/ -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~------ ~~~~~~~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ quup ~~~~~~ ~Q Disaster Risk Management Series Managing Disaster Risk in Mexico: Market Incentives for Mitigation Investment Alcira Kreimer Margaret Arnold Christopher Barham Paul Freeman Roy Gilbert Frederick Krimgold Rodney Lester John D. Pollner Tom Vogt The World Bank Disaster Management Facility 1999 Washington, D.C. Copyright ( 1999 The International Bank for Reconstruction and Development/THE WORLD BANK 1818 H Street, N.W Washington, D.C 20433, U.S.A. All rights reserved Manufactured in the United States of America First printing June 1999 The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors and should not be attributed in any manner to the World Bank, to its affiliated organizations, or to members of its Board of Executive Directors or the countries they represent. The World Bank does not guarantee the accuracy of the data included in this publication and accepts no responsibility for any consequence of their use. The boundaries, colors, denominations, and other information shown on any map in this volume do not imply on the part of the World Bank Group any judgment on the legal status of any territory or the endorsement or accep- tance of such boundaries. The material in this publication is copyrighted. The World Bank encourages dissemination of its work and will normally grant permission promptly Permission to photocopy items for internal or personal use, for the internal or personal use of specific clients, or for educational classroom use is granted by the World Bank, provided that the appropriate fee is paid directly to Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, U.S.A., telephone 978-750-8400, fax 978-750-4470. Please contact the Copyright Clearance Center before photocopying items. For permission to reprint individual articles or chapters, please fax your request with complete information to the Republication Department, Copyright Clearance Center, fax 978-750-4470. All other queries on rights and licenses should be addressed to the World Bank at the address above or faxed to 202-522-2422. Photo credits: Cover, Mexico City earthquake, 1985, photographer, Fred Krimgold; page 5, aerial view of Mexico City after 1985 earthquake, photographer, Fred Krimgold; page 8, damage from Hurricane Gilbert, Mexico, 1985, photographer, Alcira Kreimer; page 16, improved building standards for post-earthquake reconstruction, 1989, photographer, Fred Krimgold; page 24, Mexican family outside a reconstructed home, 1987, photographer, Alcira Kreimer; page 30, direct damage to infrastructure leads to economic and social disruption, 1994, photographer, Fred Krimgold; page 36, Post-1985 earthquake home reconstruction in Mexico, 1987, photographer, Alcira Kreimer. Cover design by Wendy Guyette of Communications Development. Library of Congress Cataloging-in-Publication Data has been applied for. ISBN: 0-8213-4491-9 Contents Acknowledgments v Abbreviations vi Executive Summary 1 Risk Identification 1 Risk Reduction 1 Risk Transfer 2 The Role of FONDEN 3 The Role of the World Bank 4 1. Introduction 6 2. Risk Identification: Assessing Hazards and Vulnerability 9 Hazard and Vulnerability Assessment 9 What Kind of Disasters Occur in Mexico? 10 How Are Hazards and Vulnerability Assessed in Mexico? 13 How Do Natural Disasters Affect Economic Development? 13 Recommendations 1 5 3. Mitigating Damage by Reducing the Level of Risk 17 Developing a Risk Reduction Strategy 17 Coordinating Risk Mitigation Efforts throughout the Country 17 What Steps Have Been Taken to Mitigate Risk? 19 Recommendations 2 1 The Role of the Fund for Natural Disasters 23 4. Mitigating Losses by Transferring Risk to the Insurance Industry 25 Insurance and Reinsurance Coverage throughout the World 25 Demand for Property and Casualty Insurance in Mexico 26 Supply of Property and Casualty Insurance in Mexico 27 Mitigation of Risk by Insurance Companies 28 Recommendations 28 iii iv Managing Disaster Risk in Mexico: Market Incentivesfor Mitigation Investmnent 5. Using New Financial Instruments to Transfer and Finance Risk 31 New Financial Instruments for Transferring Risk 31 Risk Transfer and the Fund for Natural Disasters 32 Using Alternative Risk Transfer Mechanisms to Deal with Uninsurable and Uninsured Properties 33 Recommendations 34 6. Summary of Findings and Recommendations 37 Identify Risk 37 Mitigate Risk 37 Transfer Risk 38 Reorganize FONDEN 39 The Role of the World Bank 39 Annexes 1. Geographic Distribution of Major Hazards in Mexico 41 2. Estimated Cost of Disaster Events in Mexico, 1980-98 42 3. Examples of Catastrophe Funds and Catastrophe Risk Pooling Used in Other Countries 48 4. List of People Interviewed for Study and Participants in Workshop 51 References 56 Tables 2.1 Forest fires in Mexico, 1993-98 13 2.2 Annual deaths and direct economic losses caused by disasters in Mexico, 1980-98 14 3.1 Responsibility for technical support to SINAPROC 18 3.2 Actions taken during various states of emergency 18 3.3 Emergency actions and technical coordinators 19 4.1 Estimate of insurance density 25 4.2 Insurance penetration and macroeconomic indicators in selected countries, 1996 26 4.3 Insurance penetration in selected developing countries, 1996 26 4.4 Reinsurance pricing of catastrophic risks in Mexico, 1994-98 27 4.5 Underwriting profit earned by offshore reinsurers, 1995-97 28 Maps 2.1 Path of Pacific Ocean hurricanes, August-November 1997 11 Al Active volcano risks and hurricane paths 41 Acknowledgments This report summarizes the findings of a World Bank insurers, reinsurers, banks and bank associations, and mission on disaster management, mitigation, transfer, regulators. It also met with experts in natural disaster and financing conducted in Mexico, October 19-28, management, including researchers working at the 1998. It incorporates the insights generated at a work- National Autonomous University of Mexico and at inter- shop on the preliminary findings of the mission held national agencies represented in Mexico City. The active in Mexico City, February 23-24, 1999. interest and support from the team's many interlocu- The report was prepared by a team led by Alcira tors is deeply appreciated. (A list of agencies and indi- Kreimer. Team members included Margaret Arnold, viduals met is provided in annex 4.) Christopher Barham, Rodney Lester, and John Pollner The staff of the World Bank's local office in Mexico (World Bank); Tom Vogt (International Finance Corpo- City provided tremendous support and assistance to ration); and Paul Freeman, Roy Gilbert, and Fred Krim- the team during the October mission and the Febru- gold (consultants). TomDavid and FranciscoJavierPerez, ary workshop. Special thanks are owed to Country from Impact Forecasting, L.L.C. (a subsidiary of the Aon Director Olivier Lafourcade; .Agriculture and Environ- Group), participated in the mission as private sectorpart- ment Sector LeaderAdolfo Brizzi; and to Mission Coor- ners in the Market Incentives for Mitigation Investment dinator Dora Posada, for their cooperation and support. (MIMI) initiative; we acknowledge their helpful inputs Cooperation and partial funding for the study were to the study. Natural disaster management experts Daniel provided by the govemment of the Netherlands, as part BitrTn Bitran, Ricardo Toledo, and Luis Enrique Gutier- of the World Bank/Netherlands Environment Program. rez Alcaraz also made valuable contributions to the study The study team thanks them for their generous support. The report was edited by Barbara Karni and laid out by The team also thanks the private sector partners of Suzanne Luft of Communications Development. the MIMI initiative-the Aon Group, Gerling Global The mission benefited from the close collaboration Reinsurance, Munich Reinsurance, and Renaissance of the Mexican President's Office; the Ministry of Finance Reinsurance, Ltd.-for their support of and collabora- and Public Credit; the Ministry of Environment, Natural tion on the MIMI initiative. Responsibility for the report's Resources and Fisheries; the Ministry of the Interior; and conclusions and recommendations (as well as any remain- the Federal District Government of Mexico City The ing errors) rests solely with the authors and should not study team would like to express its gratitude for their be attributed to any of the interviewees or supporters cooperation and support. The mission met with private of the initiative. -V Abbreviations CATEX Catastrophe Exchange CENAPRED National Center for Disaster Prevention (Centro Nacional de Prevenci6n de Desastres) CNA National Water Commission (Comisi6n Nacional del Agua) FONDEN Fund for Natural Disasters (Fondo para Desastres Naturales) LIBOR London interbank offer rate MIMI Market Incentives for Mitigation Investment NGO Nongovernmental organization PAHO Pan American Health Organization PEMEX Petroleos Mexicanos SEMARNAP Ministry of Environment, Natural Resources, and Fisheries (Secretaria de Medio Ambiente, Recursos Naturales y Pesca) SINAPROC National Civil Protection System (Sistema Nacional de Proteccion Civil) v,i Executive Summary Natural disasters have had a formidable impact on devel- identify the principal hazards it faces and assess the vul- opment in Mexico. Disasters have destroyed human, social, nerability of principal settled areas, infrastructure, and and physical capital, and they have derailed economic economic assets. The analysis should include esti- development, as funds are reallocated from ongoing pro- mates of the direct, indirect, and secondary impacts of grams to finance relief and reconstruction assistance. natural disasters. The consequences for development and economic growth Mexico has highly competent intellectual and insti- are ominous. tutional resources for hazard and vulnerability assess- The aim of this study is to assess the current capac- ment at the National Center for Disaster Prevention ity of Mexico to deal with disaster risk and to identify (CENAPRED) and the National Autonomous University ways in which the impacts of catastrophes on the of Mexico. CENAPRED is a unique institution that economy can be reduced. The study analyzes the three bridges the gap between academic researchers and the main components of a comprehensive disaster risk man- government by channeling research applications devel- agement strategy: risk identification, risk reduction, and oped by university researchers to the Ministry of the risk transfer and financing. Interior. Neither CENAPRED nor the National Mexico is vulnerable to a variety of natural disasters, Autonomous University receives the budgetary atten- including earthquakes, hurricanes, and volcanoes. Despite tion it deserves, however. the frequency with which these disasters strike, howev- er, inadequate investment is made in mitigation efforts, and insufficient funds are set aside to pay for relief and Risk Reduction reconstruction efforts. As a result, when a disaster occurs, the government is often forced to use funds that had been The damage caused by disasters can be reduced by allocated to other programs, disrupting the operations changing perceptions and behavior so that all mem- of those programs. The effect is to reduce growth and bers of society place a high priority on safety in plan- derail important development efforts. ning and development. All stakeholders must contribute The Mexican government could use mechanisms to to the creation of a culture of safety manage risk so that ongoing programs are not disrupt- Civil defense in Mexico (as in many other countries) ed following a disaster. Doing so involves identifying the has focused mainly on monitoring, preparedness, and risks the country faces, mitigating the damage caused by response to disasters. These e fforts are crucial to miti- those risks, and transferring the risk to other parties gating the effects of disasters. However, the core of a (namely, insurance companies and the capital markets). mitigation program should consist of more "upstream" measures, such as the safe lccation, design, and con- struction of structures, infrastructure, and settlements. Risk Identification Mexico has taken some important steps in this direc- tion. Scientific advisory committees have been estab- Formulation of an effective risk management strategy lished, standards for civil works have been set, engineering must begin with the identification of risk. Mexico should advances have been made, schools have been retrofitted 1 2 Managirg Disaster Risk in Mexico: Market Incertivesfor Mitigation Investment to withstand earthquakes, and a program for certifying part of the relief effort following a disaster could be hospitals that meet disaster readiness standards has been used to implement mitigation measures, such as strength- put into effect. Much more could be done, however, to ening buildings in earthquake- and wind-prone areas, reduce vulnerability to disasters in the long term. The improving irrigation facilities in drought-prone areas, government should develop strategies for- improving resettling communities away from hazardous sites, clear- education about disaster mitigation, creating incentives, ing stream beds and drainage channels, constructing and adopting regulations that will encourage individu- flood protection, and participating in reforestation. als and businesses to reduce the risks they face and pro- mote a culture of social protection. Active programs of dissemination, targeted profes- Risk Transfer sional education, and broad public education on disas- ter risk and mitigation should be developed and delivered The Mexican property and casualty (or "general") insur- throughout Mexico. Programs should include dissemi- ance market is underdeveloped, with the great major- nation of information on natural hazard risk, mitigation, ity of property owners uninsured against maj or natural and insurance (possibly through a collaborative effort catastrophes. The lack of property and casualty insur- by the govemment and the insurance industry); the inclu- ance coverage in Mexico appears to be caused largely sion of disaster preparedness and mitigation materials in by demand-side rather than supply-side factors. Although elementary and high school curricula; and special pro- some factors distort the supply of property and casual- grams to target low-income communities, which are often ty insurance in Mexico, excess catastrophic reinsurance worst hit by disasters. capacity is available and a potentially large distribution The regulatory approach to encouraging disaster mit- system has been developed. igation in Mexico requires a comprehensive reassessment In order to increase insurance awareness and pene- of the formal land use and building regulatory process- tration among small and medium-size enterprises and es. New approaches to land use management and build- the general population, affordable insurance products ing quality management must be developed that effectively must be available to the public. Capacity could be contribute to public safety in the informal sector. In recog- increased and prices reduced in several ways: nition of the fact that most mitigation decisions are * Catastrophic premium rates could be formally made at the community level, resources must be allo- deregulated, with a maximum tariff set at the current cated to increasing the capacity and authority for risk level. management and disaster mitigation at the state and com- * The discontinued 10 percent compulsory surcharge munity levels. on earthquake premiums could be reintroduced New initiatives should also focus on positive incen- and extended to all catastrophe classes (including tives, such as provision of community-level technical civil disturbances). These funds could be placed in assistance for disaster mitigation, establishment of a central pool managed by the Mexican Insurance "low-risk districts" or development of historic cen- Association, which would help stabilize reinsurance ters that include tax incentives for those that invest in costs and cover catastrophe-induced insolvency Com- mitigation, granting of financial assistance for engi- panies could continue to be required to accumulate neering analysis of risk, and establishment of certifi- catastrophic reserves up to an agreed-upon stan- cation programs. The private sector could consider dard retention level. donating engineering design assistance or providing * This industry-run pool could be required to offer it at low cost. catastrophe-only coverage, which would be distrib- Disaster relief and recovery assistance, distributed uted by insurance companies. through the government's Fund for Natural Disasters * In order to raise insurance awareness, the mandatory (FONDEN) can also play an important role in post- automobile liability insurance requirement, which disaster mitigation in areas in which disasters are recur- was postponed, could be implemented as soon as rent. Temporary employment programs established as possible. Executive SLrmnirary 3 . The insurance industry, in cooperation with the that have already been insured for noncatastrophic government, could be encouraged to educate the risks. Limit agricultural coverage to farmers who are public about insurance. unable to afford private insurance; farmers who are * The government could consider providing targeted able to afford private insurance should be obligated incentives to expand insurance density to purchase it. Subsume welfare payments to sub- In addition, private sector insurance companies could sistence farmers under the victims fund or route them be used to help Mexico deal more effectively with risk through other social safety net programs. in the public sector. The public sector should purchase * Reconstitute FONDEN as a legal entity that is capa- insurance through international brokers and risk ble of transferring risk. transfer specialists; individual agencies should not be * Inventory all vital infrastructure and estimate purchasing insurance on their