet due to the low growth and inflation rates associated with the cur-
rent recession, the government has in recent years borrowed close to US$200 million

10
In an assessment of financing options following the 2001 El Salvador earthquakes, the World Bank/IMF
team stated: “Under any monetary system, a country needs to maintain a strong underlying fiscal position
and a sound credit policy, with an adequate cushion of net international reserves, to preserve macroeconomic
stability. Expanding the money supply or reducing the central bank’s net international reserves is never op-
tional sources of financing for reconstruction costs.”

55
from the central bank. Sources disagree over whether the central bank will be a viable
alternative for funding reconstruction in the future.

In Colombia, domestic credit is also considered extremely scarce and capital markets
shallow (IDB, 2000a). In 1998, Colombia experienced a severe commodity price shock
as world market prices for oil and coffee fell. External financing became very expensive,
and the Colombian government turned to the domestic market resulting in the already
high real interest rates rising to 20 percent, depressing economic activity (World Bank,
2002). No extra domestic credit is assumed for Colombia.

In the Dominican Republic, investor confidence in domestic government bonds is low

because bond interest payments in the past were delayed. To tap domestic sources af-
ter a disaster, the central bank can legally force commercial banks to loan funds to the
government. After Hurricane Georges, the government obtained US$150 million in this
way. Sources estimate that another such credit could be obtained in the future.

As domestic debt tends to be short-term, high interest rates will be charged on all gov-
ernment debt and the cost of repayment will go up. Domestic interest rates are already
high in the case study countries. In El Salvador, it is estimated that there is no addi-
tional domestic market as investors invest directly in external markets.

Based on these assumptions, estimates of additional domestic credit available (or con-
straints on this credit) after a catastrophic event would be US$100 million in Bolivia,
US$150 million in the Dominican Republic, and none in Colombia or El Salvador.

Accessing Additional External Credit

Accessing additional external credit increases future debt service obligations and re-
duces a country’s ability to take on additional debt. Constraints for external credit come
from the demand side as well as the supply side. Demand is restricted by external debt
sustainability. The Highly Indebted Poor Countries Initiative (HIPC) assesses on a regu-
lar basis the debt sustainability for developing countries. The main indicator used in the
HIPC is the ratio of the net present value of debt to exports. A ratio of less than 150
percent is generally regarded as a sustainable value for this indicator. Another impor-
tant indicator is the debt service/exports ratio for which a value above 20 percent sig-
nals a problematic debt situation. For the case study countries, indicators for 2001 are
given in table 2.6.

Table 2.6. Estimates of Additional Domestic Credit Available

(Percent)
Country Debt service/ Net present value
exports of debt/exports
Bolivia 13 133
Colombia 48 222
Dominican Republic 10 74
El Salvador 8 98
Sources: IMF/WB (2001a, 2001b); IDB (2002a, 2002b).

According to the HIPC metric, in 2001 Bolivia, the Dominican Republic, and El Salvador
had sustainable levels of debt, whereas Colombia’s debt had reached unsustainable lev-
els. Table 2.6 also shows that Bolivia, after qualifying for significant debt relief in 2001,
is now under the ceiling of 150 percent (IMF and World Bank, 2001b). For the Domini-

56
can Republic, external indebtedness is less severe and also projected to stay roughly
constant. This is also the case for El Salvador, despite additional financing needs after
the earthquakes of 2001. This analysis assumes that extra borrowing is allowed only to
the extent that these two indicators remain below critical levels. The more binding indi-
cator is generally the net present value of debt to exports. For Colombia, no extra bor-
rowing is assumed to be feasible due to the critical debt situation.

On the supply side, multilateral financial institutions offer loans at more generous terms
than borrowing at market conditions. Whereas the supply of international loans is po-
tentially unconstrained for the purposes of reconstruction financing, the availability of
loans from multilateral financial institutions is limited by the willingness of the donor
community to grant subsidized credit. Eligibility for highly concessional loans, that is,
loans offered at better than market rates, is determined by per capita income. The
World Bank offers highly concessional loans to the poorest low-income countries with a
per capita income of less than US$885 in 2000. Countries with higher incomes have to
borrow on significantly less favorable terms.

All case study countries are classified as lower-middle-income countries (World Bank,
2001a). However, Bolivia has the lowest per capita income (US$980 in 2001), and the
World Bank defines it as a blend country, that is, it is eligible for concessional condi-
tions: 0.75 percent interest rate, 35-year maturity, and a 10-year grace period.

For El Salvador, the conditions outlined in the IMF and World Bank (2001a) assessment
of the 2001 earthquakes were used: 7.5 percent interest rate, 20-year maturity, and 5-
year grace period. These conditions were also assumed to hold for Colombia and the
Dominican Republic.

Capital market conditions for issuing bonds are contingent on sovereign ratings from
agencies such as Moody’s. The lower the rating, the higher the risk premium and total
interest on bonds will be. According to this rating, domestic Colombian bonds and El
Salvadoran domestic and international bonds are still in the investment grade category,
whereas the others fall below this rating. The ratings imply a risk premium or spread on
top of the risk-free benchmark interest rate of 30-year U.S. treasury bonds, which, as
of March 11, 2002, was at 5.7 percent. Bolivia, as part of the HIPC agreements, has
renounced its ability to take out additional debt at commercial rates.

The final estimates of additional external credit available to the case study countries
indicates that all of them have credit available through the IDB and the World Bank of
up to US$200 million, while only the Dominican Republic and El Salvador have addi-
tional market credit of up to US$800 million.

Summary of the Resource Gap for the Four Case Study Countries

It is useful to compare the estimates on the upper limits and constraints on the sources
of financing to the demand for financing derived from the earlier disaster risk estimates
and analysis of the government’s liabilities. This generates estimates of the potential
resource gap for each of the four case study countries. Note that the resource gap does
not take into account the costs of the financing sources, but rather the upper bounds or
constraints on the availability of these sources. The estimates of the resource gap for
the four case study countries are presented in table 2.7. Caution should be exercised,
however, in the interpretation of the results since they are based, in the absence of ex-
act information, on many assumptions.

57
Table 2.7. Calculated Resources Gap for the Four Case Study Countries

20-year event 50-year event 100-year event

Item El El El
Bolivia Colombia Dom. Salvador Bolivia Colombia Dom. Salvador Bolivia Colombia Dom. Salvador
Rep Rep Rep

Direct Damage 200 2,000 1,250 900 600 5,000 3,000 3,000 1,000 8,000 6,000 4,500
Government responsibility 100 1,000 625 450 300 2,500 1,500 1,500 500 4,000 3,000 2,250

Aid 2 17 11 8 5 43 26 26 9 69 52 39
Insurance payments 5 50 31 23 15 125 75 75 25 200 150 113
Budget reallocation 250 1,500 500 250 250 1,500 500 250 250 1,500 500 250
New taxes 0 500 0 90 0 1,000 0 180 0 1,500 0 270
Domestic credit 100 0 150 0 100 0 150 0 100 0 150 0
External credit IDB/WB 100 100 100 200 200 200 200 200 200 200 200
External credit market 0 0 800 0 0 800 800 0 0 800 800

Resource gap None None None None None None None None None 531 1,148 579
Resource gap w/o IDB/WB None None None None None None None 169 116 731 1,348 779

Additional debt 0 0 83 80 30 0 899 969 216 200 1,150 1,000

Note: Figures are in millions of US dollars. The numbers are based on data in tables 2.1 – 2.6 and suppositions explained in the text. They are used in this study to show how
tradeoffs in policies could be understood, not to demonstrate the actual tradeoffs for each country.
It appears that low hazard and access to World Bank and IMF loans protects Bolivia
from shortfalls in reconstruction financing. It is important, however, to note that with-
out access to World Bank and IMF loans, there is an estimated 1-in-100 chance annu-
ally of a shortfall of over US$100 million due to natural disaster risk.

In the case of Colombia, there is an estimated resource gap associated with 100-year
events. This shortfall is not as dramatic as calculated for El Salvador and the Dominican
Republic because of a certain demonstrated ability of Colombia to raise taxes. Under
these assumptions, Colombia could raise more than US$2 billion after a major disaster,
instead of the estimated 15 percent increase of US$1.5 billion, the resource gap for Co-
lombia would disappear.