own. More cost-effective FONDEN's risk exposure. Catastrophe modeling- coverage should be purchased by increasing the which analyzes the frequency of hazard events, the deductible level and purchasing higher levels of cata- severity of the events, the vulnerability of assets that strophe coverage. are the responsibility of FONDEN, and the value of those assets-could be used to make these estimates. The Role of FONDEN * Develop financial models that allow various policy options to be evaluated. Financial modeling based on Funds earmarked for relief and reconstruction after cat- the data developed in the catastrophe modeling astrophe events in Mexico are maintained in FONDEN, would yield various options and prices for the trans- which provides funds for the repair of uninsured infra- fer of risk. structure, immediate assistance to restore the produc- * Group similar risks together in funds so they can be tivity of subsistence farmers, and relief to low-income aggregated, priced, and transferred separately Allo- victims of disasters. Budgetary allocations for FONDEN cate sufficient dedicated capital and premium-paying have not been sufficient to cover the fund's obligations capacity to FONDEN to support its risk placements since 1996, and shortfalls are made up by diverting in the insurance or capital. markets. funds from other government budgets. The government * Modify FONDEN's mandate so that it is permitted to could explore several innovative techniques to make disburse funds for studies and training on mitigating more efficient use of the assets it allocates following dis- risk before a disaster occurs. Funding for studies and asters. These include purchasing reinsurance, making training would represent an addition to the annual use of new capital markets instruments, and using budgets of executing agerncies; the actual financing contingent financing. of mitigation works would presumably be carried out Before the government can make use of risk transfer as part of the public investment programs executed and risk-financing tools, it must modify the way FONDEN by the various government departments. is organized. Changes should include the following: * Consider linking the terms of FONDEN financing for * Refocus FONDEN's efforts by absorbing only those reconstruction following a disaster to the extent to residual catastrophic risks that cannot be absorbed which mitigation measures had been implemented by third parties, encouraging economically worth- before the disaster (possibly through varying the cost- while mitigation by the public sector, and fostering sharing arrangements with states and municipalities). the development of private sector risk reduction * Establish guidelines to be followed by other gov- and transfer mechanisms. ernment agencies in managing the risks they bear. * Clearly define the risks for which FONDEN bears FONDEN could, for example, specify the types and responsibility beginning with the infrastructure fund. acceptable limits for insurance coverage for various Limit coverage to those infrastructure items that are types of infrastructure. It could also help federal, essential to support the country's economic and social state, and local authorities purchase property insur- affairs following a natural disaster and those items ance on a group basis. 4 lManaging Disaster Risk in Mexico: Market Incentivesfor Mitigation Investment The Role of the World Bank providing contingent loans to finance a credit layer in a risk management package. The study has helped the World Bank define more clear- * Partial risk guarantees. The Bank operates a guaran- ly ways in which it could help developing countries tee program that supports private investment in manage disaster risk more effectively. Instruments that borrowing countries by guaranteeing against risks could be of value include the following: associated with policy or contractual performance * Loansfor investment in mitigation and reconstruction. by the government. Under the prograrr, the gov- Loans could be provided to finance retrofitting plans ernment counterguarantees the Bank guarantee by aimed at reducing losses (to schools, hospitals, pub- agreeing to repay the Bank for any payments it makes lic buildings) and the reconstruction of infrastructure to investors under the terms of the guarantee. In prin- damaged by natural disasters. An important feature ciple, this program might provide credit enhance- could be to link the terms of reconstruction financing ment for catastrophe bonds and commercial to the extent of mitigation carried out by the final contingent credit lines, thereby improving their terms beneficiaries of the financing. A technical assistance and attractiveness to investors. component could support capacity building for vari- In summary, there is much that Mexico can do, even ous aspects of disaster management. in the short term, to improve its ability to reduce and Contingent loans for reconstruction. To make sure that manage natural disaster risk. Leadership on lhazard and funds are accessible immediately when disaster strikes, vulnerability assessment, mitigation research and diffu- the Bank could establish a credit line that could be sion, insurance reform, and improved public sector drawn upon under certain conditions. Establishing risk management programs are among the key actions such a line would speed the delivery of funds during Mexico should take. This case study and future discus- emergencies. sions with the Mexican government should help refine * Contingent loansfor risk financing. Catastrophe risk these options so that they address Mexico's prmary con- is normally "packaged" into "layers" and financed cerns and contribute to more sustainable social and or transferred through a variety of arrangements, economic development. including primary insurance, reinsurance, credit financing, and bond and derivative financing.' Note The combination of instruments used depends on 1. A "layer" constitutes a commitment to cover losses in excess specific circumstances, including the relative costs of an agreed-upon minimum value and up to an agreed-upon of the various instruments. The Bank could consider maximum value. : WI SlI 'l .. SS ^d w g _ I _S ^ - S 9S | F"", :¢ t;t §|. -1!' i' '' Chapter 1 Introduction The World Bank established the Disaster Management natural disasters and because it is considering signif- Facility in July 1998 to provide proactive leadership in icant public policy changes to its insurance regula- coordinating efforts to introduce disaster prevention and tions. The mission to Mexico, carried out in October mitigation practices in the range of development-related 1998 at the request of the Mexican government, found activities. One of the facilitys key activities is the Market that Mexico can take steps to reduce natural disaster Incentives for Mitigation Investment (MIMI) initiative. losses through mitigation and to finance losses more MIMI seeks to create public-private partnerships to advance effectively when they do occur. This report synthesizes disaster risk awareness and to provide effective incen- the findings of the October 1998 mission and incor- tives for institutional and individual investment in disas- porates the comments and proceedings of a work- termitigationmeasures. Theinitiative isimportantbecause shop held in Mexico City, February 23-24, 1999. increasing World Bank and member country capacity to The specific objective of the study is to assess the mitigate risks would save development resources and pro- current capacity of Mexico to deal with disaster risk tect future investments. Privatization of disaster risk, where and to identify ways in which the impacts of catastro- applicable, would allow governments to shift funding from phes on the economy can be reduced. The study ana- emergency relief and reconstruction activities to more lyzes the three main components of a comprehensive effective and sustainable disaster mitigation investment, disaster risk management strategy: risk identification, Partnerships are key to the MIMI initiative. As the man- risk reduction, and risk transfer and financing. ager of the MIMI initiative, the Disaster Management Chapter 2 examines risk identification, the first step Facility has mobilized team members from vanous parts of comprehensive risk management. Risk identification of the World Bank Group, including regional country is the process of assessing potential losses clue to nat- departments, the Financial Sector Development Depart- ural disasters. Estimation of potential losses entails haz- ment, and the International Finance Corporation. Team ard assessment andvulnerability assessment. The chapter members also include academic institutions (Virginia includes recommendations for expansion and improve- Polytechnic Institute and the International Institute for ment of natural disaster risk assessment in Mexico. Applied Systems Analysis) and private sector insurers Chapter 3 examines risk reduction, which ir.cludes the (the Aon Group, Gerling Global Reinsurance, Munich physical measures to reduce exposure or increase Reinsurance, and Renaissance Reinsurance, Ltd.), resistance to natural hazards. Disaster mitigation mea- which provide valuable expertise, information, and sup- sures are those that reduce the physical impacts of dis- port to the team's research and policy development work. asters. The chapter describes current disaster mitigation The linchpin of the MIMI initiative is its country case activities in Mexico and suggests ways in which the pub- studies. These studies document current disaster expo- lic and private sectors can provide incentives for effec- sure, insurance coverage, and the use of mitigation prac- tive risk reduction. Chapter 4 examines how Mexico tices and assess the potential for mobilizing mitigation could transfer risk to the insurance industry It describes investment through the insurance and financial markets. the role of insurance and reinsurance and explains Mexico was chosen for the first appraisal mission why insurance coverage in Mexico is low. It proposes because of its extensive experience with losses from a series of steps that could help Mexico use private insur- 6 Introduction 7 ance companies to improve the public sector's ability It also looks at the structure of FONDEN, making rec- to deal with risk. Chapter 5 examines how Mexico might ommendations on ways the fund could be reorganized. be able to use new financial instruments, such as cata- Chapter 6 summarizes the study's main findings and strophe bonds and options to transfer or finance risk. recommendations. I ,4 :'l 2 mgE |~~~~~~~~~~~~~~~~~~~~~~~~~~~i~j Chaptei 2 Risk Identification: Assessing Hazards and Vulnerability In many countries, including Mexico, devastating Hazard and Vulnerability Assessment natural disasters are a frequent occurrence. A natur- al disaster occurs when an extreme event (flood, earth- The process of disaster risk management begins with quake, drought) affects vulnerable structures, such risk identification. Risk identification includes hazard as buildings and infrastructure, and the consequences assessment and vulnerability assessment. These activi- surpass a society's ability to respond. Extreme nat- ties are essential for the definLtion of strategies to man- ural events such as fires, floods, earthquakes, and age disaster risk, including risk reduction (mitigation droughts, have always been part of the natural of the impact of disasters) and estimation of potential cycle; virtually all parts of the world are exposed to losses necessary for financing or transferring risk. Haz- them. ard assessment requires scientific understanding of But accelerated changes in demographic and eco- relevant natural phenomena and interpretation of his- nomic trends in Mexico, as in other countries, have torical records of the occurrence of extreme events. disturbed the balance between ecosystems, increas- Hazard assessment provides the basis for the identi- ing the risk of human suffering and death. Rapid pop- fication of hazard zones, which can be presented on ulation growth increases pressures on natural resources maps at various scales. Such maps may indicate the and the environment and raises the consequent risks expected peak intensity of an event (as is done on earth- associated with human activities (Kreimer and Munas- quake zone maps) and the frequency of occurrence in inghe 1991). Major disasters not only damage capi- a particular area (as is done on floodplain maps). See tal assets but have long-term effects on the economy annex 1 for a map of some of the risks facing Mexico. as well. Vulnerability assessment focuses on the targets of nat- Despite the repeated experience of disasters, pub- ural hazards. It involves the evaluation of expected lic officials often react with surprise at the results of performance of structures, infrastructure, and institu- an "act of God." Awareness of risk to extreme natur- tions under the load exerted by extreme natural al events is not well established in the process of events. Unreinforced masonry buildings, for example, public and private investment decisionmaking. In- are known to perform poorly in particular types of earth- adequate provision of disaster response and recon- quakes. Buildings with poorly secured roofs are known struction funds before disasters often leads to ad hoc to suffer damage in hurricanes. More sophisticated dif- fund transfers, disrupting and postponing planned ferentiation of expected performance is possible for a development activities. Such diversion of develop- range of classes of structures and infrastructure. ment funding postpones progress toward long-term When hazard assessment and vulnerability assess- economic and social improvement. ment are combined, it is possible to develop estimates While the precise scale, time, or location of disaster of potential losses. With established probabilities of nat- events cannot be predicted, areas of hazard and vulnera- ural event occurrence and expected structural perfor- bility can be identified. Aggregate losses can be antici- mance it is possible to attach estimates of probability pated at the national level, and provision can be made for to specific loss scenarios. Modeling of expected disas- extreme events. ter losses represents a valuable tool for public and private 9 10Q Managing Disaster Risk in Mexico: Market Incentivesfor Mitigation Investment decisionmakers. Investors and insurance companies can and 1998 Mexico experienced 84 major earthquakes evaluate the security of their investments and the extent (earthquakes measuring more than 7.0 on the Richter of their exposure to disaster risk. Governments can use scale). such models to plan for development, evaluate options To estimate the potential for catastrophic loss for mitigation or risk reduction investments, and plan from earthquakes in Mexico, the country's seismici- for response needs before a disaster occurs. ty needs to be analyzed in conjunction with patterns The foundation of hazard assessment is the histori- of human and industrial habitation in disaster-prone cal record of extreme events, including their location, areas. While significant research has focused on earth- time, and intensity. The interpretation of this record in quake prediction, it is difficult to estimate earthquake light of scientific understanding of the particular phe- frequency. nomenon provides the basis for hazard mapping and Earthquakes occur frequently in Mexico because the estimation of event frequency. country sits atop four tectonic plates, and movement between the plates increases the risk of earthquakes. Epi- centers of earthquakes stronger than 7.0 on the Richter What Kind of Disasters Occur in Mexico? scale are located on the Pacific coast of the states of Jalisco, Colima, MichoacTn, Guerrero, and Oaxaca. The Mexico has a long history of natural disasters. Earthquakes most active seismic region in Mexico is the Guerrero and floods figure prominently in Mexican tradition and coast, where scientists believe a major seismic event is folklore. According to one legend, the modem age began long overdue. The Guerrero coast is a highly developed after a powerful earthquake destroyed the area's original area, with a significant concentration of population, struc- inhabitants, the Quinametin giants. Another legend claims tures, and infrastructure. The loss potential from a seis- that Quetzalcbatl sent a hurricane to destroy the city of mic event in the region is thus very high. Cholula. The major earthquakes of 1908, 1909, 1911, 1987, Between 1980 and 1998 Mexico experienced 79 dis- and 1988 occurred in the northeast section of the "seis- asters. These events included weather-related events (hur- mic gap." Most earthquake damage has occurred in ricanes, storms, floods, droughts, and forest fires); the regions between Jalisco, Chiapas, and the Isth- geology-related events (earthquakes, volcanic eruptions, mus of Tehantepac. Major earthquakes have also and landslides); and events caused by human action (indus- occurred in Veracruz, Puebla, the State of Mexico, trial accidents, chemical and oil spills and explosions, and and Baja California, especially near the border with structural fires). More than half of the events were weath- the United States. er related, with hurricanes and flooding as the leading The complexity of seismic events is illustrated by causes of damage. the characteristics of the earthquake that struck Much work has already been done in Mexico. Risk Mexico City in September 1985, an event that took atlases, for example, identify vulnerable areas within more than 6000 lives and damaged the housing of cities. Such tools can help policymakers develop effec- about 180,000 families. The ground movement res- tive strategies for reducing or transferring risks in those onant cycle coincided with the natural vibration peri- areas. No comprehensive analysis of vulnerability exists, od movement of the 5- to-12-story buildings that however. Some sense of Mexico's vulnerability