In the case of El Salvador, however, the resource gap shown for 100-year events is less
likely to be filled. The magnitude of this gap depends primarily on the assumption about
additional credit the country could access in external markets. The current assumption
sets external credit limits at an additional US$1 billion. Only if El Salvador could raise
more than an additional US$579 million could it, too, avoid a resource gap at least 99
percent of the time. If El Salvador has less access to credit markets than estimated, it
would not only face a resource gap for 100-year events, but also for 50-year events. If
El Salvador could not tap into US$169 million in World Bank and IMF loans after a 50-
year event, it would have a higher annual probability of experiencing a resource gap
(table 2.7).

Financing Reconstruction

This section reviews the specific policies used by each of the four countries to finance
post-disaster reconstruction. Two comments about this review deserve mention. First,
all of these countries are in the process of dramatically changing their national strate-
gies to cope with natural disasters. In the Dominican Republic, the national assembly
has been considering new legislation. The country is redirecting its civil defense opera-
tions. El Salvador has recently passed new legislation and is in the process of reorganiz-
ing its internal government structure for coping with disasters (2003). Bolivia has also
just passed new legislation, and is in the process of designing regulations to implement
the new proposed structure. During the time of the country visit, the new legislation
had not yet been officially published. Finally, Colombia has modified its long-standing
legislation and is also now reorganizing its administrative structure. While all these
changes are important for each country, and considerable data were collected on the
new realignments of responsibility, the purpose of this study is to examine how recon-
struction is financed. Often, that decision is outside the expertise of the newly reorgan-
ized or recently created agencies in each country.

Second, the technical details about existing financing strategies formed the basis for
much of the work done in calculating the resource gap for each country. The discussion
of the variables used to determine the resource gap included much of the information
gleaned from the country visit. Those assumptions will not be repeated here. Rather,
this section highlights some of the important findings from the country visits.

Current Government Strategies for Financing Reconstruction

Reliance on Loan Diversions

All of the studied countries depend heavily on loan diversions for post-disaster financ-
ing. Most recently, El Salvador reoriented US$300 million in existing loans, including a

59
US$75 million IDB loan that had been approved but not ratified. It was ratified in the
hours following the first 2001 earthquake and was immediately used. In addition to re-
directing loans from international financial institutions, deferral of commercial credits
also played a role. The Dominican Republic received a one-year moratorium on its debt
service payments from the Paris Club after Hurricane Georges.

Role of the Central Bank

The central bank plays a critical role in some of the selected countries in providing im-
mediate liquidity. Bolivia and the Dominican Republic rely on short-term loans from
their central banks to fund immediate post-disaster assistance. Neither Colombia nor El
Salvador is permitted to borrow from the central bank.

Budget Reallocation

Bolivia, Colombia, and El Salvador have all experienced significant difficulties in reallo-
cating current budgets to finance reconstruction projects. Sources in El Salvador com-
mented on how fortunate they were that the 2001 earthquakes occurred at the begin-
ning of the fiscal year, when there was still some flexibility in the national budget. Some
flexibility also exists within individual ministries. INVIAS, the Colombian agency respon-
sible for road construction and maintenance, reallocates its annual budget between
maintenance and post-disaster reconstruction every year.

The Dominican Republic has a highly flexible budget. Pursuant to Section 1402 of the
Constitution, the Dominican Republic allocates 25 percent of all discretionary budget
funds allotted to the president. This generally amounts to 12.5 percent of total govern-
ment expenditures each year. These funds are kept in an accumulating fund for emer-
gencies. In practice, this provides a discretionary fund for the office of the president
that is often used to meet immediate needs after a disaster. Outside this special fund,
there is little discretionary spending in the national budget.

Increasing Taxes

Bolivia, the Dominican Republic, and El Salvador are limited in their ability to raise
revenues through tax increases. Typical sources of additional revenues are reductions in
tax exemptions (as seen in El Salvador after the recent earthquakes) and mandatory
employee contributions (as seen in the Dominican Republic after Hurricane Georges).

The use of increased tax revenue to fund reconstruction has historically been used in
Colombia. Colombians have relatively higher per capita incomes, and the disaster risk in
Colombia is geographically diversified across the entire country, so there is little risk
that major production centers would all be affected simultaneously. Combined with a
strong feeling of social solidarity, Colombians have effectively found tax revenue
sources to fund post-disaster reconstruction. It was the feeling of the finance ministry
that it would be able to access additional taxes in the event of future emergencies.

International Aid

International aid is important for all the countries in the immediate post-disaster period.
Generally, this aid is directed at relief efforts, not reconstruction. Most of the relief aid
is gifts in kind: food, machinery, volunteer services, clothing, and the like. The cash
portion of the aid is less than 10 percent of the face value of contributed aid. El
Salvador had specific figures related to the cash and noncash portions of the aid. In

60
discussions with the other countries, they confirmed the cash portion of their aid.
Overall, international aid makes up a small fraction of the funds needed for
reconstruction work.

External Debt

El Salvador and the Dominican Republic both use external commercial debt to provide
additional needed financial resources. For both countries, the majority of reconstruction
in recent years has been financed by issuing bonds in external commercial markets. The
Dominican Republic recently borrowed US$500 million in the commercial market. El
Salvador has one of the lowest debt ratios in Latin America, and its sovereign bonds
have been investment grade since 1997. Its government took on US$700 million in new
loans to finance reconstruction after the 2001 earthquakes. Bolivia is limited from ac-
cessing new commercial external debt by its arrangements with the international finan-
cial community as a component of its debt relief package. Colombia has relatively high
levels of current debt.

Insurance Proceeds and Reserve Funds

There is current experience with both insurance and reserve funds as a source of fi-
nancing in all of the countries. They all require some level of insurance on government-
owned property. Both the Dominican Republic and El Salvador have seen substantial
increases in the use of insurance for disaster risk. Last year in both countries, premium
revenue from property and liability insurance increased by more than 10 percent. Bo-
livia has seen a dramatic increase in premium revenue, but this is primarily associated
with a new mandatory automobile program. In all the countries, the major industrial
enterprises are insured against natural hazard risk.

There is some experience in the region with reserve funds. Colombia has a reserve fund
at the national level designed explicitly for coping with natural disasters. The fund was
established in 1984 after the Popayán earthquake, and its annual funding has been sub-
ject to political will. At the end of the 1980s and the beginning of the 1990s, it was an
average of US$5 million a year, but at the time of the 1999 earthquake it was ex-
hausted. Municipal governments across Colombia also keep reserve funds. There unfor-
tunately is limited information sharing about which cities have reserve funds and how
much is contained in each one. One source estimates that the reserve funds held in Bo-
gotá and Medellín are each larger than the one the national system maintains centrally.

El Salvador has an account consisting of the proceeds received by the government from
the privatization of telecommunication assets. The rules governing disbursement from
this fund are vaguely defined, but are primarily intended for road maintenance. After
the recent earthquakes, US$100 million of an initial fund of US$500 million was used for
reconstruction.

Bolivia has passed legislation to set up a natural disaster reserve fund. The goal of the
fund is to guarantee that Bolivia has sufficient resources to meet the contribution
needed to access borrowing from the international financial institutions. The initial
US$10 million is being provided from unused funds held at the housing ministry. This
ministry played a critical role in designing the fund. It is hoped that annual funding will
amount to US$8 million.

61
Inherent Tradeoffs in Ex Ante Risk Management Strategies

An alternative set of tools can be used to manage disaster risk: ex ante risk
management tools. As noted by their name, these tools are operative before an event
occurs. They are directed at either reducing the risk (mitigation and risk prevention) or
guaranteeing resources for funding losses if a hazardous event happens (insurance).
Tools addressed at providing financial resources after an event are generally termed
risk-financing instruments.

By their very nature, ex ante risk management tools are complicated. The benefit of
these tools lies in an understanding of probability; they are valuable for planning
against an unknown future. These instruments require monies to be spent today to re-
duce the consequences of an unknown, but probably certain, future event. If the future
event does not occur, the value of the money spent to protect against the event looks
lost. Even worse, the perceived benefit of spending the funds on another project whose
benefit is immediately apparent is also lost. To use ex ante risk management tools, a
policymaker must bridge the psychological gap of weighing the cost of current expendi-
ture against future unknown but predictable consequences. This is often a hard gap to
cross.