can be make up the city's dense historic center, making this gained from examining the frequency and magnitude earthquake the most destructive in the hemisphere's of various kinds of disasters. history Poorly built tenements housing low-income families in overcrowded conditions suffered the worst damage. Also affected were 340 office buildings, in which 145,000 government workers were employed; Mexico is a seismically active country It is located along 1,200 small industrial shops; 1,700 hotel rooms; 1,200 the world's "fire belt," where 80 percent of the earth's schools; and 2,000 hospital beds. The economic loss- seismic and volcanic activity takes place. Between 1900 es exceeded $4 billion. Rish Iderntiftcuitton: Assessing Hazards and Vu1ncrabiltty I I Tropical Storms and Hurricanes Campeche portion of the Gulf of Mexico, the Eastern Mexico is one of the countries most severely affected Caribbean, and the tropical Atlantic Ocean. The Gulf by tropical storms. It is one of the few regions in the of Tehuantepec becomes turbulent around the end of world that can be affected simultaneously by two inde- May Most hurricanes head out to sea, but some make pendent cyclone regions, the North Atlantic and the landfall. The Sonda de Campeche portion of the Gulf North Pacific. The frequency of hurricanes would jus- of Mexico becomes turbulent in June. Winds head tify a well-organized system to deal with this hazard, north, affecting Veracruz and Tamaulipas. Hurricanes something that is not in place (Rosengaus 1998). of great intensity occur in the Eastern Caribbean between Between 1980 and 1998 Mexico suffered 43 weather- July and October, affecting Yucatan. The Atlantic Ocean related disaster events. That period may have been between 80 and 120 latitude north becomes active in atypical, however. During the 10 years between 1967 August. Winds head westward, penetrating the and 1977, Mexico suffered 157 severe weather events, Caribbean Sea, YucatAn, Tamaulipas, and Veracruz. including 57 hurricanes that caused serious damage in The precise location of these four regions is unstable six states. Many parts of Mexico are affected by cyclones and can vary as a result of ocean currents and their and tropical storms during the summer months (June temperatures. to October). During the winter, nontropical storms can The states of Baja California Sur, Michoacan, Sinaloa, lead to heavy rainfall in some parts of the country. Sonora, and Tamaulipas have been worst hit by hurri- Other parts of the country are frequently affected by long canes, with hurricanes making landfall on average every periods of drought, leading to the loss of cattle and crops. two to four years (map 2.1). About 40 percent of the The hurricanes affecting Mexico come from four population, or some 4 million people, live along the main regions: the Gulf of Tehuantepec, the Sonda de coast in these regions and are exposed to hurricanes. Map 2.1 Path of Pacific Ocean hurricanes, August-November 1997 I a Nora Ignacio Tijuana Linda Canc(rn Kevin mw Hlilda _CMarty %" ~ Jinena 12 Managing Disaster Risk in Mexico: Market Incentives for Mitigation Investment Hurricanes can be devastating. In 1997 Hurricane volcanic ash is irritating to the lungs and, eyes, and Pauline claimed hundreds of victims, caused flooding people are advised to take shelter until the air clears. and landslides, and cut electricity, communications, and Falling ash can be dangerous for air traffic, which should water. Acapulco, Mexico's most famous tourist desti- be suspended during volcanic eruptions. Ash can nation, was badly hit when homes in squatter settle- damage motors, the moving parts of machines, and ments on hillsides were destroyed, leaving many people infrastructure as well, blocking drains and causing elec- homeless. Pauline had such devastating effects because tricity networks to short circuit. Excessive accumula- it followed the coastline of Oaxaca and Guerreo, whip- tion of volcanic ash on rooftops can weaken structures. ping up strong waves and currents. The region's uneven topography also made it vulnerable to flooding and landslides following the hurricane. Rapid urbanization in coastal areas in recent years has increased the poten- Flooding caused by rivers overflowing their b anks occurs tial damage that hurricanes like Pauline can cause to almost every year in Mexico. Heavy rainfall also causes large population centers. erosion and landslides. Flooding is naturally less fre- quent in semi-arid regions, but it can be particularly dev- Volcanoes astating when it does occur there (as it did in Monterrey in 1988, for example). Mexico has dozens of volcanoes, 14 of which have erupt- Floods are rarely caused by heavy' rains alone. Soil ed in recent history. Volcanic eruptions can produce ash- erosion caused by deforestation, inadequate agricultural fall, tidal waves, lava and mud flows, and volcanic practices, and increasing urbanization all increase the risk earthquakes. These phenomena vary from one volcano of flooding. Despite considerable investment in drainage to another. infrastructure, Mexico City experiences losses every year The most recent volcanic eruption that resulted in from flash flooding. The occurrence of flooding is increas- catastrophic losses in Mexico was el Chichon in 1982, ing in flood plain areas that have been urbanizecd, as changes which completely destroyed eight communities and in land use broaden the capture of rainfall, producing claimed 2,000 lives. The eruption lasted nearly sixhours flows that the natural basin cannot cope with. Such and caused $117 million worth of losses to crops, cat- urban flooding affects assets of considerable economic tle, and damage to cultivable land within a 50-kilometer value. radius (Gutierrez 1998). The loss from flooding in Mexico has been high. Between 1993 and 1998 Popocatepetl, el Colima, el Between 1973 and 1990 more than 1,800 people died Tacana, and el Everman were active, with all but el Ever- and about $7.7 billion in economic losses were incurred man threatening settled areas. A major eruption of (Gutierrez 1998). Popocatepetl would affect settlements within a 20-30 kilometer radius, according to researchers at the Nation- al Autonomous University of Mexico. The volcano is being Forest Fires carefully monitored to allow for preparedness measures, Mexico experiences forest fires every year. The magnitude including evacuation of the communities in the vulner- of these fires depends in part on climatic conditions. In able area. 1998, for instance, excessively dry conditions, strong In major cities-such as Puebla, Cuernavaca, and winds, and high temperatures resulting from the El Nifno Mexico City, where the closest volcano is 43 kilo- phenomenon increased the number and severity of for- meters away-the effects of a volcanic eruption are like- est fires. Damage from forest fires was particularly severe ly to be limited to falling volcanic ash. Falling ash lasting in 1998, when about 850,000 hectares, includ-ng 198,000 for many hours can darken the sky and lead to electri- of forested land, were destroyed (table 2.1). WVorst affect- cal storms that can interfere with telecommunications. ed were the states of Chihuahua, Oaxaca, and Durango. Day suddenly turning to night can also have negative Slash and burn agriculture was responsible for more psychological effects upon inhabitants. While not toxic, than half of all forest fires in Mexico in 1998, according Risk Identification: Asscssing Hazxards and Vu[nerability 13 Table 2.1 Forest fires in Mexico, 1993-98 Directorate of Civil Protection of the Ministry of the Inte- 1ear Number ofnres Hectacrcs affctctd rior represents a major contribution to natural hazards 1993 10,251 235,020 data collection and analysis. CENAPRED was established 1904 7,830 141,502 with financial and technical assistance from the Japan- 1995 7,860 309,097 ese government. It is located on the campus of the Nation- 1996 95216 2481765 al Autonomous University of M\exico and benefits from 1998 14,445 849.632 close cooperation with the university Sotrce. Ministrv of Environment, Natural Resources, and Fisheries 1998 CENAPRED is a unique institution that bridges the gap between academic researchers and the govern- to the Ministry of Environment, Natural Resources and ment by channeling research applications developed by Fisheries. Natural causes accounted for only 3 percent university researchers to the Ministry of the Interior. It of all forest fires. is involved in both research and information dissemi- nation. Its collaboration with academic researchers and practical application of scientific knowledge rep- How are Hazards and Vulnerability Assessed resent a model that could be adopted by other disaster- in Mexico? prone countries. Risk identification is composed of hazard assessment and vulnerability assessment. Vulnerability is generally a How Do Natural Disasters Affect Economic function of location or construction. Assessing vulnera- Development? bility compares the resistive capacity or strength of a structure to the expected natural hazard loads associated An important component of vulnerability assessment is with the structure's location. Buildings and infrastructure the analysis of disaster damage in order to determine the are classified by structural type and material. Based on resistive capacity or strength of a structure to the laboratory tests and actual disaster damage experience, impacts of a particular natural hazard. Understanding the engineers are able to estimate the expected performance full economic impacts of natural disasters is also essen- of structures subjected to such factors as ground shaking tial to developing disaster risk management policy intensityorwindspeed. Surveyandevaluationofthestock Over the past 25 years, a growing body of research of buildings and infrastructure can provide estimates of has analyzed the economic impacts of disasters in potential damage and can help to identify weaknesses in developing countries (Benson 1997; Benson and criticalstructuresorsystems. The vulnerabilityassessment Clay 1998). The United Nations Economic Council provides the basis for risk reduction activities. Once vul- for Latin America and the Caribbean (ECLAC) has car- nerable structures and infrastructure are identified, strate- ried out systematic studies of the short- and long-term gies can be developed to strengthen or replace them. economic consequences of disasters. ECLAC's method- Mexico is fortunate to have highly competent intel- ology distinguishes three types of losses: direct loss- lectual and institutional resources for hazard and vul- es, indirect losses, and secondary effects. While some nerability assessment. For hurricane, flood, coastal work has been done in Mexico on direct losses, very flooding, and drought, hazard assessment is carried out little information is available on indirect and secondary by the National Water Commission. For earthquake, losses. volcano, and landslide hazards, assessment is handled by the National Center for Disaster Prevention (CENAPRED) and units of the National Autonomous Direct and Indirect Losses University of Mexico. Direct losses include losses of physical assets owned Following the 1985 Mexico City earthquake signifi- by individuals, businesses, and the government. They cant progress was made in identifying natural disaster include damage to structures and infrastructure as well risk. The establishment of CENAPRED in the General as loss of inventory and agricultural produce. 14 Managing DisaLster Risk in Mexico: Market Incentivesfor Mlitigation Investment Indirect losses are losses caused by the disruption of 1. Ad hoc reallocation of expenditurefollowing a disas- resource allocation across sectors and within markets. ter can jeopardize long-term development goals. Emer- These losses include the forgone production of goods gency reconstruction requirements in the immediate and services as a result of interruptions in utility ser- aftermath of a natural disaster often demand a sub- vices, transport, labor supply, suppliers, or markets. Indi- stantial reallocation of resources and government expen- rect losses also include the cost of providing interim diture in the short run. In Mexico funds that had been services while the original operating capacity is being committed to other programs have been reallocated to restored. provide disaster assistance. Funds supported by World Most estimates of disaster loss in Mexico have exam- Bank lending have also been diverted to deal with emer- ined onlythe direct losses oflife and property (table2.2). gency needs. As much as 30 percent of WoVorld While direct damage is an important measure of Bank-financed funds earmarked for water sector pro- impact, it does not capture the full impact of the event jects in Mexico, for example, are believed to have on the economy been reallocated to disaster-related programs. Even if funds are well spent on disaster-related operations, a sudden unplanned reallocation can undernine, if not Secondary Economic Effects sacrifice altogether, the longer-term development objec- Secondary effects include impacts on such macroeco- tive of the project that was bled of its resources. In agri- nomic variables as economic growth, the balance of culture, for instance, production targets may not be payments, public spending, and inflation. Estimating met. Indivisible infrastructure projects-such as power these costs is difficult because of data limitations.' These plants, highways, and ports-may not be built at all. effects can hurt long-term economic development in Failure to fund these projects reduces economnic growth. several ways: 2. The need to respond to natural disasters may under- minefinancial planning and budgeting as an instrument of Table 2.2 Annual deaths and direct economic losses economic and social development. In countries prone to caused by disasters in Mexico, 1980-98 natural disasters, ad hoc reassignment of government Direct economic !osses funds inevitably becomes frequent, if not routine, unless Yea Deathls (milior,s of u.s dollars) efforts are made to mitigate the effects of such disas- 1980 3 310.4 ters. If budgetary allocations are regularly pillaged to 1982 50 314.0 deal with natural disasters, confidence in the govern- 1984' 1.000 26.3 ment's commitment to loncg-term economic and social l985 6,043 4,159.91 D 1986 0 1.5 development and its ability to pursue such goals will 1987 6 0.3 be undermined. Unpredictable budget commitments 1988 692 2,092.9 are also likely to undermine efforts by state and munic- 1989 0 648.0 1990 391 94 5ipal governments that contribute to long-term economic 1991 11 167.5 development. Disasters are normal events in Mexico, 1992 276 192.5 occurring an average of three times a year between 1980 1993 28 125.6 and 1998. The government thus needs to take these 19954 364 6896 events into account in its development planning. Dis- 1996 224 5.3 aster mitigation and risk transfer measures can help 1997 228 447.8 the government restore the sustainability of its sup- 1998 199 2,478.8 port-through the budget-for economic and social Total 9515 11, 758.7 development. Note: Figures for 1981 and 1983 were not available. Figures are based on data available on particular events and do not represent a comprehensive 3. Expenditures on disaster recovery can cause budget economic assessment of all direct losses. Figures in U.S. dollars are based deficits. Reallocation of budgetary assignments to disaster- on conversion rates in the vear the losses were incurred. r r * ~~~~~~~~relatedl reconstruction may not be enough to meet all a. Includes $515 million in indirect losses. Soturce: Bitran Bitran 1999. emergency needs; additional funds may be needed to Risk fderntifhcation: Assesswng Hazards arnd Vulnerabilkty 15 meet the needs of people affected by the crisis. The dif- Recommendations ficulty of funding such needs may be exacerbated by the fact that public sector revenues may be interrupt- Before it can develop an effective risk management strat- ed by the disaster. Pressures to meet such needs can egy, Mexico must identify the principal hazards it thus increase public sector deficits, possibly driving up faces and assess the vulnerability to those hazards of interest rates and the rate of inflation. Such deficits its major settled areas, its infrastructure, and its eco- undermine international investors' confidence in the nomic assets. A good starting place would be a com- continuity of government policy, reducing both domes- plete inventory of federal government assets. The risk tic and foreign investment. Measures to mitigate the assessment should include estimates of direct, indirect, effects of natural disasters can help attenuate such effects. and secondary losses from disasters in order to provide 4. Natural disasters can adversely affect the balance of a better understanding of the most appropriate policy payments. Natural disasters are likely to hurt a coun- options. try's balance of payments, particularly if the area hit is Efforts already under way at CENAPRED and the a major exporting region. Developing models that clas- National Autonomous University of Mexico should receive sify regions in terms of their effects on the balance of additional support so that these institutions can con- payments could help governments pinpoint where dis- tinue to operate as centers of excellence in natural haz- aster mitigation efforts could be most beneficial in terms ard research. The valuable role CENAPRED plays in of preventing deterioration of the trade deficit.2 identifying earthquake and volcano hazards should be 5. Investors are likely to demand higher rates of return expanded to include the other principal hazards affect- following a disaster to compensate them for higher per- ing Mexico. ceived risks. Changes in the perception of risk can adverse- ly affect an economy if investors demand higher rates of Notes return. Entrepreneurs who do not believe they are 1. Given the enormous size of Mvlexico's economy-GDP was being adequately compensated