At the country level, the tradeoff is usually framed as a being between growth (a result
of more money being spent now) and stability (a guarantee of funds to pay for future
losses). The tradeoff between stability and growth can be seen in the growth trajecto-
ries in figures 2.6 and 2.7, which show various computer runs of alternative simulated
GDP growth paths based on projected levels of growth for El Salvador. The growth tra-
jectories account for the likelihood of natural hazard losses. Figure 2.6 represents a
simulation of growth trajectories over a 10-year period generated by sampling from El
Salvador’s disaster risk distribution with frequencies proportional to the probability of
given disaster events occurring. Figure 2.7 is the same simulation with one change: the
government decides to purchase insurance against disaster losses.

The purchase of insurance slightly lowers the average growth rate trajectory from 2.53
to 2.38 percent, and the median trajectory from 2.66 to 2.40 percent, because the
funds spent on insurance cannot be spent on other growth encouraging activities. In-
surance reduces the number and magnitude of worst-case scenarios by guaranteeing
funds to rebuild. Without insurance, economic growth rates range from a maximum of
2.99 to a low of 1.11 percent; with insurance, the highest possible 10-year return is
2.55, but returns are guaranteed to never drop below 1.85 percent. In essence, maxi-
mum growth is reduced by 0.44 percentage points, but minimum growth is guaranteed
to be at least 0.75 percentage points higher. This illustrates the tradeoff inherent in an
insurance purchase: a reduction in funds spent on current growth permits a govern-
ment to protect itself against extreme future losses. This is a common, well-known fea-
ture of insurance. Applying the model to the other case study countries generates simi-
lar results.

62
Figure 2.6. Projected Annual Economic Growth for El Salvador without Insurance
3

2.8

2.6

2.4

2.2
Growth rate

1.8

1.6

1.4

1.2

1
1 2 3 4 5 6 7 8 9 10 11

Year

Figure 2.7. Projected Annual Economic Growth for El Salvador with Insurance
3

2.8

2.6

2.4

2.2
Growth rate

1.8

1.6

1.4

1.2

1
1 2 3 4 5 6 7 8 9 10 11

Year

63
Costs and Benefits of Different Ex Ante Tools

A range of ex ante tools can provide security against resource gaps. All the tools involve
an opportunity cost of resources foregone in the current period for the benefit of guar-
anteed future resources. The exact costs and benefits are specific to each tool. Table
2.8 compares the costs and benefits of three tools: reserve funds, insurance, and con-
tingent credit.

Table 2.8. Costs and Benefits of Ex Ante Financing Tools

Cost, benefit Reserve fund Insurance Contingent credit

Cost before Contribution x number Premium x number Holding fee x number of

event of years before event of years before years before event
event
Benefit after Only reserved funds All funds available All needed funds avail-
event and interest available for covered losses able
Cost after event None None Additional debt service
and reduced ability to
take out future debt

Reserve funds involve setting aside funds in highly liquid accounts held either domesti-
cally or abroad. In theory, the annual contribution to that fund should be equal to the
annual expected loss based on the risk the fund is designed to cover. The cost of these
funds is primarily the opportunity cost of not investing the funds elsewhere. Highly liq-
uid accounts offer only a 5-6 percent rate of return compared with the 16 percent rate
of return frequently attributed to investment in development projects.

Annual insurance premiums are also based on annual expected losses for the risk in
question, but they include additional administrative, capital, and uncertainty costs. In
the case of disaster risk in a developing country, insurance premiums can be twice the
cost of the annual expected loss. In contrast to reserve funds, however, insurance ar-
rangements guarantee the entirety of the covered risk regardless of how long the policy
has been in place. With reserve funds, it can take many years to build up sufficient re-
serves.

Contingent credit agreements are similar to insurance policies in that they also guaran-
tee access to sufficient funds immediately post-event. These credit arrangements, how-
ever, do not transfer the risk but simply postpone and spread it. The cost of a contin-
gent credit arrangement is very low before an event: a holding fee may cost only 0.5
percent of the amount guaranteed, but the amount disbursed will need to be repaid
with interest in the years following the event.

Mitigation is a fourth ex ante instrument but is not presented in Table 2.8. In its most
recent World Disaster Report, the Red Cross indicates that investments of US$40 billion
in disaster preparedness, prevention, and mitigation would have reduced global eco-
nomic losses in the 1990s by US$280 billion (IFRC, 2001). It is very difficult, however,
to calculate the benefits of mitigation on a countrywide basis. Currently, the benefits of
mitigation are generally demonstrated on a project-by-project basis.

In evaluating other ex ante tools, in addition to the immediate cost and benefits of each
tool, it is useful to consider the extent to which each tool provides incentives for

64
investment in mitigation. Insurance could provide these incentives directly through
reduced premiums.

The pros and cons of the different ex ante instruments can be highlighted with a nu-
merical example. In the case of El Salvador, say that after evaluating the
growth/stability tradeoff, the government decides it is willing to deal in an ex post man-
ner with all events that require less than US$1.5 billion of government funds, that is,
with all events that on average are expected to return more frequently than once every
50 years. In addition, say that the government is indifferent to all events that occur on
average less frequently than once every 100 years, that is, to all the very rare events
that would cost it more than US$2.2 billion in reconstruction. The government decides
to use an ex ante tool to cover the US$700 million in funds that it would need to avoid
a resource gap over that range of events.

Table 2.9 gives the pros and cons of the different ex ante tools in covering that risk for
El Salvador.11 The tables shows that reserve funds cost the least but also provide the
least guaranteed benefit: a reserve fund alone will only provide as much benefit as an
insurance or contingent credit arrangement if it is able to accumulate for 22 years be-
fore the first large event occurs.12 In comparing a contingent credit agreement with in-
surance, table 2.9 shows that the relative cost depends on when the event occurs and
how future payments are discounted relative to current ones. If the event does not oc-
cur at all, or occurs in the distant future, contingent credit will be considerably more
cost-efficient. If the event occurs in the near future, the 20 years of loan repayments
will cost significantly more than insurance payments that are calculated by averaging
costs over a much longer time horizon.

Table 2.9. Example of Costs and Benefits of Ex Ante Tools for El Salvador

Cost, benefit Reserve fund Insurance Contingent credit

Cost before US$19 million x number US$38 million x num- US$3.5 million x
event of years before event ber of years before number of years be-
event fore event
Benefit after US$19 million x number US$700 million avail- US$700 million avail-
event of years before event + able immediately able immediately
5 percent interest com-
pounded over that pe-
riod
Cost after None None US$56 million a year
event for 20 years after the
event in debt service
payments
Note: The numbers are based on the assumptions explained in the text.

To compare and contrast the ex ante instruments, a two-stage decision model was de-
veloped. Among others, the modeling presents the tradeoff between the probability of
having a resource gap and the cost of the tools to avoid those gaps. It generates cost-
efficient strategies for reducing the probability of a resource gap.

11
The government faces expected liabilities of US$48 million annually from its exposure to 20 to 100-year
disaster events. Of that figure, the expected loss for all 50 to 100-year events is approximately US$19 mil-
lion. The debt service payments for the contingent credit agreement are based on 5 percent interest and a

Accumulated funds in year n are 19 * Σ (1 + (1+i) + (1+i)2 + … + (1+i)n).

five-year grace period.
12

65
Figure 2.8 represents the cost of reducing the probability of a resource gap for each
tool, based on data from the Dominican Republic. The figure shows that insurance most
drastically reduces the probability of a resource gap: investing 12 percent of the annual
budget fully eliminates any such danger. A reserve fund also reduces the probability of
a resource gap, but only in the long run. It takes time to build up sufficient funds in a
reserve account to guarantee that the resource gap for a 100-year risk is eliminated.
Investments in mitigation could be considered similar from a financing point of view.
However, it is not really comparable, because mitigation measures reduce the damages
and thus the losses. The amount of up-front investment in reserve funds required to
guarantee covering the potential losses in the short run is inordinately high: 10 percent
of the budget invested in insurance provides the same loss reduction as 50 percent of
the budget invested in reserve funds.

Figure 2.8. Cost of Reduced Probability of a Resource Gap for the Dominican Republic
20%

18%
Probability of a natural hazard resource gap

16%

14%

12%

10%

8%

6%

Insurance Reserve fund

4%
Mitigation
2%

0%
0% 10% 20% 30% 40% 50% 60% 70%

Annual ex-ante payments

(% budget)

The results from the modeling example presented above for the Dominican Republic
also hold true for El Salvador. Since both Colombia and Bolivia have small or no re-
source gaps, there is no need to examine modeling for them looking solely at the re-
source gap.

Another important observation from tables 2.8 and 2.9 is the impact of contingent
credit on reducing the amount of future credit available. As noted in table 2.7, the re-
duction of the resource gap required a significant increase in debt levels in Bolivia, the
Dominican Republic, and El Salvador. It may be that it is prudent policy for countries to
also reduce their dependence on borrowing as the primary tool to generate needed
post-disaster financing resources. Insurance and sufficiently large reserve funds reduce
the dependency of countries on ex post borrowing to finance reconstruction and pre-
serve their buffer of potential future credits for other eventualities.

66
Ex Ante Financing Instruments in Practice

Reserve Funds

There are many differences among the ex ante options and, in particular, issues that
arise with their implementation.

A number of countries have been exploring the use of reserve funds as a means to pro-
vide post-disaster funding. The best-known fund is FONDEN in Mexico. This fund is an
annual budgetary allocation for natural disaster expenditures. It does not accumulate
from year to year (Kreimer and others, 1999). There is a new fiscal stabilization fund in
Peru, but no record exists as to its effectiveness (World Bank, 2001c). There is no re-
serve fund currently in use that is set aside solely for natural disasters. As demon-
strated in the modeling, reserve funds alone do not appear as the most cost effective
way of reducing the resource gap. The problem is that the time frame involved in ac-
cumulating a sufficient fund is so long that the fund will not effectively protect against
large events that occur in the first years of accumulation.

In the case study countries, the topic of reserve funds was often mentioned. In many
countries, the budgetary process does not allow the accumulation of reserve funds be-
tween periods. Bolivia is exploring a new model for establishing a reserve fund. It plans
on making annual expenditures to a fund held outside the country and monitored by an
international financial institution. The issue with disaster reserve funds is similar to the
problems surrounding reserve funds established for pensions or other purposes. It may
be that a disaster reserve fund lacks a natural political constituency to protect it as it
accumulates.

Contingent Credit

Contingent credit agreements offer, for a small annual fee, the option to borrow a given
amount immediately at a previously determined interest rate. To measure the benefits
of contingent credit arrangements, would require estimating the cost of having to wait
several months after a disaster to reorient funds from existing projects and acquire new
loans, versus having the funds on hand in a matter of days.

Relative to the other ex ante options, these credit arrangements require a smaller fee
paid up front, but they increase the debt burden after an event. Low interest ex post
borrowing would be more attractive than contingent credit agreements if it were quickly
available after an event. If expedient ex post event loans and rates were guaranteed to
a country through some other mechanism, there would be no value in paying the fee for
a formal contingency credit agreement. On the other hand, if available loans or associ-
ated rates were subject to deteriorating terms after an event, or to a lengthy process of
negotiation and processing, then the holding fee may be worthwhile. The case study
countries did not experience any deterioration of their terms of credit after recent disas-
ters.

Insurance

The modeling results here show insurance as a cost-effective option for reducing the
probability of a resource gap and protecting countries from worsening their debt posi-
tion after disaster events. Insurance also could provide current-period incentives for in-
vestment in physical mitigation through risk-adjusted premiums. The best example of
the use of private insurance as a viable means of reducing societal risk comes from the

67
mutual insurance companies, founded in early nineteenth century New England, that
covered factories (Bainbridge, 1952). These mutual companies offered factories protec-
tion against potentially large losses from fire in return for a small premium. In order to
reduce risk, the mutuals required inspection of a factory both prior to issuing a policy
and after one was in force. Customers who were regarded as poor risks had their poli-
cies canceled; factories that instituted loss prevention measures received premium re-
ductions.

As the mutual companies gained experience with fire risks, they set up research de-
partments to determine what factors caused fires and how to reduce losses by concen-
trating on those factors. For example, the Boston Manufacturers’ Mutual Company
worked with lantern manufacturers to encourage them to develop safer designs, and
required policyholders to purchase lanterns only from companies whose products met
their specifications. Manufacturers’ Mutual hired researchers to find ways to reduce the
risk of fire, for example, by developing nonflammable lubricating oils. It then shared
these findings with key trade associations, and distributed educational pamphlets on
preventing fires to textile mill owners.

In many cases, mutual companies would only offer insurance to companies that
adopted specific loss prevention methods. For example, Spinners Mutual only insured
factories that installed automatic sprinkler systems. Manufacturers’ Mutual in Provi-
dence, Rhode Island, developed specifications for fire hoses and advised mills to buy
only from companies whose hoses met those specifications. By researching and requir-
ing loss prevention techniques and inspecting facilities before issuing or renewing a pol-
icy, nineteenth century insurers were able to reduce losses dramatically and provide
coverage against risks for which there had previously been no protection.

In these ways, insurance companies provide incentives for mitigation. Insuring national
infrastructure would provide the contractual obligation to maintain the structures and
disincentives for placing new structures in high-risk areas. Bringing the cost of disaster
risk into the current period also provides political impetus to allocate monies from the
current budget funds for mitigation and prevention works.

The benefits of insurance naturally raise issues regarding the status of insurance in
Latin America. Since the beginning of the decade, Latin America has gone through a
significant transformation in its insurance markets. As Swiss Reinsurance Company
(2000) notes, “growth was three times as high as in the industrialized countries.” Eco-
nomic reforms at the beginning of the 1990s created new business opportunities across
the board for both domestic and foreign insurance companies. From 1990 to 1998,
premium incomes in the nonlife sector increased on average by 4.6 percent annually in
real terms in the region’s six largest countries (Swiss Reinsurance Company, 2000). In
2000, nonlife insurance premiums increased in the region by another 4.6 percent to
US$27.1 billion. Increasing premium rates, especially for natural hazards, contributed
to this growth. Both the Dominican Republic and El Salvador showed double-digit
growth in their premiums. Colombia increased premiums by 4.7 percent in 2000 (Swiss
Reinsurance Company, 2001). The premium volume for Bolivia was too low to be sepa-
rately captured in worldwide surveys, but according to the insurance department, a
dramatic increase in premiums occurred in 2000 due to a new mandatory automobile
insurance program.

While this growth is impressive in the region, it is based on a small, beginning insurance
market. By all measures, the insurance penetration in Latin America is small. Even
among emerging markets, the countries in this study all ranked among those with the

68
lowest penetration. On average, less than US$50 per capita is spent on insurance and
insurance purchases constitute less than 1 percent of GDP. Latin America constitutes
less than 3 percent of the worldwide market for nonlife insurance premiums (Swiss Re-
insurance Company, 2001).

Why is insurance not more prevalent? Insurance requires a complex series of laws,
regulations, and administrative agencies. The requirements to operate an effective
regulatory and supervisory scheme are complex. In addition to the regulatory issues,
there are issues related to the fundamental structure of the market for insurance. For
example, many countries may be too small to provide adequate risk diversification to
properly support a national insurance scheme. Proposals to create regional insurance
markets hope to increase risk diversification and potential market size, thereby making
the market more attractive for the insurance industry and lowering the cost of insur-
ance. A larger potential market subject to a uniform regulatory scheme may encourage
the international insurance industry to help develop viable markets. Regional proposals,
like the World Bank initiative for a Central American insurance market, are based on
overcoming impediments to the supply of insurance.

Additional limitations to insurance relate to broader market cycles. The availability of

insurance for natural hazard risk in developing countries is dependant on a series of
factors that impact the world insurance market. The events of September 11, 2001
have influenced the availability and cost of all insurance for catastrophes, natural or
manmade, worldwide.

Professional risk bearers, like insurance companies, are fully capable of modifying their
products to adapt to local needs. However, there will be little willingness on their part to
do so if no demand exists for the modified products. Some countries make insurance in
certain sectors mandatory. For example, Turkey adopted this strategy by requiring
homeowners to insure themselves. Another approach to stimulate the insurance market
is to demonstrate the benefits of insurance by taking out policies at the government
level, for example by insuring government-owned buildings and infrastructure. To some
extent, this has been the approach taken by the four studied countries.

Despite the limitations of markets in developing countries, the modeling would indicate
that policies directed at increasing the efficiency of insurance markets in the developing
world could yield large dividends.

Other Risk Transfer Instruments

A new strategy for dealing with risk shifting of disaster loss has developed since 1996:
catastrophe hedges. These instruments bring the same benefits as insurance, but they
are set up in a way that brings risk directly to the capital market, bypassing the tradi-
tional path of insurance. Since the cost of catastrophe insurance is dominated by capac-
ity limitations, and the capital markets lack a capacity constraint, pricing theoretically
should be competitive for these products in the long term (Doherty, 1997; Pollner,
2000).