for the risk they face may $335 billion in 1997-most studies have been unable to detect decide to close their businesses or move to a less risky the effects of disasters on macroeconomic variables. The sole location. If a business closes, the national economy suf- exception is the 1985 Mexico City earthquake, which adverse- fers a loss of income; if a business relocates within the ly affected macroeconomic variables in 1986. country, income is merely transferred from one region 2. Some attempts have been made to estimate the effect of nat- to another. Both cases have negative impacts on the local ural disasters in Mexico on the balance of payments. According economy and could serve as an inducement for local to one estimate, the 1985 earthquake m Mexico City increased governments at the state and municipal level to take mea- Mexico's balance of payments defcit by S8.6 billion in the five sures to mitigate the effects of natural disasters. years following the quake (Jovel undated). Chapter 3 Mitigating Damage by Reducing the Level of Risk Natural disasters result when an extreme natural event or irrigation systems. Resistance seeks to reduce the coincides with a vulnerable human settlement. Put sim- damage caused by hazards by constructing settlements ply, disasters are a product of hazard and vulnerability: that can withstand their effects. Both strategies are essen- tial to reducing disaster risk. Avoidance strategies should Disaster = hazard x vulnerability be used in land use and development planning. Resis- tance strategies should be used in building safe struc- The risk of disaster can be reduced by reducing either tures. the hazard factor or the vulnerability factor (or both). The first priority in natural disaster vulnerability In the case of some hazards, such as floods and land- reduction must be to ensure that new structures, infra- slides, engineering technologies can reduce the effect structure, and settlements are structurally sound and of a hazard. For other hazards, such as earthquakes located in areas that are not vulnerable to hazards. Doing and hurricanes, no effective means are available to mod- so will require clear delineation of hazard zones and ify the intensity, frequency, or location of a hazard event. the establishment of constnrction standards that cor- In the short term, real time hazard monitoring can pro- respond to expected hazard exposure. More problem- vide useful information that can help communities atic are existing structures, which may be difficult to prepare for disasters, mobilize resources, and evacuate relocate and expensive or impossible to reinforce. populations. Such monitoring can reduce deaths caused The estimation of potential disaster losses provides a by flooding, hurricanes, and volcanoes. useful reference for evaluating the cost and benefits of Civil defense in Mexico (as in many other countries) mitigation expenditures. Building codes, for example, has focused mainly on monitoring, preparedness, and can be analyzed to compare the added costs of con- response to disasters. While effective preparedness struction, supervision, and regulation with the poten- and response strategies are crucial to mitigating the tial savings from reducing the expected damage from effects of disasters, the core of a mitigation program future extreme events. should be more "upstream' measures, such as the safe location, design, and construction of structures, infra- structure, and settlements. Coordinating Risk Mitigation Efforts throughout the Country Developing a Risk Reduction Strategy The Mexican government established the National Civil Protection System (SINAPROC) in 1986 as the main The most important means of reducing disaster risk is mechanism for interagency coordination of disaster reducing vulnerability Two basic strategies are used: efforts. SINAPROC is responsible for mitigating the loss avoidance and resistance. Avoidance seeks to reduce of lives and material and the interruption of essential the effects of hazards on settlements by banning society functions caused by disasters. Responsibility for building in hazard zones or modifying the pattern of the system lies with the General Coordinating Body for occurrence of the hazard with structures such as dams Civil Protection in the Ministry of the Interior. The 17 18 MVlanaging Disaster Risk in Mexico: Market Incentivesfor Mitigation Investment system synchronizes the technical work of various min- SINAPROCs Program for Civil Protection 1995-2000 istries (table 3.1). is set up as a national regulatory and operational instru- In 1990 the National Council for Civil Protection (Con- ment. The program defines the prevention anc response sejo Nacional de Protecci6n Civil) was added to SINAPROC. actions required to deal with natural disasters. It also The council is an advisory, planning, and coordinating defines the authority, responsibilities, and relationships committee for civil protection. It is headed by the Presi- among different offices and organizations. The program dent of Mexico and made up of 12 ministers and the mayor covers prevention, rescue, and recovery of the federal district of Mexico City SINAPROC issues public wamings based on- the level At the subnational levels, governors of states and of emergency response required (table 3.2). It then presidents of municipalities are responsible for imple- coordinates responses by different ministries (table 3.3). menting and coordinating local Civil Protection Sys- Once an emergency is over, responsibility for recon- tems. These State and Municipal Councils for Civil struction reverts to the ministry responsible for the sec- Protection are responsible for planning, prevention, tor affected. and implementation of recovery actions after natur- To help prepare for and mitigate the effects of dif- al disasters. These offices provide requested support ferent kinds of potential disasters, SINAPRCC has set to state and municipal governments. Each state and up mechanisms for monitoring disasters. Th-is moni- municipal council also has its own technical unit for toring, which covers earthquakes, cyclones, and volca- civil defense. The autonomy of these subnational coun- noes, is performed by various institutions: cils sometimes conflicts with the effective integra- Earthquakes. The National Broad Band Seismology tion and coordination of national efforts by SINAPROC. Network was designed by the Engineering Institute of the National Autonomous University of Mexico Table 3.1 Responsibility for technical support to monitor earthquake activity in Mexico. A new to SINAPROC network has also been developed to improve the Iype oft !IaZad Tecd.hnical coordiiator understanding of the earthquake hazard in Mexico. G(eologic Ministry of Social Development Cyclones. Working with regional hurricane centers Hvdro-me:eorolog cal Ministry of Environmental, in Miami and San Francisco, Mexican authorities Natural Resources, and analyze tropical cyclones in order to help forecast Fisheries the movement and intensity of hurricanes head- CheLnical Ministry of Trade and Industrial ing toward Mexico. These forecasts are made avail- Development able to the Ministry of the Interior's Civil Defense Healtli Ministry of Environmental, Directorate and state representatives of tlle (nation- Natural Resources, and al) Ministry of Agriculture and Water Resources. Fisheries Mexico also monitors tropical cyclones, through Sociat-o7--anzational Ministry of the Inter-ior --cihaa Ministy of he Ineriorits own information network. Between May and Source: Ministry of the Interior. Table 3.2 Actions taken during various states of emergency Err,ergc,Cv sate Definitior, Action Pre-warning Occurrence of catastrophic phenomenon possible. Prevention actions taken by local authorities. \\'arning Occurrence of catastrophic phenomenon imminent- Preparedness program readied for search and rescue. Alarm Catastrophic phenomenon produces damage. Local emergency declared and emergency program executed. Emergency Health and security risk exceeds state capacity to State of emergency declared and emergency program respond executed. Disaster More than one state suffers severe damage affecting President declares disaster. social structure and vital systems. Federal response measures required. Source: SINAPROC. Mitigctting Daoingc by Reducing the Level of Risk 19 Table 3.3 Emergency actions and technical coordinators What Steps Have Been Taken to Mitigate Risk? Action Tcckitvol coordirnator Warning Ministry of the lnter.or Important steps have already been taken in Mexico to Damainge asessmnt Ministry of the Interior mitigate disaster risk. Scientific advisory committees Darnage assessment Ministry of the Interior (CENAPRED) have been established, standards for civil works have Emergency plans Ministry of the Inteior been set, engineering advances have been made, schools (National Center of Operations) have been retrofitted to withstand earthquakes, and a Emergencv coordination Ministry of the Intenor program for certifying hospitals that meet disaster readi- (National Center of Operations) ness standards has been put into effect. Security Ministry of the Interior Search and rescue Ministry of National Defense Scientific Advisory Committees Logistical support Ministry of the Interior In response to the active state of the Popocatepetl vol- Health Ministry of Health cano since December 1994, CENAPRED set up scientif- Provisions Ministry of Trade and Industrial ic advisory committees to aid the monitoring and Development decisionmaking process. Four committees, each made Emergency communications Ministry of the Interior up of unpaid experts in the appropriate field, were Source: Ministry of the Interior. established to study geology-related events, weather-relat- ed events, chemical-related events, and the social impli- November, observation and related telecommuni- cations of an eruption of Popocatepetl. Each of these cations networks constantly monitor weather sys- committees has contributed to the mitigation effort in tems, providing updated satellite imagery every 30 Mexico: minutes. Bulletins and warnings of tropical cyclones . The scientific advisory committee on geology-relat- are passed on to the Ministry of the Interior's Civil ed events has four functions. It strengthens efforts Defense Directorate three times a day, allowing civil to enhance the safety of dikes securing water reser- defense directors to alert civil defense units in at- voirs, undertakes the necessary studies and moni- risk zones. toring of Mexico's active volcanoes, promotes and Volcanoes. Four volcanoes are currently being mon- coordinates the updating of local building regula- itored in Mexico. CENAPRED and the National tions and their enforcemen-it, and supports detection, Autonomous University of Mexico monitor monitoring, and warning systems for tsunamis in Popocatepetl; the University of Colima monitors Col- Mexico. ima; the governments of Mexico and Guatemala mon- . The scientific advisory committee on weather-relat- itor Tacana ; and CENAPRED and Vera Cruz University ed events provides information on weather condi- monitor Pico de Orizaba. Emergency response pro- tions and forecasts, especially information relating grams have been designed for dealing with eruptions to tropical storms and hurricanes; assesses the vul- of Popocatepetl and Colima. nerability of bridges to being swept away by floods; SINAPROC provides a structure for coordinating and calibrates and uses meteorological radar in emergency preparedness and response activities at the Mexico. It analyzed the government's performance federal and subnational levels. Long-term disaster risk before and during Hurricane Pauline. reduction, however, requires the mobilization of dif- . The scientific advisory committee on chemical-related ferent institutions and skills from those appropriate for events monitors the processing and transportation emergency response. Long-term mitigation must be of dangerous substances, forest fire management, com- incorporated into strategies for sustainable development mand systems and transport and distribution of and tested against overall social and economic goals. toxic waste; assesses the vulnerability of communi- Efforts for longer-term risk reduction in Mexico are ties to industrial accidents; and fosters mutual industrial discussed below. help groups. 20 Maraging Disaster Risk itt M'lexco: Markec Incentives for .kitigation Investment The scientific advisory committee on social sciences sity faculty have worked with the Mexican Society for has undertaken studies on the social problems faced Earthquake Engineering to bring up-to-date training to by communities at risk from an eruption of Popocatepetl practicing engineers in other parts of Mexico. and created a council made up of chairpersons of the The Pan-American Health Organization has created scientific advisory committees. a program of voluntary certification of disaster readi- ness for health facilities interested in preparing for and mitigating the effects of disasters-a concept that B °dding and Readiness Standards could be extended to other types of institutions, such The Federal Electricity Commission has played an impor- as schools, government agencies, municipal govern- tant role in natural hazard mitigation in Mexico by ments, and private companies. The prograrr. provides developing the Manual of Civil Works. This 10-volume guidance on appropriate measures and certrfies com- work includes current engineering standards for pliance with those measures based on expert evaluation earthquake-, wind-, and flood-resistant designs. The of structural safety, provision for functionality follow- manual serves as a voluntary reference standard for design ing a disaster, programs for mass care, special staff train- engineers in all parts of Mexico that lack modern and ing, and organizational and administrative orientation. complete design standards. Regrettably, the research team Certification enhances an institution's reputation and that has updated the manual has been disbanded because indicates its level of disaster preparedness. Despite the of budget constraints, and the future of this critical ref- benefits of certification, however, less than 5 percent of erence is uncertain. hospitals have undergone the process. Thanks to studies of the seismic risks facing Mexico City and seismic engineering instruments that have been installed, much is now known about the characteris- Urban in as tics of ground motion during earthquakes. Current building regulations in Mexico take into account sub- Mexico's urban areas are home to 74 percent of the coun- soil types in different parts of the Mexico Valley to deter- try's population. Rapid and uncontrolled growth of the mine the kind, size, and use of construction allowed. major cities has increased the vulnerability of their inhab- Mexico City's construction norms are now very strict, itants-especially those with low incomes-to the effects resulting in much safer buildings than those built before of disasters. A recent example was Hurricane Pauline, 1985. Programs for retrofitting and reinforcing dam- which caused great human losses, not through high aged buildings with weak structures have been devel- winds or tidal surges but through extraordinarily high oped, with priority given to hospitals and schools. rainfall, which led to runoff down steep slopes of moun- All school buildings in seismic areas of Mexico were tain sides that had been settled illegally retrofitted to comply with the Manual of Civil Works in Mexico's National Plan of Urban Development 1989. The project was initiated and managed by Daniel (1995-2000) details the country's policy for the sector. Ruiz, the former director of the Earthquake Engineer- The plan makes proposals concerning the location of ing Institute of the National Autonomous University of population and economic activity at the national and Mexico and current director of the National Water Com- subnational levels taking into consideration the historic mission. The effort represented an extension of the pro- formation of human settlements in Mexico, the inter- gram financed by the 1986 World Bank Emergency nal organization of population centers with regard to Reconstruction Loan, which financed the strengthen- land use, and the basic infrastructure requirements of ing of schools in Mexico City. All school buildings in the population. Mexico are now insured by the federal government. Land management policy at the national level ties the The Earthquake Engineering Institute of the Nation- objectives of settlement policy to those of urban devel- al Autonomous University of Mexico has played a par- opment. This is done in a framework of coordination ticularly important role in the revision of seismic design that aims to promote local economic activities, gener- standards following the 1985 earthquake. The univer- ate employment, improve living conditions, and increase Mitigating Damuiagc by RcdLucing the Levcl of Risuk 21 equity in public sector management. The National subsidies for mitigation. Private sector incentives could Plan of Urban Development (1995-2000) incorpo- include reduced insurance premiums or reduced inter- rates strategies in four areas: est rates on construction loans. * Urban development of 100 cities * Consolidation of large metropolitan cities * Land use planning and support of urban development duc * Promotion of participation of all stakeholders in urban The educational approach to changing attitudes is based development by establishing consultative councils at on knowledge. Knowledge about natural hazards, vul- all levels and providing technical assistance to these nerability, and mitigation is based on scientific and engi- councils neering research. Continued strong funding for the Urban design is an essential tool of disaster mitiga- high-quality research on natural disaster risk reduc- tion. Planning, land use management, and building reg- tion conducted at such institutions as the National ulations by themselves can make an enormous Autonomous University of Mexico and CENAPRED is contribution to mitigating the worst effects of natural critical to the success of the educational approach. and other disasters. In practice, however, the imple- Dr. Roberto Meli, director general of CENAPRED, mentation and impact of these regulations has been lim- has highlighted four main research areas that can con- ited. Disaster risk management should be integrated tribute to mitigation efforts: into urban planning to ensure the safety of human set- * Evaluation of the management of recent Mexican disasters. tlements and their infrastructure. Recent disasters can provide valuable lessons on improv- ing disaster prevention and mitigation strategies. Inves- tigation of contributing factors, damage evaluation, and Recommendations emergency management and recovery measures in three recent disasters-Hurricane Pauline (1997), the 1998 Disaster risk can be reduced if high priority is placed Chiapas rain storms, and the 1998 forest fire season- on safety in all planning and development. Doing so could prove particularly instructive. requires a change in attitudes and behaviors, however. Modernization of building codes and quality verification Three approaches can help effect such change: programs. Building codes in Mexico need to be updat- • Education. Accurate, understandable information on ed. They also need to be made more uniform and disaster risk and mitigation measures must be pro- enforced throughout the country. Model codes should vided to key decisionmakers and to the public at be developed that correspond to local hazard risks large. and construction practices. Mechanisms to ensure * Regulation. Those who disregard public safety should code enforcement should also be studied. be restrained in the public interest. Permission to build *Methodologies for local community risk management in known hazard zones should be denied, and safe studies. Sound disaster mitigation programs can be construction should be required by law Local gov- prepared and implemernted only on the basis of ernments, which regulate urban planning and land detailed risk studies in which local physical and social use, must ensure that hazardous sites are not occu- conditions are taken into account to build alterna- pied by punishing violators who put public safety at tive damage scenarios. Such studies should use a risk. Structures built in contravention of the law, includ- common methodology leading to the development ing structures in hazard zones and structures that do of mitigation measures and emergency management not meet safety standards, should be demolished. programs tailored to local communities. A pilot "micro- * Incentives. Individuals and corporations that engage zonation" study could be performed to demonstrate in appropriate behavior (choosing safe locations and the proposed methodology constructing safe facilities) should be rewarded. Where * Formulation of megadisaster scenarios. The occur- economicallyjustified, public sector incentives could rence of a rare unanticipated natural disaster could include fiscal measures, such as tax credits or direct produce much more severe losses than those covered 22 Managing Disaster Risk in Mexico: Market Incentivesjor Mitigation Investment by Mexico's disaster relief fund. To establish insur- igation are not available in many parts of the coun- ance coverage needs and prepare mitigation strate- try Lack of training in seismic, wind, and hiydrolog- gies, it is necessary to estimate the maximum ic design means that many practicing engineers have losses that could be incurred given the occurrence only a limited understanding of special design pro- of various events with different probabilities of tran- visions. The Mexican Society for Earthquake Engi- spiring. Some possible megadisasters include a neering has worked to promote education about Mitch-like hurricane hitting the Yucatan Peninsu- earthquakes in some parts of the country Training la and the Mexican Gulf Coast, damaging the on mitigation must be provided throughout bhe coun- Campeche offshore oil fields, the Tabasco-Ver- try, and it must cover all potential disasters. Provid- acruz petrochemicals facilities, and highly popu- ing this training will necessitate additional funding lated areas; an 8.2 Richter magnitude earthquake for CENAPRED and the National Autonorr.ous Uni- in the Guerrero gap, damaging Acapulco and Mex- versity of Mexico. ico City; and a Plinian eruption of the Popocatepetl volcano, with large pyroclastic flows destroying Regulation nearby inhabitated areas and heavy ashfalls cover- ing a radius of 80 kilometers from the crater (includ- The traditional approach to natural hazard mitigation is ing much of Mexico City). coercive regulation. Building codes and zoning laws are Convincing the public of the importance of taking promulgated by public agencies and enforced through action to mitigate natural disasters is difficult, because the legal system, writh varying degrees of effectiveness. most people are unaware of the level of risk they face There is no provision for choice of risk exposure by and are unwilling to invest in programs that may yield individuals. benefits only in the long run. Information and educa- This approach is feasible if there is an organized tion are key to overcoming these barriers. Active pro- and effective plan review and field inspection process grams of dissemination, targeted professional education, and if the population accepts the rationale of the reg- and broad public education on disaster risk and mitiga- ulation and is able and willing to pay the costs of tion should be developed and delivered throughout Mex- enforcing it. Where these conditions are not met, the ico. These efforts should include a broad program of effectiveness of the regulatory approach may be public information on natural hazard risks, risk mitiga- reduced. tion, and insurance, undertaken as a collaborative effort In Mexico, as in many other countries, regulation by the government and the insurance industry. The objec- and enforcement are not applied uniformly To improve tive of this program should be to create a 'culture of the effectiveness of this approach, policymakers should safety" conduct a comprehensive reassessment of regulations Including disaster preparedness and mitigation in governing land use and building construction. At the elementary and high school curricula is also recom- same time, new approaches to land use and building mended. Helping to develop a culture of safety among quality management should be developed that effec- young children would provide long-term benefits. tively contribute to public safety Recognizing that most Special efforts should also be made to educate low- mitigation decisions are made at the community level, income communities, which are typically the most resources must be allocated to increase capacity and vulnerable to disasters, on disaster mitigation. authority for risk management and disaster initigation Mexico has made substantial progress in mitigat- at the state and community levels. ing disaster risks since its 1985 earthquake, thanks in part to the world-class experts at CENAPRED and the National Autonomous University of Mexico. Efforts have not been uniform throughout the country, The incentive approach involves both public and pri- however. Specialized training in earthquake engi- vate sector incentives for engaging in behavior that neering and other specialties related to disaster mit- reduces risk. The public sector could encourage desir- Mitigating Darnage by Reducing the Level of Risk 23 able behavior by providing community-level technical The Role of the Fund for Natural Disasters assistance for disaster mitigation, establishing low-risk (FONDEN) districts that include tax incentives for mitigation invest- ment, providing financial assistance for engineering The Mexican government allocates budgetary funds analyses of risk, and establishing certification programs, for disaster relief and reconstruction efforts by placing such as the Pan-American Health Organization program them in the Fund for Natural Disasters (FONDEN). for health facilities described above. FONDEN is not currently permitted to use its funds The private sector could also encourage risk mitiga- for pre-disaster mitigation purposes. It can, however, tion. Engineering firms could donate design assistance allocate funds to postdisaster mitigation efforts in areas or provide such assistance at low cost. Most important, where disasters are recurrent. Such postdisaster miti- insurance companies could play a much larger role than gation could help break the cycle in disaster-prone areas. they currently do in Mexico. Following a disaster, FONDEN could help mitigate In industrial countries the insurance industry is one future disasters by providing grant funds worth 5-10 of the primary sources of information and expertise on percent of the relief allocatior. to be used for approved dealing with risk. Households and businesses covered mitigation projects in impacted areas or states. Tempo- by insurance enjoy greater security both because they rary employment programs, which are generally part of adopt risk reduction measures and because they trans- the relief effort, could be structured to implement miti- fer the risk to the insurer (and are therefore ensured that gation measures. Temporary workers could help strength- their losses will be replaced). Insurance coverage would en buildings in earthquake- and wind-prone areas, improve thus be of great value to people in disaster-prone irrigation facilities in drought-prone areas, resettle com- developing countries, where most of the population munities away from hazardous sites, clear stream beds lacks reliable information on the risks to which they and drainage channels, construct flood protection, and are exposed and knowledge of effective means to mit- reforest damaged areas. igate those risks. Mitigation programs that require some financial input Insurance density-defined as the share of the from beneficiaries through cost recovery mechanisms potential property loss that is insured-is generally could help educate the popuLation about the risk and low in developing countries, where many people can- cost of disasters and the need to mitigate them. Such not afford to purchase insurance (see chapter 4). Reg- programs could be developed and overseen by ulators should support greater penetration of insurance CENAPRED and assisted by student mitigation teams. in Mexican society and greater access to the insurance FONDEN could also require municipalities or states to mechanism for people and enterprises of modest participate in mitigation projects by conditioning assis- means. Regulation should also encourage the broad tance on participation in such projects. Support, for provision of technical assistance for loss management example, could be conditioned on the state's commit- and premium pricing to provide incentive for miti- ment to review and update building and zoning codes gation investment. and to carry out a comprehensive hazard assessment. |-: ~~~~~~~~~~~~~~ ga, ~ ~ ~ ~ ~ 44 * - 0; ::: iTs .tX_XSe Chaptei 4 Mitigating Losses by Transferring Risk to the Insurance Industry In most countries in which potential losses are insured, sionally, reinsurance coverage is canceled altogether catastrophe risk is transferred from the national insur- after a disaster (as it was in some small island coun- ance system into worldwide risk-sharing pools. These tries hit by severe tropical cyclones). Often some risk pools are managed by international reinsurance com- mitigation has been requirecL to reattract the reinsur- panies and backed by substantial capital resources, ance industry sometimes supported by generous tax regimes that A number of mechanisms have been developed in allow the accumulation of reserves against future cat- wealthier countries in recent decades to cope with astrophes out of pre-tax income. In countries without large risk exposures. National disaster funds have been well-developed direct insurance systems and countries established in Turkey and Mexico. Compulsory state- in which the potential loss is very large, much of the mandated insurance-based pools-often backed by catastrophic risk remains within the country, where it high-level reinsurance or capital market instruments- has the potential to adversely affect macroeconomic have been established in Florida, France, New Zealand, variables. and Norway In Tokyo financial instruments have been developed that rely on objective measures of the sever- ity of the event (rather than the loss). All of these mea- Insurance and Reinsurance Coverage sures could potentially be applied in developing countries. throughout the World (For a description of some cf these mechanisms, see annex 3.) One measure of the extent of insurance coverage with- in a country is insurance density, the estimated Table 4.1 Estimates of insurance density proportion of private property (buildings and con- (percent) tents) insured against natural hazard. Insurance den- F' - i Evcnlt/llcatioi7l Tnlsar-cmcr n/salt sity varies across countries and across types of events (table 4. 1). E--a tl,,quakc6 (table 4.1). ~~~~~~~~~~~~~~~~South A-rica 95 As table 4.1 indicates, a high level of income is nec- Svdney Australia 90 essary for high insurance density. Even in high-income Israel 90 countries, however, density can be low if the risk and Nancouver. Canada 50 Lalilornia, United States 40 possible severity of an event is high (as in Japan). Italy 20 The reinsurance markets have historically adopted Belgluti 10 an implicit financing approach under which losses N4exieL <10 incurred in one period are "paid back" in subsequent Tokc Japan 5 periods through increases in the cost of reinsurance. Flood As a result, reinsurance prices have fluctuated wide- United Sntor 190 ly, and these fluctuations are passed on to the con- Germany <10 sumer. Rate increases after natural disasters can be Netherlands <5 large enough to affect the viability of businesses. Occa- Source: Swiss Reinsurance Co. 1997. 25 26 Managing Disaster Risk in Mcxico: Mfarket Irncentivesfor Mitigation Investment Demand for Property and Casualty Insurance then. In 1996 insurance density in Mexico was among in Mexico the lowest in the world for countries with similar income levels (table 4.3). The Mexican property and casualty (or "general") insur- The limited capacity to consume insurance in Mexi- ance market is underdeveloped, with the great major- co may reflect the high cost of catastrophic insurance ity of property owners uninsured against major natural coverage in many parts of the country In area highly catastrophes. According to the 1990 population and prone to earthquake, including Mexico City annual prop- housing census, 8.2 million of Mexico's 16 million hous- erty coverage costs about 0.5 percent of the amount es (51 percent) were built with solid materials. Assum- insured. Insuring a house worth $20,000 wvould thus ing 85 percent of these homes had potable water and cost about $ 100 a year-about 3 percent of the average drainage, about 7 million houses in Mexico were Mexican's annual income. It is unrealistic to expect insurable in 1990. Given an estimated annual rate of low-income families in developing countries to allocate increase in housing of 1.85 percent since 1990, about this large aproportion of their income to insurance when 8.3 million houses were insurable in 1998. Of that num- most of their income is devoted to food, shelter, and ber, only 150,000 homes, or 1.8 percent of the market, transport. were actually insured in 1998, according to industry In contrast, the country's major corporate and nation- sources. More than 8 million homes in Mexico are thus al public sector assets employ world standard risk man- insurable but uninsured. agement methodologies. About 90 percent cf industrial Themainreasonsforlackof coveragearelackof aware- enterprises and 50 percent of commercial enterprises ness and understanding of the role of insurance, the belief are insured. The rate of coverage is much lower in the that the govemment will fund any serious losses arising small business and private property sector, where only from natural disasters, and lack of income. While more about 2 percent of the market is insured. research needs to done on the relationship between income The public sector accounts for about a third of all levels and insurance consumption, the data suggest that property insurance premiums in Mexico. Federally owned high and evenly distributed income is correlated with assets appear to be adequately insured. State and munic- increased insurance consumption (table 4.2). ipal governments appear to be insured on an ad hoc Insurance consumption in Mexico rose before the basis, however, and their coverage may be insufficient. 