Considerable interest has developed regarding the use of derivatives as replacements or

complements to catastrophe insurance. The strategy in developing these derivatives is
to create a financial instrument that is negatively correlated (has a negative covariance
value) with both the risk of a portfolio of stocks or bonds and the risk of loss from a
catastrophic event. By creating a derivative with a return negatively correlated with a
portfolio of stocks, the ownership of the derivative is attractive to a stock portfolio

69
owner. The correlation matrix between catastrophe risk and other financial assets may
be as low as –0.13 (Hodgson, 1997). In addition, these derivatives can act like insur-
ance. Their payment is based on some measure of realized catastrophe risk. As a result,
these hedges should be of interest to both portfolio investors and the owners of catas-
trophe risk.

The new capital market hedges for catastrophe risk are varied in how they attempt to
hedge risk. In essence, the hedges vary on two primary variables: whether they are
issued as equity or debt, and whether they pay based on indemnifying losses or on the
occurrence of a specified physical event.

The proceeds from an equity hedge need not be repaid. The owner of the hedge would
book the proceeds from the hedge as an asset. A debt hedge would require that the
proceeds received by the owner of the hedge must be repaid. The proceeds from the
hedge would be booked as a liability, to be repaid over some predetermined future pe-
riod. A contingent surplus would be a debt hedge.

Insurance is an indemnity instrument, which is designed to reimburse incurred dam-

ages. For example, fire insurance will reimburse homeowners for fire damages subject
to the deductible and limits of the policy. If a homeowner has a fire, but no damage is
caused, the homeowner is not entitled to any reimbursement because there is no dam-
age to be indemnified.

An event hedge pays if a specified physical event occurs, regardless of whether dam-
ages are suffered. For example, weather derivatives based on the number of cold days
in winter pay the owner of the derivative solely based on the number of cold days, re-
gardless of the damages incurred by the owner of the derivative. Table 2.10 delineates
these hedges.

Table 2.10. Overview of Noninsurance Catastrophe Hedge Instruments

Type of instrument
Basis for loss
payments Equity Debt

Indemnity Catastrophe bonds Contingent surplus notes

Exchange traded Catastrophe equity puts
Event Catastrophe options
Weather derivatives Catastrophe swaps

The new capital market instruments on the market fall into six broad categories:

• Catastrophe bonds. These are “Acts of God” bonds that pay investors high yields,
but are subject to default if a defined catastrophe occurs during the life of the bond.
The investor appeal comes from a high yield with a low probability of default. Funds
obtained from the sale of these bonds are normally invested in risk-free instru-

•
ments, and the interest earned reduces the net cost of the bond to the issuer.
Contingent surplus notes. These notes are essentially “put” rights that allow the
owner of the note to issue debt to prespecified buyers in the event of a catastrophic
event. The owner of the note pays a fee to the potential debt buyers for their com-

•
mitment to buy the debt.
Exchange-traded catastrophe options. The property claims service options that trade
on the Chicago Board of Trade provide for the purchaser of the option to demand

70
 payment under an option contract if the claims index surpasses a prespecified level.
The indexes used cover different areas of the United States and reflect aggregate

•
reported claims by the insurance industry.
Catastrophe equity puts. Equity puts are another form of an option that permits the
insurer to sell equity shares on demand after a major disaster. The insurer pays a

•
fee for the put option. These instruments are similar to contingent surplus notes.
Weather derivatives. Weather derivatives are contracts that provide payouts in the
event a specified number of days occurring with temperatures or rainfall above or

•
below a specified trigger point.
Catastrophe swaps. These derivatives use capital market players as counterparties.
An insurance portfolio with potential payment liability is swapped for a security and
its associated cash flow payment obligations. An insurer takes on the obligation to
pay an investor periodic payments on a specified portfolio of securities that the in-
vestor was originally liable to pay, while the investor assumes the liability of the in-
surer to make payments in the event of a catastrophe.

The range of instruments is both diverse and growing (Pollner, 2000). Since 1996,
nearly US$4 billion in catastrophe hedges have been placed in the capital markets. The
most significant limitation to using noninsurance hedges, however, is their cost. Securi-
ties markets are notoriously complex, and the tools they develop are often expensive
for single transactions (Chichilnisky and Heal, 1997). Their high transaction costs cur-
rently make “Acts of God” bonds much more expensive than insurance. By some esti-
mates, they are twice as expensive (Swiss Reinsurance Company, 1999). For this rea-
son, their application may be limited to large transactions that may exceed the capacity
of the insurance market to provide protection. It is estimated that insurance is still
more than 95 percent of the market for catastrophe risk shifting (Swiss Reinsurance
Company, 1997).

Mitigation

Initial investments in physical mitigation can significantly and cost-effectively reduce

vulnerability of a construction project to natural phenomena. After basic precautions are
taken, however, considerably larger expenditures may be required to achieve additional
reductions in risk. As shown in figure 2.9, at some point it may be more cost effective
to transfer the remaining risk or find alternative ways to finance it than to attempt to
mitigate it away. For example, constructing a new structure to code is relatively cheap,
whereas constructing the perfect structure to resist all shocks is inordinately expensive.
It is therefore difficult to estimate on a national level the aggregate benefits of physical
mitigation without knowing where that mitigation will be applied.

For the purpose of this modeling, it is assumed that the first dollars invested in physical
mitigation have a large impact, reducing losses by four to five or more times the
amount invested in mitigation. After these initial investments have been made, addi-
tional investments have a decreasing marginal benefit. As a result, investment in physi-
cal mitigation is beneficial only up to a point. The value of this point depends entirely on
the curve chosen to describe decreasing returns to mitigation. Afterward there is a re-
sidual risk that mitigation does not efficiently cover.

The analysis also shows that physical mitigation takes a longer time to close the re-
source gap than does insurance. On the other hand, there are some significant benefits
to mitigation not captured in the modeling. If mitigation is integrated as a component of
a project, the marginal cost for each project may be small. The cumulative effect of in-
cluding mitigation as a piece of each new project may be considerably less expensive

71
than investing in physical mitigation on a countrywide basis for existing structures. And
various risk prevention measures, like land-use planning and building codes, may be
inexpensive to implement and may substantially reduce long-term risk.

Figure 2.9. Decreasing Marginal Returns to Investment in Physical Mitigation

Risk (in $)

Expenditure on
Mitigation (in $)

Policy Recommendations

In making policy recommendations based on this research, a warning needs to be em-

phasized. The models contained herein are not intended to dictate policy decisions.
Rather, the models are tools in a methodology for exploring alternative policy options,
describing the working assumptions and data used in arriving at the results highlighted,
but not the level of detail required to make effective final policy recommendations. Con-
siderably more attention to the structure of the models and the required data will be
needed to frame policy decisions.

Another critical factor in the modeling is that it is based on the assumption that the high
levels of post-disaster support currently provided to the case study countries will con-
tinue at existing levels. Both the international aid and international finance communities
have expressed considerable concern about their ability to sustain their current level of
funding. The models assume that funding will be available in the same proportional
amount relative to disasters as has been the case to date. If this assumption were
wrong, it would dramatically impact the results of this analysis, particularly for the most
vulnerable countries.

The analysis of the cases suggests that small countries with historically high incidence
of natural disasters may face the possibility of significant shortfalls in their ability to fi-
nance post-disaster reconstruction. This is the situation in the Dominican Republic and
El Salvador, cases analyzed in this study. For large countries with more modest or di-
versified disaster risk, the study suggests a greater ability to absorb losses from disas-
ters. This has been the case for Bolivia, which also has had sufficient resources to re-
spond, thanks to traditional access to low interest loans from multilateral institutions.
On the other hand, in the similarly geographically diverse but more populous and rela-
tively high per capita income country of Colombia, the government has been able to
expand tax revenues to cover disaster losses. But for any country, changes in their vul-
nerability (increasing urbanization in disaster prone areas, for example) or economic
situation should compel a reexamination of past financial solutions to finance potential
future disaster losses.