1994-95 economic shock, but it has dropped since Increasing demand for insurance in Mexico will Table 4.2 Insurance penetration and macroeconomic Table 4.3 Insurance penetration in selected cleveloping indicators in selected countries, 1996 countries, 1996 Insuurance Per copita Insurarice premnnres consurnmpon of pr-em u o GDP per capitao GnP cs Gi percentage rnsuratncc" as a pe7rcertage Canoutry 'IJ.S dloZlars) coejfcicntt! of GDP" Country (U.S. doloars) of GDP' L.nited Sta es 26,030 <35 4.8c Panama 82 2.70 United Kingclom 19,720 <35 3.4 Costa Rica 61 2.28 Spain 14,810 <35 2.9 Columbia 39 1.77 Sweden 28,271 <35 2.3 Slovakia 62 1.76 Argentina 8,584 >45 1.2 Poland 56 1.62 Mexico 3,512 >45 0.8 Tunisia 32 1.51 Philippines 1,197 >45 0.8 Venezuelab 46 1.49 Zambia 325 >45 0.6 Thailand 40 1.32 Russia 28 0.95 a. The Gini coefficient is a measure of income distribution. The lower the Mexico 28 0.81 coefficient the more equally income is distributed. Peru 17 0.68 b. Exclusive of life insurance. Peru_ 17 _0.68 c. The consumption figure is heavily influenced by the private health a. Exclusive of life insurance. insurance system. b. Figures include some social insurance. Source: Swiss Reinsurance 1998. Source Swiss Reinsurance Co. 1998. Mitigating Losses by Transferriog Risk to tlte Insurance Intlnstiy 27 depend on several factors. Income levels must rise so to place a policy with a foreign insurer. Second, the that more of the population will be able to afford prop- Mexican insurance industry is currently preoccupied erty insurance. Greater awareness of insurance prod- with developing the pension and life insurance mar- ucts must also be created, and the population must come kets. Relatively little attention has been paid to the to understand that the government cannot meet all needs property and casualty market. Third, the leading insur- following serious catastrophes. ance companies in Mexico are privately owned and may have a more conservative approach to under- writing and growth than traded companies wiLh access Supply of Property and Casualty Insurance to strong capital markets. in Mexico Mexican insurance companies do not offer stand-alone catastrophic insurance, which must be purchased as an Lack of property and casualty insurance coverage in extension of a basic fire policy. About 50-80 percent of Mexico appears to be caused largely by demand-side Mexicans buying fire coverage purchase catastrophic cov- rather than supply-side factors. Although some factors erage. While most insurance is priced competitively, earth- distort the supply of property and casualty insurance quake coverage is subject to an industrywide price in Mexico, excess catastrophic reinsurance capacity is agreement approved by the government. In practice, how- available and a potentially large distribution system ever, earthquake rates are discounted. has been developed. Moreover, by 2000 the market will be open to unconstrained competition. Reinsurance Premiums for property and casualty insurance against natural catastrophes reached 1.6 billion pesos (about Mexico has the most concentrated insurance industry $200 million) at the end of 1997, 37 percent of the in Latin America. In 1997 the top five insurers wrote amount spent on fire and catastrophic premiums. About 70 percent of all property insurance policies. (Compa- 1. 1 billion pesos of this amount was transferred to rable figures for other Latin American countries were reinsurance companies. 27 percent in Argentina, 51 percent in Brazil and Reinsuring domestic risk makes sense, because it Chile, 50 percent in Colombia, and 43 percent in helps local insurers spread risks over a worldwide pool Venezuela.) Although concentration in Mexico has of insurers. However, catastrophe reinsurance pricing dropped since 1997 as a result of the influx of new com- was very volatile in the late 1.990s (table 4.4). The very panies, it still remains much higher than in other high catastrophe reinsurance premium rates in 1994 countries in the region. and 1995 reflected the enormious losses incurredby the Development of the property and casualty insur- global reinsurance sector in 1992 through 1994 and the ance sector is inhibited by several factors. First, all tight worldwide market for reinsurance that followed. policies written in Mexico must be underwritten by With increasing capacity attracted by high profits, prices local underwriters unless special authority is granted have dropped in recent years. Given Mexico's heavy and Table 4.4 Reinsurance pricing of catastrophic risks in Mexico, 1994-98 (percent) itemii 1994 1995 1996 1997 1 998 Catastrophe reinsurance coverage as a percentage of reference loss' Up to 122 7-156 3-118 2-140 2-170 Rate on line paid Ipercent sum at risk for reinsurers)5 4.4 4.5 3.0 1.7 1.0 Pure risk rate on line (percent stim at risk for reinsurers) 0.7 1.6 2.0 1.5 1.2 a. Reference loss is the estimated insured loss for which primary insurers are liable in the event of a major natural disaster. b. Rate on line is the reinsurance premium paid as a percentage of the layer of catastrophe reinsurance purchased. Source: Mexican Insurance Association. 28 Managing Disaster Riski in Mexico: Market Incentives for Mitigation Investment growing reliance on this method of risk transfer, Sale of Insurance mechanisms may need to be created to cushion the Insurance is sold largely through agents and brokers insuring public from price shocks (which may result licensedbytheinsurancesupervisor,theNationallnsur- from disasters occurring outside Mexico, most likely ance Commission. About 20,000 agents and brokers are in the United States or Japan). licensed, although the majority work part time or are The total underwriting profit made by offshore rein- inactive, according to industry sources. surers in recent years from fire insurance (including Insurance can also be purchased through banks in catastrophic coverage) has been very high relative to Mexico. Experience elsewhere in the world suggests that premiums, despite the large number of earthquakes sales through banks are likely to represent the fastest and hurricanes, some of which caused considerable means of expanding market penetration amnong small damage (see annex 2). This demonstrates the market businesses and individuals, especially if an active mort- failure of the insurance mechanism in Mexico. The rein- gage market is developed by the main commercial banks. surance mechanism can be economically effective only if the underlying insurance industry covers a high proportion of risks in the country (table 4.5). Mitigation of Risk by Insurance Companies Compulsory Reserves Mexican insurance companies have done li:tle to con- tribute to the prevention or mitigation of damage from Another source of additional potential market distortion disasters. Higher premiums for earthquake coverage in Mexico arises from the requirement that all insurers are charged for buildings without seismic resistant design set aside 35 percent of their cumulative earthquake features. Because quality control during construction is premiums as reserves. For the industry as a whole, often lax or nonexistent, however, lower premiums for more than $400 million was held in reserves in 1998, buildings that are supposed to be earthquake-resistant and these are required to be invested in liquid, short- do little to mitigate disasters. term government bonds. Collaboration between the Mexican insurance com- The legal status of these reserves remains unclear. panies and the National Autonomous University of Mex- Mexican regulators interpret the law to mean that earth- ico has taken place, but the work has focused only on quake reserves are held in perpetuity against infrequent tariff setting and reserves. Contacts with C] NAPRED but large catastrophes. The industry position is that have been superficial, and no actions hav2 resulted they are part of capital and revert to shareholders in from them to date. the event that the business is terminated or ceases to write catastrophic risk policies. Rules are currently being drafted to clarify the status of these funds and Recommendations ensure a more effective relationship between risk expo- sure and prudential management of insurers' balance Private insurance companies in Mexico are better able sheets. than the public sector to deal with risk. Much needs to Table 4.5 Underwriting profit earned by offshore reinsurers, 1995-97 (iv mors of U S. do lars) jt, i1, , 999 196 1997 Prope[nx ( l(e' premiums Written in Mex.co 45t.5 448.5 488.2 Prern:ums remnsured 315.1 305.6 336.8 L_r.de-writing profit o foreign reinsurers (before reserves) 105.1 183.6 159.6 .rsurance claims paid 265 7 40 \umnher of earthqiaakcie 3 1 0 Numlber of windstormrs (inciuc ing hurricanes1 3 3 2 Source Mexican Insurance Association and Nat onal Insurance Commission. Mitigating Losscs by Transfer-iing Risk to the Insurance hidoLsry 29 be done, however-including getting support from mul- * In order to raise insurance awareness, the mandato- tilateral agencies-before insurance can play an impor- ry automobile liability insurance requirement, tant role in preventing and mitigating the effects of disasters. which was postponed, should be implemented as Private insurers should be encouraged to deal with soon as possible. At the same time, the insurance the risks faced by small and medium-size enterprises. industry, in cooperation with the government, should Letting the private sector deal with this part of the mar- be encouraged to educate the public about insurance. ket would reduce the load on the public sector at the * The govemment may wish to consider providing some time of a natural disaster, and it would generate foreign incentives for householders and small businesses to exchange inflows at a time when the balance of payments buy insurance, possibly through the use of (means- may be under pressure. Private sector involvement would tested) vouchers or tax rebates. also lead to speedier claims assessment and settlement. Improvements in insurance coverage could also be In addition, insurers in a developed market generate use- achieved by changing the public sector's procurement ful risk management information for businesses and indi- process. The governments nmandatory tendering pro- vidual homeowners because the price of insurance curement process has resulted in a suboptimal level of rates varies according to the characteristics of the risk. insurance coverage for all but the most sophisticated To increase insurance awareness and density, insur- purchasers of insurance (such as PEMEX). Given that ance products must be made available to the public at 90 percent of all govemment business (other than PEMEX) affordable prices. Catastrophe insurance can currently is currently placed with the two leading private insur- be purchased in Mexico only as part of fire insurance, ance companies, the current individual government and the cost of such insurance is prohibitively high for agency approach appears to be inefficient. Consolida- most Mexicans, especially those in Mexico City A num- tion of purchasing combined. with the employment of ber of changes could be made in the short term to increase international brokers and risk transfer specialists should capacity and reduce prices: be considered for all public sector risks. * Catastrophic premium rates could be formally dereg- The efficiency of insurance as a risk transfer mecha- ulated, with a maximum tariff set at the current level. nism could he increased if pLblic sector agencies pur- Deregulating premiums would mean that consumers chased higher levels of catastrophic coverage while could benefit from low reinsurance costs. absorbing higher levels of initial losses. The Ferrocarriles * The discontinued 10 percent compulsory surcharge Nacionales de Mexico (national rail carrier), for example, on earthquake premiums could be reintroduced and purchased a $20 million mul:i-risk policy, which stipu- extended to all catastrophic classes (including civil lated that only the first $1 rnillion were available for disturbances). These funds could be placed in a cen- property losses. The overall level of coverage was gross- tral pool managed by the Mexican Insurance Associ- ly inadequate and failed to cover much of the damage to ation for use in smoothing out reinsurance pricing railway tracks incurred in the 1998 flooding in Chiapas. cycles and as a guarantee fund in the event that any Insurance advisers putting the insurance programs insurer became insolvent following a major disaster. together set prices so low that it is not possible to design Companies could continue to be required to accu- properly researched risk management programs that mulate catastrophic reserves up to an agreed upon take a holistic approach to mitigation, loss control, and standard retention level. rsk transfer. Information provided to insurers reported- • If the insurance industry is not prepared to provide ly includes only details of a few peak risks for some agen- catastrophe-only coverage, an industry-run pool could cies; very little information is provided on other exposures. also be required to offer such coverage, which would Insurers should be provided with adequate information be sold through private insurers. The pool would be on risk exposure, and large intemational brokers should available only to insured risks, however. be permitted to participate in the tendering process. ,..,,, a E| l~~~ Chapte- 5 Using New Financial Instruments to Transfer and Finance Risk Dealing with the effects of naLural catastrophes has proved internationally through insurance and reinsurance. In extremely disruptive to the budgetary process in Mexico. developing countries a combination of low incomes and What strategies can the Mexican government adopt to lack of awareness about insurance means that less than help it deal with these cnses more effectively? How can 10 percent of private property is typically insured. As effective disaster relief be provided without derailing other a consequence, the global reirnsurance market cannot programs? This chapter examines the options available be accessed. Catastrophe boncLs may provide a means to the govemment for transferring and financing risk, sug- of circumventing the underdeveloped state of the insur- gests some steps to take to pursue these options, and makes ance industry in many developi-ng countries exposed to recommendations for administering the Fund for Natur- catastrophe risk. al Disasters (FONDEN) more effectively Catastrophic risk transfer alternatives have been used for a variety of risks, including hurricane and earthquake risk in the United States and Japan, flood risk in most of New Financial Instruments for Transferring Risk the world, and storm risk in Europe. These tools could be extended to developing countries if the right structure Peak catastrophe risks throughout the world far were created to transfer risk and the right information pro- exceed the capacity of the insurance and reinsurance vided to professional risk takers. The development of both industries to insure them. Available capacity falls far appropriate institutional structures and proper modeling short of the estimated damages from even a single peak of the risk exposure are essential if these alternative risk catastrophic event (such as a major earthquake on the transfer mechanisms are to be used in Mexico. New Madrid fault in the United States, which could cost up to $110 billion in direct losses). In such an event it Catastvophe Bonds and Options is estimated that more than a third of the U.S. insur- ance industry would become insolvent, even after allow- Although the insurance risk exchanges have yet to ing for reinsurance of their exposures (Insurance Services develop a major presence, they are growing, with total Office 1999). risk transfer amounting to about $3 billion to date. Recognizing the inability of the insurance industry The Chicago Board of Trade lrst launched catastro- to cover all catastrophic risks, insurers and corporations phe futures and options in December 1992. Since then have sought to spread these risks to the capital markets the Catastrophe Exchange (CATEX) and the Bermu- by issuing catastrophe bonds and options and other da Commodities Exchange have begun trading similar instruments. The development of computer mod- insurance risk. In 1999 the Chicago Mercantile Ex- eling of natural perils and the growing understanding change introduced temperature-extreme derivatives, of catastrophic risk among institutional investors is which had been trading over-the-counter among ener- increasing market capacity of these new financial instru- gy utilities in the United States. These derivatives are ments (Froote 1998). designed to enable energy utilities to smooth income In industrial countries, where a high proportion of when power usage is low because of mild winters or property is insured, much catastrophic risk is spread cool summers. 31 32 Managing Disaster Risk in Mexico: Market Incentives for Mitigation Investment Act of God bonds were first issued in 1996. These Risk Transfer and the Fund for Natural Disasters bonds-which are often structured as reinsurance arrangements for legal and tax purposes-are nothing Funds earmarked for relief and reconstructicn after catas- more than regular bonds issued with embedded con- trophe events in Mexico are maintained in FONDEN. tingent options that provide the right to withhold The fund has three main strategic objectives: some or all of the principal and accrued interest in the * Complement the systems of civil protection and assis- event of a defined trigger event. Between 1996 and 1998 tance to victims. this market (including related instruments) issued more * Finance repairs of damage to infrastructure caused than $2.5 billion of paper to about 50 institutional bydisasters,withoutaffectingthebudgetedprograms investors. Spreads varied according to structure and and projects of the federal administration. exposure but averaged about 4 percent over the Lon- * Promote state and municipal participation to create don interbank offer rate (LIBOR). a partnership among the three levels of government A catastrophe bond need not rely on measures of to deal with the negative impact of extreme events. actual loss. The trigger could be an objective measure FONDEN is composed of three separate funds. The of the intensity of an event, such as an earthquake or infrastructure fund provides for the repair of unin- hurricane. One such bond has been issued for Tokyo sured infrastructure. The agriculture fund provides earthquake risk, which pays out according to location immediate assistance to restore the productivity of and magnitude, as measured by the Japanese Meteoro- low-income (subsistence) farmers. The ass.stance fund logical Agency Countries could set up a catastrophe provides relief to low-income victims of d:sasters. fund to issue such bonds, the proceeds of which Budgetary allocations for FONDEN have not been would be held in escrow pending a trigger event. sufficient to cover the fund's obligations since 1996. In No investor has yet lost money invested in these 1998, for example, FONDEN's budget was about $227 instruments. Some observers believe that the first real million, while its expenditures were expected to exceed natural catastrophe will see the end of the market, as $500 million. The shortfall is made up by diverting investors lose money and retreat from the market. funds from other government budgets. This need not be the case if investors include partic- ipation in these instruments as part of a diversified portfolio. Indeed, investors have shown interest in the new instruments because the natural risks cov- One way that Mexico could transfer its dlisaster risk ered are typically uncorrelated with major financial out of the country would be for the government to indices, thereby providing new opportunities for purchase catastrophe insurance. A hypothetical exam- risk diversification. ple demonstrates how this option would function. For purposes of illustration, assume that the government Contingent Credit wants to be able to cover up to $500 million in infra- structure losses caused by hurricanes, floods, and earth- Another instrument used to finance large catastrophic quakes and that FONDEN is willing to absorb $25 losses is contingent credit with long maturities. Under million in losses per incident, up to a maximum of S50 a contingent credit arrangement, the lender charges a million. FONDEN could purchase such coverage for, fee, which is paid as long as the trigger event does not say, S48 million a year (depending on the pricing cycle). occur (the International Bank for Reconstruction and Doing so would put a ceiling of $98 million on its expen- Development charges 0.75 percent). If an event does ditures ($48 million in premiums plus $50 million in occur, the borrower can rapidly draw down funds. To maximum uninsured losses). It would provide cover- be able to repay the credit, the borrower must have a age to segments of the society that would not have had sufficient stock of reserves or future premium income. the resources to purchase insurance themselves, and it Tf it does not, the government could assume the risk of would obligate both FONDEN and the insurer to take repayment through the taxation base. actions to mitigate the level of risk. LUsbig New Financtal Inst rLinents to T-anse l- anid Finanice Risk 33 Transferring Risk through the Capital Markets on the ad hoc capture of funds allocated to other agen- Another way in which FONDEN could transfer risk cies to meet its current requirements. To reduce the would be to issue a bond in the international capital uncertainty and disruption to the budgetary process, markets with characteristics similar to reinsurance. The FONDEN could use its current and future budgetary federal government could, for example, issue a $500 funding allocations to finance extraordinary costs when million bond to international investors, holding the pro- disaster-related damages costs are high. ceeds of the bond in an escrow account. Funds held in The most promising option would be to establish a escrow would be released to Mexico upon the occur- line of credit with a financial institution, which would rence of catastrophes that cause more than a given provide funds in the event that catastrophic events level of damage to infrastructure. If no disaster occurs, exceeded FONDEN's annual budgetary funding. Repay- investors would retain the principal as well as the ment of the credit line would take place over a long peri- interest earned on the bond. If a catastrophic event od of time (perhaps 12-15 years if multilateral funding occurs, funds advanced to Mexico under the bond would were obtained). Although the government would have be partially forgiven; interest on the advanced amounts to pay a commitment fee for the credit line, access to would be partially or totally forgiven. the line would ensure an immediate source of funding, Catastrophe bonds can be flexible and are designed obviating the need to divert funds from other govern- to meet the particular needs of the insurer. Pricing of ment agencies. Credit line arrangements-which have these instruments has been higher than the pricing of been used by insurance companies concerned about the reinsurance, although it has been consistent with the impact on their cash flow of a major catastrophe event- average pricing of reinsurance during the 1990s. are a useful tool because they involve no transfer of risk. Given the novelty of these instruments and the risks The insurability of the risk is thus irrelevant. faced in Mexico, a bond guarantee might be necessary to attract investor interest in such instruments. The World Bank's guarantee facility might be deployed to Using Alternative Risk Transfer Mechanisms to provide credit enhancement and augment the attrac- Deal with Uninsurable and Uninsured Properties tiveness and security of catastrophe bonds. A partial risk guarantee would consist of a contract-signed by Because insurance coverage in Mexico is very limited, the government, the World Bank, and investors-that the govemment is faced with large fiscal obligations fol- would commit the Bank to pay investors the interest or lowing natural disasters. Some public sector properties principal due in the event that the government could and infrastructure (such as utilities, certain roads, and not do so. If the guarantee were paid out, it would water systems) can be insured at the state and municipal become an obligation due from the government to the levels. Other assets, however, are uninsurable or are dif- Bank. The lower interest that Mexico should offer on ficult for the insurance industry to evaluate and price. the bond as a result of the Bank guarantee would more For these reasons, altematives to insurance, such as risk- than compensate for the modest guarantee fee charged transfer and financing instruments, should be considered. by the Bank. Transferring Risk Borne by the Public Financing Risk with Commercial Bank Credit Lines and Business Sectors If the Mexican government cannot or prefers not to The govemment should seek to adopt efficient risk man- transfer risk, it can explore various options for financ- agement techniques, including loss control, mitigation, ing that risk. Because the size and frequency of the nat- and risk transfer. Catastrophe bond and option contracts ural catastrophes covered by FONDEN vary from year could be used to insure state and municipal assets and to year, annual funding requirements fluctuate widely. infrastructure (sewerage systems, uninsurable buildings) Since supplementary funding from the government can- that may not attract the interest of the insurance indus- not be requested during the year, FONDEN has relied try (because of their unusual nature or lack of compliance 34 Managing Disaster Risk ir MNexico: Mfarket Incentives for Mitigation lrvesrmert with building codes, for example) or may command mex operates on the basis of strict underwriting and too high a premium for local governments to afford. They claims adjustments that are verified by a iarge team of could also be used to cover small and medium-size field technicians. Since 1990 Agroasemex has been able enterprises for which insurance is available but is too to place a relatively low-level stop loss program (75 per- expensive for owners to purchase. The covering of such cent loss ratio in excess of the first 75 percent loss enterprises through a financial instrument would free ratio, that is, any losses in excess of 75 percent of the the government from having to extend financial assis- premium amount are paid by reinsurance up to a max- tance followving a natural disaster. Bond and option instru- imum of 150 percent of the premium). Mexico is exper- ments could be used to cover medium-size cash crop imenting with a number of other leading-edge ideas, farmers. (In principle, such farmers could be insured including remote sensing of soil moisture content and through the government's crop insurance program. In the use of French-style cooperatives to reduce moral fact, however, such farmers remain uninsured.) hazard and information asymmetry. The government could potentially bypass the entire The main drawback of all such individual assessment insurance loss evaluation process by using an investment approaches is that they are extremely expensive to admin- instrument that triggers or limits certain payments to ister and almost always require some forrn of govern- investors based on objective indicators. Indicators, ment subsidy New approaches to crop-rela.ted disasters such as earthquake tremors and hurricane wind speeds, in which payouts are related to objective measures of can sometimes be transmitted to central satellite stations. rainfall and soil moisture rather than to crop yield itself are being examined by a number of countries and by the World Bank. Such approaches are potentially Transferrnng Risk Borne by the Agricu1tlral Sector much less expensive to administer and still leave an Many countries compensate farmers after natural dis- important role for the insurance sector. asters, almost always doing so out of current bud- getary resources. The degree of formality of these compensation arrangements varies widely In Belgium Recommendations and France, for example, a separate national agricul- tural disaster fund compensates farmers. In Australia, During the past decade the insurance and capital mar- Denmark, and Germany, compensation is distributed kets have developed several innovative techniques for largely on an ad hoc basis. In the United States farm- transferring the risk associated with natural hazard ers are eligible for free catastrophe insurance (subject events. The Mexican government should investigate to a $50 administrative charge). Despite the low cost, how it could use these new products to make more however, up to 20 percent of farmers remain uncovered efficient use of the assets it allocates following disas- (politicians generally bail them out following disasters ters. By transferring risk to the private sector or financ- anyway). ing the costs of disaster-related damage, Mexico could As the damage done by natural disasters has grown, reduce the disruption to the budget process caused by however, particularly in agricultural areas, governments unanticipated expenditures. In order to do so, it will have found themselves unable to cover the costs of com- have to take several steps: pensation, and politicians have come to recognize that * Reconstitute FONDEN as a legal entity that is capa- disasters can inflict long-term damage on the economic ble of transferring risk. base. In a number of countries (Australia and Morocco, * Group together similar risks in funds, so that they for example) governments are attempting to find ways can be aggregated, priced, and transferred separate- to reduce their contingent liabilities, by transferring catas- ly Allocate sufficient dedicated capital and premium- trophic crop risk using private sector mechanisms. paying capacity to FONDEN to support its risk In Mexico a well-developed agricultural insurance placements in the insurance or capital markets. system is available through Agroasemex, the state-owned * Inventory all vital infrastructure, and estimate agency set up to replace Anagsa in the late 1980s. Agroase- FONDEN's risk exposure. Catastrophe modeling- Using New Financial Ins:ruwncnts to Transfer- and Finance Risk 35 which analyzes the frequency of hazard events, the should be screened for economic viability in accor- severity of the events, and the vulnerability and value dance with the standard cost-benefit criteria used to of assets that are the responsibility of FONDEN- evaluate public investments. could be used to make these estimates. * Consider linking the terms of FONDEN financing for Develop financial models that allow various policy reconstruction following a disaster to the extent to options to be evaluated. Financial modeling based which mitigation measures had been implemented on the data developed in the catastrophe modeling before the disaster (possibly through cost-sharing would yield various options and prices for the arrangements with states and municipalities). transfer of risk. Clearly define the risks for which FONDEN bears Risk management could also be improved by chang- responsibility, beginning with the infrastructure fund. ing the way FONDEN operates. Specific recommenda- Limit coverage to those infrastructure items that are tions include the following: essential to support the country's economic and social * Refocus FONDEN's efforts by absorbing only those affairs following a natural disaster and those items residual catastrophe risks that cannot be absorbed that have already been insured for noncatastrophic by third parties, encouraging economically worth- risks. Limit agricultural coverage to farmers who are while mitigation by the public sector, and fostering unable to afford private insurance; farmers who are the development of private sector risk reduction able to afford private insurance should be obligated and transfer mechanisms. to purchase it. Welfare payments to subsistence farm- * Modify FONDEN's mandate so that it is permitted ers could be subsumed under the victims fund or to disburse funds for research and training on miti- paid through other social safety net programs. gating risk before a disaster occurs. Funding for * Establish guidelines to be followed by other gov- research and training would represent an addition ernment agencies in managing the risks they bear. to the annual budgets of executing agencies; the actu- FONDEN could, for example, specify the types and al financing of mitigation programs would presum- acceptable limits for insurance coverage for various ably be carried out as part of the public investment types of infrastructure. It could also help federal, programs executed by the various govemment depart- state, and local authorities purchase property insur- ments. Investment programs for mitigation works ance on a group basis. I - i I S , j g I 1 i | i I _1 I I I I _ I _ f , 1 I [ | . I Ff-iA:,!tit.si,! .w g | r-0P > ;T A; ft50Xa g S I L'=Aji.i,,# | | ,E LO: | ; _|y I _ _W_ t _ E I W _WkWtH _1q1 | ! Xt'\u __ _ W ., .. ;(.;S-:. Chaprei- 6 Summary of Findings and Recommendations Mexico is vulnerable to a variety of natural disasters, Mexico has a good core of institutional capability to including earthquakes, hurricanes, and volcanoes. Despite manage natural disasters, led by the Ministry of the Inte- the frequency with which these disasters strike, how- rior and the Ministry of Environment, Natural Resources, ever, inadequate investment is made in mitigation efforts, and Fisheries. High-level coordination and strategic and insufficient funds are set aside to pay for relief and planning for disaster management need to be improved, reconstruction efforts. As a result, when a disaster occurs, however. the government is often forced to use funds that had been allocated to other programs, disrupting the oper- ations of those programs. The effect is to derail impor- Mitigate Risk tant development efforts and possibly reduce growth. The Mexican government could improve the way it The damage caused by disasters can be reduced by chang- manages its risks so that ongoing programs are not dis- ing perceptions and behavior so that all members of soci- rupted following a disaster. Doing so involves identify- ety place a high priority on safety in planning and ing the risks the country faces, mitigating the damage development. This is important from a social and eco- caused by those risks, and transferring the risk to other nomic standpoint, since, in the absence of ample insur- parties (namely, insurance companies and the capital ance penetration in the private sector, a large proportion markets). of catastrophic risks are de facto borne by the govern- ment and society at large. All stakeholders must con- tribute to the creation of a culture of safety Toward that Identify Risk end, the government should develop strategies for improv- ing education about disaster mitigation, creating incen- Formulation of a risk management strategy must begin tives, and adopting regulations that will encourage with the identification of risk. Mexico should identify the individuals and businesses to reduce the risks they face. principal hazards it faces and assess the vulnerability of Information and education are key to raising aware- principal settled areas, infrastructure, and economic assets. ness of risk and the importance of investing in mitiga- The analysis should include estimates of the direct, indirect, tion measures. Active programs of dissemination, targeted and secondary impacts of natural disasters. professional education, and broad public education on Adequate funding must be maintained for the risk disaster risk and mitigation should be developed and identification efforts already under way. The research delivered throughout Mexico. Programs should include capacity of CENAPRED and the National Autonomous dissemination of information on natural hazard risk, University of Mexico is excellent, but neither institu- mitigation, and insurance (possibly through a collabo- tion receives the budgetary attention it deserves. Given rative effort by the government and the insurance indus- the quality of their work on identifying risk and miti- try); the inclusion of disaster preparedness and mitigation gating damage from volcanoes and earthquakes, their materials in elementary and high school curricula; and mandate should be expanded to include hurricane and special programs to target low-income communities, flood mitigation. which are often worst hit by disasters. 