72
Keeping the above precautions in mind, a number of policy recommendations can be
made:

• Governments should analyze the risk of natural hazard events. The modeling tech-
niques for measuring risk exist and most countries have the necessary data to
evaluate hazard exposure and vulnerability. What is lacking is the time and re-
sources to integrate the known information, thus limiting the ability of the govern-
ment to plan for disasters, instead of only responding to them of course, modeling is
expensive and obtaining detail is very time consuming. At a minimum, working es-
timates of potential losses should be developed. The modeling should be done at the
national, regional, and municipal levels for all essential infrastructure and buildings.
Schools, hospitals, bridges, and roads are all examples of assets for which models

•
can be developed.
Governments should guarantee the institutional capacity to avoid rebuilding expo-
sure after a disaster occurs. Land-use planning, building codes, and proper recon-
struction standards should be developed before disaster occurs. Reconstruction after
a disaster should provide the opportunity to implement the proper risk reduction

•
policies.
Each government needs to create a clear inventory of risks for which it is responsi-
ble. If the government is responsible for housing reconstruction, this should be
made clear and the obligation budgeted. If the government does not assume re-
sponsibility for some private sector risk, it should examine strategies to assist the

•
private sector to assume that risk on its own behalf.
Multilateral financial institutions and their client countries should look at the options
imbedded in existing loans to convert to reconstruction financing at the time of a
disaster. The terms of the conversion should be openly discussed in advance, and
not be left to a decision process made in the immediate post-disaster period. In this
way, issues of loss prevention and mitigation can receive the policy hearing often

•
missing in the crisis atmosphere of the immediate post-disaster period.
Governments should evaluate their current strategies to purchase insurance against
natural disaster risk. Insurance purchases should be part of an integrated risk man-
agement strategy. The governments need to understand the different ways public
entities may purchase insurance, and explore the means to make such purchases
more efficient. Considerable savings or substantial increases in insurance protection
could be accomplished in each country if it systematically reviewed its insurance
purchasing activities.

73
74
 Bibliography

Albala-Betrand, J.M. 1993. Political Economy of Large Natural Disasters with Special
Reference to Developing Countries. Oxford: Clarendon Press.

Ames, B., W. Brown, S. Devarajan, and A. Izquierdo. 2001. Macroeconomic Policy and
Poverty Reduction. Washington, D.C.: World Bank.

Andersen, T. 2001. “Managing Economic Exposures of Natural Disasters: Exploring Al-

ternative Financial Risk Management Opportunities and Instruments.” Special
report prepared for the IDB. Internal document.

Anderson, M., and P. Woodrow. 1989. Rising from the Ashes: Development Strategies
in Times of Disaster. Boulder, CO: Westview Press.

Applied Technology Council (ATC). 1985. Earthquake Damage Evaluation Data for Cali-
fornia. ATC-13 Report. Redwood City, CA, and Washington, D.C.: ATC.

Asociación Colombiana de Ingeniería Sísmica (AIS). 1996. Estudio General de Amenaza

Sísmica de Colombia. Bogotá: AIS.

_____. 1997. Normas Colombianas de Diseño y Construcción Sismo Resistente NSR-98,

Ley 400 de 1997, Decreto 33 de 1998. Bogotá: AIS.

Awareness and Preparedness for Emergencies on a Local Level (APELL). 1995. APELL
Worldwide. New York: UNEP.

Bainbridge, J. 1952. Biography of an Idea: The Story of Mutual Casualty and Fire Insur-
ance. Garden City, NY: Doubleday & Co.

Baker, J.L., L. Christiaensen, C. Scott, and Q. Wodon. 2001. “Pro-poor public

expenditures: Development targets, social programs, and fiscal sustainability.”
Washington, D.C.: World Bank.

Bate, P. 1999. “Insurance against the next Mitch.” IDB America (August).

Bender, S. 1991. Primer on Natural Hazard Management in Integrated Regional Devel-

opment Planning. Department of Regional Development and Environment. Ex-
ecutive Secretariat for Economic and Social Affairs. Washington, D.C.: Organiza-
tion of American States.

Bendimerad, F., F. Wenzel, B. Wisner, and A.G. Green. 1999. “The earthquakes and
megacities initiative.” Science International 5-7.

Benson, C. 1997a. “The Economic Impact of Natural Disasters in Fiji.” London: Over-
seas Development Institute.

_____. 1997b. “The Economic Impact of Natural Disasters in Viet Nam.” London: Over-
seas Development Institute.

75
_____. 1999. “The Economic Impact of Natural Disasters in the Philippines.” London:
Overseas Development Institute.

Bhandari, R.K. 2000. “National Natural Disaster Knowledge Network for India.” Asian
Disaster Management News, 6 (July-December 2000): 3-4.

Bhatt, M. 2001. “Vulnerability, Innovations, and the Poor: The Demand Side.” In Inno-
vations In Managing Catastrophic Risks: How Can They Help The Poor? Financial
Sector Department and Disaster Management Facility, World Bank, Washington,
D.C.

Blaike, P., T. Cannon, I. Davis, and B. Wisner. 1994. At Risk, Natural Hazards, People’s
Vulnerability and Disasters. London: Routledge Publishers.

Bolivia HIPC document: http://www.worldbank.org/hipc/country-cases/bolivia/Bolivia-

CP.pdf.

Caballeros, R., and R. Zapata Marti. 2000. “A Matter of Development: How to Reduce
Vulnerability in the Face of Natural Disasters.” Joint ECLAC-IDB paper presented
at the seminar Confronting Natural Disasters: A Matter of Development, New Or-
leans, March 25-26.

Cardona, Omar Dario. 2001. “Diseño del Sistema Nacional de Prevención, Mitigación y
Repuesta - PMR (Gestión de Riesgos).” Unidad Ejecutora Sectorial del Subpro-
grama de Prevención de Desastres, Secretariado Técnica de la Presidencia, Re-
pública Dominicana.

Cardona, O.D., and L.E. Yamín. 1997. "Seismic Microzonation and Estimation of Earth-
quake Loss Scenarios: Integrated Risk Mitigation Project of Bogotá, Colombia."
Earthquake Spectra 13(4): 795-814.

Cardona, O.D., M.G. Ordaz, L.E. Yamín, and S. Arámbula. 2001. "Evaluación de pérdi-
das por sismo en Colombia: aplicación a la industria aseguradora." VIII Semina-
rio Internacional y I Congreso Nacional de Ingeniería Sísmica, Universidad de
Los Andes, Bogotá. August 21-25.

Carter, W.N. 1992. Disaster Management: A Disaster Manager’s Handbook. Manila:

Asian Development Bank.

Carter, W.N., J.M. Chung, and S.P. Gupta. 1991. “South Pacific Country Study.” Disas-
ter Mitigation in Asia and the Pacific. ADB, ed. Manila: Asian Development Bank.

Charvériat, Céline. 2000. Natural Disasters in Latin America and the Caribbean: An
Overview of Risk. IDB Research Department Working Paper #434. Inter-
American Development Bank, Washington, D.C.

Chichilnisky, Graciela, and G. Heal. 1997. “Managing Unknown Risks: The Future of
Global Reinsurance.” Paine Webber Working Paper Series, Columbia Business
School (August).

Coburn, A.W., R.J.S. Spence, and A. Pomonis. 1991. Vulnerability and Risk Assessment.
New York: United Nations Development Program (UNDP) and Disaster Relief Or-
ganization (UNDRO).

76
Corporacion Andina de Fomento (CAF). 2002. http://www.caf.com/espanolold/
01b02e.asp.

Cuny, F.C. 1983. Disasters and Development. New York: Oxford University Press.

Davis, I., and S.P. Gupta. 1990. “Disaster Mitigation in Asia and the Pacific.” Report
prepared by the Asian Disaster Preparedness Center of the Asian Institute of
Technology.

Devarajan, S., M.J. Miller, and E.V. Swanson. 2000. “Development Goals: History, Pros-
pects and Costs.” Washington, D.C.: World Bank.

Doherty, N.H. 1997. “Financial Innovation for Financing and Hedging Catastrophe Risk.”
In Financial Risk Management for Natural Catastrophes, Neil R Britton and J.
Oliver, eds. Brisbane: Aon Group Australia Limited.

Economic Commission for Latin America and the Caribbean (ECLAC). 1998. “Honduras:
Evaluación de los daños ocasionados por el Huracán Mitch: Sus implicaciones pa-
ra el desarrollo económico y social y el medio ambiente.” ECLAC publication.

Elahi, Shirin. 2001. “The Great Hanshin Earthquake, Kobe, Japan.” Paper prepared for
the TSUNAMI Project on the Uninsured Elements of Natural Catastrophe Losses,
IIASA, Laxenburg, Austria.

EQE. 1995. The January 17, 1995 Kobe Earthquake. An EQE Summary Report.
http://www.eqe.com/publications/kobe/kobe.htm.

Evans, Arthur F. 1996. “Insurance, Reinsurance and Catastrophe Protection in the Car-
ibbean.” USAID-OAS Caribbean Disaster Mitigation Project, Organization of
American States, Unit for Sustainable Development and Environment. Washing-
ton, D.C.: OAS.

Federal Emergency Management Agency (FEMA). 2000. “Project Impact: Building Disas-
ter Resistant Communities.” Washington, D.C.: Federal Emergency Management
Agency.

Ferranti, D., G.E. Perry, and others. 2000. Securing Our Future in a Global Economy.
Washington, D.C., World Bank.

Fischer, Stanley, and W. Easterly. 1990. “The economics of the government budget
constraint.” The World Bank Research Observer 5(2):127-42.

Foster, H.D. 1980. Disaster Planning: The Preservation of Life and Property. New York:
Springer-Verlag.

Freeman, P.K., and H. Kunreuther. 1997. Managing Environmental Risk through Insur-
ance. AEI Press (soft back), Kluwer Academic Publishers (hardback).

Freeman, P.K., L. Martin, R. Mechler, and K. Warner. 2001. “Catastrophes and Devel-
opment: Integrating Natural Catastrophes into Development Planning.” Washing-
ton, D.C.: World Bank. www.proventionconsortium.org.

77
Freeman, Paul K., Landis Mechler, Leslie Martin, Reinhard Mechler, and Koko Warner.
2000. Integrating Natural Catastrophe Exposure into Development Planning:
Case Study of Natural Catastrophe Risk in Honduras. Laxenburg, Austria: Inter-
national Institute for Applied System Analysis.

Froot, K. 1999. The Financing of Catastrophe Risk. Chicago: University of Chicago

Press.

Garb, S., and E. Eng. 1969. Disaster Handbook. New York: Springer.

Gilber, C., and C. Gouy. 1998. “Flood Management in France,” In Flood Response and
Crisis Management in Western Europe: A Comparative Analysis. Uriel Rosenthal
and Paul Hart, eds., 15-54. Berlin: Springer.

Government of Fiji. 1985. “Budget 1986.”

_____. 1995. “Fiji National Disaster Management Plan.” Government of Fiji National
Disaster Management Council, Suva, Fiji.

Gulkan, P. 2001. “Rebuilding the Sea of Marmara Region: Recent Structural Revisions in
Turkey to Mitigate Disasters.” Presented at the Wharton-World Bank Conference
on Challenges in Managing Catastrophic Risks: Lessons for the U.S. and Emerg-
ing Economies, Washington, D.C.

Hays, W.W., ed. 1981. Facing Geologic and Hydrologic Hazards. Earth-Science Consid-
erations. Geological Survey Professional Paper 1240-B. Washington, D.C.: U.S.
Government Printing Office.

Heavily Indebted Poor Countries Debt Initiative (HIPC). 2002. World Bank initiative.
http://www.worldbank.org/ hipc/about/hipcbr/hipcbr.htm.

Hodgson, A. 1997. "The Pricing of Catastrophe Risk-Capital Market and Insurance Per-
spectives." In Financial Risk Management for Natural Disasters. N. R. Britton and
J. Oliver, eds. Brisbane: Aon Group Australia Limited.

Hoogeveen, J.G.M.H. 2000. “Risk and Insurance by the Poor in Developing Nations.” In
Managing Disaster Risk in Emerging Economies. A. Kreimer and M. Arnold, eds.
Washington, D.C.: World Bank.

Horváth, G., S. Kisgyörgy, J. Sendzimir, and A. Vári. 2001. “The 1998 Upper Tisza
Flood, Hungary: Case Study Report” From the IIASA study on Flood Risk Man-
agement Policy in the Upper Tisza Basin: An Integrated Approach. IIASA, Laxen-
burg, Austria.

Inter-American Development Bank (IDB). 2000a. “Colombia: Economic Situation and

Perspectives.” Washington, D.C.: IDB.

_____. 2000b. “Facing the Challenge of Natural Disasters in Latin American and the
Caribbean: An IDB Action Plan.” Washington, D.C.: Inter-American Development
Bank.

_____. 2002a. “Country Data for Dominican Republic.” Washington, D.C.: IDB.

78
_____. 2002b. “La Sostenibilidad de la Politica Fiscal en America Latina: El Caso de Co-
lombia.” Washington, D.C.: IDB.

International Federation of Red Cross and Red Crescent Societies (IFRC). 2001. World
Disasters Report: Focus on Recovery. Geneva: International Federation of Red
Cross and Red Crescent Societies.

International Monetary Fund and World Bank (IMF/WB). 2001a. “El Salvador: IMF and
World Bank Staff Assessment of the Macroeconomic Effects of the Earthquakes.”
http://www.iadb.org/exr/events/ElSalvad.pdf.

______. 2001b. “Bolivia: Completion Point Document for the Enhanced Heavily In-
debted Poor Countries (HIPC) Initiative. http://www.iadb.org/hipc/country-
cases/bolivia/Bolivia-CP.pdf.

Kreimer, Alcira, Margaret Arnold, Paul Freeman, and others. 1999. “Managing Disaster
Risk in Mexico – Market Incentives for Mitigation Investment.” Disaster Risk
Management Series. Washington, D.C.: World Bank.

Kreimer, A., and E. Echeverria. 1998. “Case Study: Housing Reconstruction in Mexico
City.” Disaster Management Facility. Washington, D.C.: World Bank.

Kunreuther, H., and R.J. Roth. 1998. Paying the Price: The Status and Role of Insurance
Against Natural Disasters in the United States. Washington, D.C.: Joseph Henry
Press.

Kuroiwa, J. 1991. “Microzonation: The key tool for formulation and implementing the
national programme for disaster prevention and mitigation.” In Regional Disaster
Management in the Latin American Region. Report and Summary of Proceedings
of the UNCRD-CISMID International Workshop on Regional Disaster Manage-
ment, Piura, Peru, 23-26 1991, 77-88. UNCRD Meeting Report Series No. 46.
Nagoya, Japan: United Nations Centre for Regional Development.

Lavell, A. 1989. “Desastres naturales y zonas de riesgo en Centroamerica: Condicionan-

tes y opciones de prevencion y mitigacion.” Estudios sociales centroamericanos
49: 79-85.

_______ 2001a. “Algunos Programas Innovadoras en torno al Problema de Riesgo y

Desastre en Centroamérica.” www.iiasa.ac.at/~warner/docs/Latin%20American
%20Links.htm.

_______ 2001b. “Disaster Risk Reduction Initiatives in Central America and the Domini-
can Republic: A Review of Recent Developments, 1997-2001.” Central American
Coordinating Centre for Natural Disaster Prevention (CEPREDENAC).

Lewis, C.M., and K.C. Murdock. 1999. “Alternative Means of Redistributing Catastrophic
Risk in a National Risk-Management System.” In The Financing of Catastrophe
Risk, edited by K.A. Froot, 51-85. Chicago: National Bureau of Economic Re-
search.

Linnerooth-Bayer, J., S. Quiano, R. Loefstedt, and S. Elahi. 2000. “The Uninsured

Losses of Natural Catastrophic Events, Seven Case Studies of Earthquake and
Flood Disasters.” Summary Report, IIASA, Laxenburg, Austria.

79
Mansilla, E. 1996. “Desastres modelo para armar: Coleccion de piezas de un rompeca-
bezas social.” La Red, Panama City.

Maskrey, A. 1993. “Los desastres no son naturales.” La Red, Panama City.

Michel-Kerjan, Erwann. 2001. “Insurance Against Natural Disaster: Do the French Have
the Answer?” Paper presented at the First Annual IIASA-DPRI Meeting on Inte-
grated Disaster Risk Management, IIASA, Laxenburg, Austria.

Mileti, D.S. 1999. Disasters by Design: A Reassessment of Natural Hazards in the

United States. Washington, D.C.: Joseph Henry Press.

Mileti, D.S., J.D. Darlington, E. Passerini, B. Forrest, and M.F. Myers. 1995. “Toward an
integration of natural hazards and sustainability.” The Environmental Profes-
sional 17: 117-26.

Mitchell, James K., ed. 1999. Crucibles of Hazard: Mega-cities and Disasters in Transi-
tion. Tokyo and New York: United Nations University Press.

Moody’s. 2002. http://www.moodys.com/moodys/cust/default.asp.