37 38 Managing Disaster Rish in Mexico: Market Incentives for Mitigation Investment The regulatory approach to encouraging disaster mit- participate in mitigation projects by conditioning igation in Mexico requires a comprehensive reassess- assistance on cost-sharing. FONDEN's support to states ment of the formal land use and building regulatory should be conditioned on a state's commitm,nt to review- processes. New approaches to land use maniagement ing and updating building and zoning codes and to and building quality management must be developed carrying out comprehensive hazard assessments. that effectively contribute to public safety in the infor- mal sector. Recognizing that most mitigation decisions are made at the community level, policymakers must Transfer Risk allocate resources to increasing capacity and authority for risk management and disaster mitigation at the Private sector insurance companies cart be used to state and community levels. help Mexico deal more effectively with risk in the pub- While new and innovative approaches to regulation lic sector. The public sector should purchase insurance are explored, new initiatives should focus on creating through international brokers and risk transfer spe- positive incentives. Such incentives could include pro- cialists; individual agencies should not be purchasing vision of community-level technical assistance for dis- insurance on their own. More cost-effective coverage aster mitigation, establishment of "low-risk districts" should be purchased by increasing the dectuctible level or historic centers that include tax incentives for those and purchasing higher levels of catastrophe coverage. that invest in mitigation, granting of financial assistance Regulation of the insurance industry should sup- for engineering analysis of risk, and establishment of port greater penetration of insurance in NMexican soci- certification programs, such as the Pan-American Health ety and greater access to the insurance mechanism for Organization program for health facilities. The private people and enterprises of modest means. Regulation sector could consider donating engineering design assis- should also encourage the broad provisior. of technical tance, providing it at low cost, or offering low-cost loans assistance for loss management and premium pricing for mitigation efforts. to provide incentives for mitigation investment. FONDEN's mandate should be modified so that it In order to increase insurance awareness and pene- can provide funds for studies and training related to tration among small and medium-size enterprises and disaster mitigation before catastrophic events occur. the general population, affordable insurance products Even if the government cannot or prefers not to mod- must be available to the public. Capacity could be ify that mandate, FONDEN can still play an important increased and prices reduced in several ways: role in postdisaster mitigation in areas in which disas- * Catastrophic premium rates could be formally deregu- ters are recurrent. Temporary employment programs lated, with a maximum tariff set at the current level. established as part of the relief effort following a disas- * The discontinued 10 percent compulsory surcharge ter could be used to implement mitigation measures, on earthquake premiums could be reintroduced such as strengthening buildings in earthquake- and and extended to all catastrophe classes (including wind-prone areas, improving irrigation facilities in civil disturbances). These funds could be placed in drought-prone areas, resettling communities away from a central pool managed by the Mexican Insurance hazardous sites, clearing stream beds and drainage chan- Association, which would help to stabilize reinsur- nels, constructing flood protection, and reforesting dam- ance costs and cover catastrophe-induced insolvency aged or destroyed areas. Companies could continue to be required to accu- Mitigation programs that require some financial input mulate catastrophic reserves up to an agreed-upon from benefnciaries through cost-recovery mechanisms standard retention level. would help educate the population about the risk and * This industry-run pool could be required to offer cost of disasters and the need to mitigate them. Such catastrophe-only coverage, which would be distrib- programs could be developed and overseen by uted by insurance companies. CENAPRED and assisted by student mitigation teams. * In order to raise insurance awareness, the mandatory FONDEN could also require municipalities or states to automobile liability insurance requirement, which Sunirpia y of Findings and Recomnmendations 39 was postponed, should be implemented as soon as on the data developed in the catastrophe modeling possible. would yield various options and prices for the * The insurance industry, in cooperation with [he transfer of risk. government, should be encouraged to educate the Group similar risks together in funds so that they can public about insurance. be aggregated, priced, and transferred separately. * The government should consider providing targeted Allocate to FONDEN sufficient dedicated capital and incentives to expand insurance density premium-paying capacity to support its risk place- ments in the insurance or capital markets. Modify FONDEN's mandate so that it is permitted Reorganize FONDEN to disburse funds for studies and training on miti- gating risk before a disaster occurs. Funding for stud- Before the Mexican government can make use of risk ies and training would represent an addition to the transfer and risk-financing tools, it must modify the annual budgets of executing agencies; the actual way FOINDEN is organized. Changes should include financing of mitigation works would presumably be the following: carried out as part of the public investment programs * Refocus FONDEN's efforts by absorbing only those executed by the various government departments residual catastrophic risks that cannot be absorbed through their normal budgetary allocations. Invest- by third parties, encouraging economically worth- ment programs for mitigation works should be while mitigation by the public sector and fostering screened for economic viability in accordance with the development of private sector risk reduction the standard cost-benefit criteria used to evaluate and transfer mechanisms. public investments. * Clearly define the risks for which FONDEN bears * Consider linking the terms of FONDEN financing responsibility, beginning with the infrastructure fund. for reconstruction following a disaster to the extent Limit coverage to those infrastructure items that are to which mitigation measures had been implemented essential to support the country's economic and social before the disaster (possibly through varying the cost- affairs following a natural disaster and those items sharing arrangements with states and municipalities). that have already been insured for noncatastrophic * Establish guidelines to be followed by other gov- risks. Limit agricultural coverage to farmers who are ernment agencies in managing the risks they bear. unable to afford private insurance; farmers who are FONDEN could, for example, specify the types and able to afford private insurance should be obligated acceptable limits for insurance coverage for various to purchase it. Welfare payments to subsistence farm- types of infrastructure. It could also help federal, ers could be subsumed under the victims fund or state, and local authorities purchase property insur- paid through other social safety net programs. A sim- ance on a group basis. ilar approach of identifying those that can afford private insurance could be developed for the victim assistance fund. The Role of the World Bank * Reconstitute FONDEN as a legal entity that is capa- ble of transferring risk. The study has helped the World Bank define more clear- * Inventory all vital infrastructure and estimate ly ways in which it could help developing countries FONDEN's risk exposure. Catastrophe modeling- manage disaster risk more effectively Instruments that which analyzes the frequency of hazard events, the could be of value include the following: severity of the events, the vulnerability of assets that * Loans for investment in mitigation and reconstruction. are the responsibility of FONDEN, and the value of Loans could be provided to finance retrofitting plans those assets- could be used to make these estimates. (at schools, hospitals, public buildings) and recon- * Develop financial models that allow various policy struction of infrastructure damaged by natural disas- options to be evaluated. Financial modeling based ters. An important feature could be to link the terms 40 iManaging Disaster Risk in Mexico: Market Incentivesfor Mitigation Investment of reconstruction financing to the extent of mitigation * Contingent loansfor risk financing. Catastrophe risk is carried out by the final beneficiaries of the financing. normally "packaged" into "layers" and inanced or A technical assistance component could support capac- transferred through a variety of arrangements, includ- ity building for various aspects of disaster manage- ing primary insurance, reinsurance, credit financing, ment. These could include data generation; risk and bond and derivative financing. The combina- assessment and modeling; development of disaster tion of instruments used will depend on specific cir- management systems and plans; redrafting of build- cumstances, including the relative costs of'the various ing codes and improvement of physical planning cri- instruments at the time the placements are made. The teria and systems; development of retrofitting plans; Bank could consider providing contingent loans to development of schemes for new risk financing and finance a credit layer in a risk management package. transfer arrangements and instruments; and design of * Partial risk guarantees. The Bank operates a guaran- improved laws, regulations, and policies to support tee program that supports private investment in bor- better risk management. rowing countries by guaranteeing against risks Contingent loansfor reconstruction. To make sure that associated with policy or contractual performance funds are accessible immediatelywhen disaster strikes, by the government. Under the program, -he govern- the Bank could establish a credit line that could be ment counterguarantees the Bank guarantee by agree- drawn upon under certain conditions. Establishing ing to repay the Bank for any payments it makes to such a line would speed the delivery of funds dur- investors under the terms of the guarant2e. In prin- ing emergencies. The credit line agreernent would ciple, this program might provide credit er hancement specify the trigger events that would make Bank funds for catastrophe bonds and commercial contingent available. It would also include guidelines govern- credit lines, thereby improving their terms and attrac- ing the use of funds (eligible expenditures; dis- tiveness to investors. bursement, procurement, and financial management This case study and future discussions wiv.h the Mex- arrangements; and on-lending terms to final benefi- ican government should help refine this menu of options ciaries, which could vary with the extent to which so that it addresses Mexico's primary concerns. The mitigation measures had been implemented before effort should enhance the quality of the Bank's lend- the disaster). Commitment fees would be paid before ing and nonlending activities in the increasingly impor- the loan was activated. tant area of natural disaster risk management. Annex I Geographic Distribution of Major Hazards in Mexico Map Al Active volcano risks and hurricane paths A Active volcanoes Seismic risk U High Medium Low L None x > ) >\ V , r Hurricane paths 'I State boundares 41 Annex 2 Estimated Cost of Disaster Events in Mexico, 1980-98 Mexico experienced dozens of disasters between killed in these events, which cost the economy about 1980 and 1998 (table Al). About 9,515 people were $12 billion in losses. Table A2 Estimated direct costs of disaster events in Mexico, 1980-98 Reportecd crect !osscs Yea-/ex eat/location Ccsaoottes and damage (ilIliors of D'S. dolla-s) 1980 Flood Tijuana and Ensenada, Baja California 3 people killed, 20 people missing, 6 towns destroyed, 30,000 victims 87.0 FAas plant fire Xalostoc, State of Mexico 4 people injured 0.8 Drought Coahuila and Sinaloa 222.6 Hurricane Allen Matamoros 25,000 people left homeless 4.3 1981 Information on losses not available 1982 Erupt'on of Chichonal volcano 42 people killed, 22,000 evacuated, 150,000 victims- including 17,000 Zoque Indians. Complete loss of cultivable land within 10 kilometer radius; partial loss of lands within 10-15 kilometer radius. Damage done to 4,000 hectares of banana crops, 15,000 hectares of corn, 905 hectares of bean crops, 10,000 hectares of coffee, and 17,000 hectares of cocoa. Eight airports temporarily shut down, access roads closed for weeks. Carpet of ash and volatile particles within 15 kilometer radius 117.0 Hurricane Paul Sinaloa 257,000 victims, 7,299 homes damaged. Complete loss of 7,334 hectares of soy crops, 2,300 hectares of corn, and 1,300 hectares of vegetables. Forty percent of 75,000 hectares of soy crops, 14,000 hectares of corn crops, and 13,400 hectares of vegetable crops damaged. Damage to industrial and commercial structures 82.4 Floods Throughout the country 8 people killed, 5 injured, 285,511 victims. Damage to homes, highways, electricity and telecommunications, and agriculture 114.6 42 Annex 2 43 Table A2 (continued) Reportedcl direct losses Year/event/llocation Casaalties and damage (millions oj U.S. dollars) 1983 Information on losses not available 1984 Industrial explosion San Juan Ixhuatepec 1,000 people killed, 5,000 victims, 200 houses completely destroyed, 150 homes left uninhabitable 26.3 1985 Earthquake Mexico City 6,000 people killed, 30,000 injured, 150,000 victims, 30,000 homes destroyed, 60,000 damaged. Damage to all productive, health, and service sectors 4,104.0a Floods Throughout the country 43 people killed, 47 injured, 73,241 victims. Damage to housing, highways, electricity and telephone system, and agriculture 39.4 Torrential rains Nayarit 48,000 victims, 22 houses damaged, 3 highways affected. Complete loss of dozens of hectares of vegetables and grains 16.4 1986 Fire Veracruz 1.5 1987 Snowfall Mexico City, States of Mexico and Hidalgo- 6 people killed. Damage to water infrastructure, electricity transformers, 1,300 hectares of crops, and vehicles (as a result of traffic accidents) 0.3 1988 Industrial explosion Ixhuapan, Veracruz 20 people killed, 15,000 left homeless, 18,000 barrels of oil burned 0.3 Forest fires 500,000 hectares of forest burned 1,250.0 Drought in agricultural zones Throughout the country 542,000 hectares of crops completely destroyed, 439,000 hectares partially damaged 168.4 Hurricane Gilbert Four states and Nuevo Leon 225 people killed, 46 injured, 139,000 evacuated, 51,000 victims, 9,739 homes damaged; 364,000 hectares of agricultural land damaged. Buildings, roads, electricity lines, highways, crops, shipping, communications, and urban services also affected 76.0 Other hurricanes Throughout the country 417 people killed, 106 injured, 192,000 victims, 31,000 homes damaged; 1.35 million hectares of agricultural land completely destroyed, 1.73 million hectares partially damaged; 74,683 head of cattle lost 597.6 (table continues on next page) 44 Managing Disaster Risk in Mexico: Market Ince-ttivesfor Mitigation Investmient Table A2 (continued) Reporredi direct losses Year/er'ur/oerc0on Casualties ar,d damage (millions rU.S. dollars) Frost Tamaulipas, Nuevo Leon, Coahuila, 50 percent of coffee crop, 250 tons Chihuahua, Veracruz, Puebla, and Tlaxcala of oranges, and 2,000 head of cattle lost 0.6 1989 80,000 hectares of agricultural land damaged c48.0 Forest fire Quintana Roo 1990 Frost Tamaulipas,Veracruz, Baja California 10,000 hectares of crops damaged in Tamaulipas 1.2 Hurricanie Diana Veracruz, Hidalgo, and Puebla 139 people killed, 56 missing, 75,000 victims. Highways and train tracks blocked in 6 states; 40,000 hectares of agricultural land lost. 90.7 Flood Chihuahua 200 people killed, 5,000 victims, 700 houses damaged 2.5 Fmosts Nuevo Leon, Baja California, Durango, 52 people killed, 500 families affected 0.1 Coahuila, Sonora, Mexico, Tamaulipas, Veracruz, and Chihuahua 1991 Industrial explosion 6 people killed, 329 injured 1 50.0 Veracruz Floods (dam breaks) Zacatecas 5 people killed, 12,000 cut off from communication, 10,500 victims, 2 bridges destroyed 0.7 W\inter stormns Baja California Sur, Sonora, Sinaloa, 153 communities flooded, 40,000 victims, 40 percent and Chihuahua of streets in La Paz, Baja California, destroyed; 50,000 agriculture hectares completely lost, 25,000 pigs lost, 160 kilometers of irrigation canals destroyed, damage to bridges (road and railroad) and highways. 16.8 1992 lndustrial explosion Guadalajara 212 people killed, 1,480 injured, 13,930 victims, 1,425 homes totally destroyed, 150 homes with minor damage; 450 businesses destroyed, contents of 802 households destroyed, 637 vehicles destroyed, 8 kilometers of streets destroyed, 300,000 cubic meters of urban . area affected 65.0 Rain and snow Sinaloa Crops lost 0.2 Tornado Reynosa More than 3,000 homes flooded 10.5 Tornado Comitan de Dominguez, Chiapas 375 homes damaged 1.3 Hurrican.e Winif-eci Manzanillo 3,000 victims, 2,000 hectares of plantain crops damaged 8.0 Annex 2 45 Table A2 (continued) Repo ntel doer;ct losses Yh rleveolt/eioraioln Cestwittes andLi dalagec (oullians of u.S. dollars, Wilnter storms Nayarit 64 people killed, 70 premises flooded, 100,000 people left homeless; 104,000 hectares of agriculture and several highways damaged 78.0 Hall storms Ttlaxcala 11 people injured, 2,000 families affected, 2,000 hectares of crops damaged or destroyed 0.2 Hail storms Guerrero and Tlaxcala 30,000 hectares of crops and orchids destroyed 2.3 Frost Zacatecas 350,000 hectares of agricultural land damaged; 27.0 Eruptions of Popocatepeti volcano (1993-98) 5 people killed, 26,000 evacuated from 23 communities. Damage from falling ash to pastures and urban centers 12.0 Hurricane Gert Veracruz, San Luis Potosi, Hidalgo, 7 people killed, 50,000 victims, 5,000 homes flooded, Guerrero, Tamaulipas, and Quintana Roo 70,000 people cut off from communications 18.1 N in.cr storms Baja California (January) 20 people killed, 10,000 victims. Communications and agricultural zones severely damaged 32.0