Morales, J.A., and H. Mueller. 1999. “Bolivia: Memorandum of Economic and Financial
Policies.” Washington, D.C.: International Monetary Fund.

Munich Reinsurance Company. 2001. Topics: Natural Disasters. Annual Review of Natu-
ral Disasters 2000. Munich: Munich Reinsurance Group.

Olson, R., J.P. Sarmiento Prieto, R. Olson, A. Estrada, and V. Gawronski. 2000. “The
Marginalization of Disaster Response Institutions: The 1997-1998 El Niño Ex-
perience in Peru, Bolivia, and Ecuador.” Natural Hazards Center Special Publica-
tion #36, University of Colorado at Boulder.

Olson, R., R. Alvarez, B. Baird, A. Estrada, V. Gawronski, and J.P. Sarmiento Prieto.
2001. “The Storms of ‘98: Hurricanes Georges and Mitch -- Impacts, Institu-
tional Response, and Disaster Politics in Three Countries.” Natural Hazards Cen-
ter Special Publication #38, University of Colorado at Boulder.

Organization for Economic Co-operation and Development (OECD). 1997. “Twenty

Guidelines For Insurance Regulation And Supervision In Emerging Economies, in
Insurance and Private Pensions Compendium for Emerging Economies.” Book 1.
Insurance Committee Secretariat. Washington, D.C.:OECD.

_____. 2001. “Net Official Development Assistance Flows from 1950 to 1999.” Paris:
OECD Development Assistance Committee. http://www.oecd.org/dac/xls/ long-
termODA.xls.

Organization of American States (OAS). 1999a. Caribbean Disaster Mitigation Project

Progress Bulletin “Promotion of Loss Reduction Incentives and Hazard Mitigation
in the Property Insurance Industry.” December. http://www.oas.org/en/cdmp/
bulletin/insur.htm.

_____ 1999b. “Final Evaluation of the Caribbean Disaster Mitigation Project (CDMP).”
http://www.oas.org/en/cdmp/document/ ofdaeval.htm.

80
Perry, Daniel. 2001. Argentina’s Plan for Dealing with Natural Disasters. Graduate
School of International Studies, University of Denver, Colorado.

Petak, W.J., and A.A. Atkisson. 1982. Natural Hazard Risk Assessment and Public Pol-
icy: Anticipating the Unexpected. New York: Springer-Verlag.

Pielke, Jr., R.A., J. Rubiera, C. Landsea, M. Molina, and R. Klein. 2001. “Hurricane Vul-
nerability in Latin America and the Caribbean.” Natural Hazards Review.

Platt, R.H. 1999. Disasters and democracy: the politics of extreme natural events.
Washington, D.C.: Island Press.

Pollner, J. 2000. “Catastrophe Risk Management Using Alternative Risk Financing & In-
surance Pooling Mechanisms: The Insurance Market and the Case of the Carib-
bean Region.” Washington, D.C.: World Bank.

Poncelet, J.L. 1997. “Disaster Management in the Caribbean.” Disasters 21(3): 267-79.

Red Cross and Red Crescent Society. 2001. World Disasters Report: Focus on Recovery.
Washington, D.C.: International Federation of Red Crescent Societies.

Red Cross, World Meteorological Organization, and Economic and Social Commission for
Asia and the Pacific (Red Cross, WMO, and ESCAP). 1977. Guidelines for Disaster
Prevention and Preparedness in Tropical Cyclone Areas. Published jointly by ES-
CAP, WMO, and the League of Red Cross and Red Crescent Societies, Ge-
neva/Bangkok.

Rokovada., J., and L. Vrolijks. 1993. “Case Study Fiji: Disaster and Development Link-
ages.” South Pacific Workshop, Apia, Western Samoa.

Sanderson, D. 2000. “Cities, disasters, and livelihoods.” Environment & Urbanization

12(2): 93-102.

Schilderman, T. 1993. “Disasters and Development: A Case Study from Peru.” Journal
of International Development 5(4).

Sen, A. 1999. Development as Freedom. New York: Alfred A. Knopf.

Silva, Carlos Daniel. 2001. “Instituciones públicas con competencia en materia de inun-
daciones.” In Inundaciones en el Area Metropolitana de Buenos Aires, Alcira
Kreimer, David Kullock, and Juan B. Valdés, eds. The World Bank Disaster Risk
Management Working Paper Series No. 3, Washington, D.C.

Smith, K. 1996. “Environmental Hazards: Assessing Risk and Reducing Disaster.”

Rutledge Physical Environment Series, University of Cambridge.

Swiss Reinsurance Company. 1997. “Too Little Reinsurance of Natural Disasters in Many
Markets.” Sigma No. 7/1997. Zurich: Swiss Reinsurance Company.

_____. 1999. “Alternative Risk Transfer (ART) for Corporations: A Passing Fashion or
Risk Management for the 21st Century?” Zurich: Swiss Reinsurance Company.

81
_____. 2000. “Emerging markets: the insurance industry in the face of globalisation.”
Sigma No. 4/2000. Zurich: Swiss Reinsurance Company.

_____. 2001. “World insurance in 2000: another boom year for life insurance; return to
normal growth for non-life insurance.” Sigma No. 6/2001. Zurich: Swiss Rein-
surance Company.

United Nations. 1984. Disaster Prevention and Mitigation: A Compendium of Current

Knowledge. New York: United Nations.

United Nations Department of Humanitarian Affairs (UNDHA, UNDP, and MWR). 1994.
United Nations Department of Humanitarian Affairs, United Nations Development
Program, and Ministry of Water Resources of the Socialist Republic of Viet Nam.
“Strategy and action plan for mitigating water disasters in Viet Nam.” New York
and Geneva: DHA (94/1).

United Nations Population Fund (UNFPA). 1999. The State of the World Population
1999. New York: UNPFA. http://www.unfpa.org/swp/1999.

Universidad de Los Andes. 1997. "Microzonificación Sísmica de Santafé de Bogotá." Bo-

gotá: Ingeominas.

Varangis, Panos, Jerry Skees, and Barry Barnett. 2002. “Weather Indexes for Develop-
ing Countries.” In Climate Risk and the Weather Market: Financial Risk Manage-
ment, edited by Weather Hedges and Robert S. Dischel. London: Risk Books.
http://www.iadb.org/idbamerica/Archive/stories/1999/eng/e899f4.htm.

Vári, A., Z. Ferencz, and J. Linnerooth-Bayer. 2001. “Flood Risk Management in the Up-
per Tisza Region: Results of a Public Questionnaire.” In Flood Risk Management
Policy in the Upper Tisza Basin: An Integrated Approach, IIASA, Laxenburg, Aus-
tria.

Vatsa, Krishna, and Frederick Krimgold. 2000. Financing Disaster Mitigation for the
Poor. In Alcira Kreimer and Margaret Arnold, eds., Managing Disaster Risk in
Emerging Economies. Washington, D.C.: World Bank (pp. 129-53).

Vermeiren, Jan. 2000. Risk Transfer and Finance Experience in the Caribbean. In Alcira
Kreimer and Margaret Arnold, eds., Managing Disaster Risk in Emerging Econo-
mies. Washington, D.C.: World Bank.

Vermeiren, J., and J. Pollner. 1994. “A Probable Maximum Loss Study of Critical Infra-
structure in Three Caribbean Island States.” Washington, D.C.: World Bank/OAS
Report. http://www.oas.org/en/cdmp/document/pml/pml.htm.

Wolfensohn, J.D., and S. Fischer. 2000. “The comprehensive development framework

(CDF) and poverty reduction strategy papers.” Washington, D.C.: World Bank.

World Bank. 1994. World Development Report 1994: Infrastructure for Development.
Washington, D.C.: World Bank.

_____. 2000. “Managing Economic Crises and Natural Disasters.” In World Develop-
ment Report 2000/2001: Attacking Poverty, 161-76. Washington, D.C.: Oxford
University Press.

82
_____. 2001a. World Bank Development Indicators. Washington, D.C.: World Bank.

_____. 2001b. Estratégia Boliviana de Reducción de la Pobreza. La Paz: World Bank.

_____. 2001c. World Development Report 2000/2001: Attacking Poverty. Washington,

D.C.: World Bank.

_____. 2002. Colombia Country Brief. http://www.worldbank.org.

World Bank and IMF. 1999. “Building poverty reduction strategies in developing coun-
tries.” DC/99-29. Washington, D.C.: World Bank and IMF.

83