Coping with losses: Options for Disaster Risk Financing in Brazil Coping with Losses: Options for Disaster Risk Financing in Brazil II Coping with Losses: Options for Disaster Risk Financing in Brazil This report was prepared by World Bank staff. The findings, interpretations, and conclusions expressed here do not necessarily reflect the views of The World Bank, its Board of Executive Directors, or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The bound- aries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Rights and Permissions Editors: Joaquin Toro, Michel Matera, Fernanda Senra de Moura and Frederico Ferreira Pedroso Photos: Joaquin Toro Design: Carlos Eduardo Peliceli da Silva The World Bank encourages dissemination of its knowledge, this work may be reproduced, in whole or in part, for noncommercial purposes as long as full attribution to the work is given. The material in this work is subject to copyright. © 2014 International Bank for Reconstruction and Development / International Development Association or The World Bank 1818 H Street NW Washington DC 20433 1 Coping with Losses: Options for Disaster Risk Financing in Brazil Table of Contents Acknowledgements 4 Abbreviations and Acronyms 5 Executive Summary 6 Fiscal Risk Assessment 8 Budgetary Process Review 8 Funding Gap Analysis 8 Disaster Insurance Market Overview 9 Discussing a National DRFI Strategy 9 Chapter 1. Introduction 10 Objectives of the Study 11 Chapter 2. Financial Disaster Risk Assessment in Brazil 12 2.1 Exposure to natural hazards in Brazil 13 2.2 Looking for the unaccounted risks: damage and losses from four recent major disasters 14 2.3 Damage assessment systems and data quality issues 17 2.4 Preliminary Fiscal Risk Assessment 18 Chapter 3. Fiscal Management of Natural Disasters 20 3.1 Federal Government’s ex-post disaster risk financing practices and arrangements 21 In-year budget reallocation 21 Taxation 23 International assistance 23 Other ex-post disaster funding resources 23 3.2 Federal Government’s ex-ante disaster risk financing practices and arrangements 24 Annual budgeting 24 Multi-year reserves 24 Contingent loans 24 Risk transfer 24 3.3 Local governments’ disaster risk financing practices and arrangements 27 Annual budgeting 27 Multi-year reserves 29 Taxation 29 Contingent loans 29 2 Coping with Losses: Options for Disaster Risk Financing in Brazil 3.4 Funding gap analysis 29 Chapter 4. Insurance Market Overview 34 4.1 Non-life insurance market overview 35 4.2 Agricultural insurance 37 Importance of the agricultural sector 37 Penetration of agricultural insurance 39 Government support programs 40 Optimizing the GoB programs through a DRFI strategy 42 4.3 Property insurance, microinsurance and temporary cash transfers 43 Property insurance market 43 Microinsurance 45 Temporary cash transfers 46 Chapter 5. Options for Disaster Risk Financing in Brazil 48 5.1 Implementing a national disaster fund 51 5.2 Optimizing subsidies and temporary cash transfer programs 53 5.3 Improving Financial Risk Assessment analysis and developing catastrophe risk models 53 5.4 Enhancing disaster data management systems and damage and loss assessment procedures 54 References 56 Annex 1. State and Municipal Disaster Funds 58 Annex 2. Funding Gap Analysis 59 Annex 3. Assessing the Risk Profile of Brazil 60 Procedure Proposed 60 Results Obtained 62 Complimentary Data: Historical Aggregate Annual Losses, current and adjusted 65 Annex 4. Disaster Related Public Expenditure 69 A.4.1 An introduction to the Brazilian budgetary systems 69 A.4.2 Disaster related public expenditure 70 Annex 5. Disaster Damage Assessment System 75 3 Coping with Losses: Options for Disaster Risk Financing in Brazil Acknowledgements This report was produced by a team led by Joaquin Toro and Michel Ma- tera (GSURR, World Bank), and comprising Fernanda De Moura (Consultant, GSURR), Frederico Pedroso (Analyst, GSURR), Vinícius Botelho (Consultant, GSURR), Rafael Proença (Consultant, GSURR), José Ángel Villalobos (Senior Insurance Specialist, GFDMR and GFDRR, World Bank), and Ésio Moreira (Consultant). Ricardo Zapata (Regional Advisor and Focal Point for Disaster Evaluation at ECLAC), and Hannah Joy Yi (Policy Analyst, Disaster Risk Fi- nancing & Insurance Program, GFDMR and GFDRR) provided guidance and comments that were crucial to the preparation of this report. The report greatly benefited from data and information provided by the Ministry of National Integration – National Secretariat of Civil Defense, Min- istries of Planning, Transports, and Finance, State Secretariats of Planning, Finance, Education, Health, and Civil Defense. The team is thankful to the peer reviewers Marc Forni (GFDMR, World Bank), Diego Arias (GFADR, World Bank) and Nikhil Da Victoria Lobo (SwissRe). The team also thanks Niels Holm-Nielsen (GSURR, World Bank) and Rafael Chelles Barroso (GMFDR, World Bank) for their comments. The team gratefully acknowledges funding from the Brazil Country Man- agement Unit and technical support from the Disaster Risk Financing and Insurance (DRFI) Program of the Global Facility for Disaster Reduction and Recovery (GFDRR). 4 Coping with Losses: Options for Disaster Risk Financing in Brazil Abbreviations and Acronyms AAL Average Annual Loss AVADAN Damage Assessment Form CEDEC Civil Defense State Coordinating Office CEPED University Study and Research Center on Disasters CNSP National Council of Private Insurance COMDEC Coordination Office of the Municipal Civil Defense CONAB National Supply Company CONDEC National Council on Civil Defense CORDEC Civil Defense Regional Coordinating Office CEMADEN Center for Monitoring and Early Warning of Natural Disasters DAC Damage Assessment Committee DaLA Damage and Loss Assessment DPL Development Policy Loan DRM Disaster Risk Management DRFI Disaster Risk Financing and Insurance DRR Disaster Risk Reduction ECLAC Economic Commission for Latin America and the Caribbean FESR Crop Insurance Stabilization Fund FONDEN Mexico’s Natural Disaster Fund FUNCAP Special Fund for Public Calamities GDP Gross Domestic Product GIS Geographic Information System GoB Federal Government of Brazil LEC Loss Exceedance Curve LOA Annual Budgetary Law MAPA Ministry of Agriculture, Livestock and Food Supply MI Ministry of National Integration MPCI Multi-peril Crop Insurance MPOG Ministry of Planning NOPRED Preliminary Disaster Notification Form PML Probable Maximum Loss PPA Multi-Year Plan PRED Disaster Response Program PROAGRO Rural Activity Guarantee Program PRONAMP National Rural Support Program for Small Producers PRRC Program of Preparedness for Emergencies and Disasters (PPED) PRVD Disaster Prevention Program PSR Government Premium Subsidies Program for Agricultural Insurance RAP Carry-Over (Restos a Pagar) SEDEC National Secretariat of Civil Defense S2ID Disaster Information System SINDEC National Civil Defense System SOF Secretariat of the Federal Budget SUSEP Superintendence of Private Insurance 5 Coping with Losses: Options for Disaster Risk Financing in Brazil Executive Summary This study presents options for disaster risk financing in Brazil, draw - ing from international experience. The study presents a series of complementary options for disaster risk financing, based on a pre- liminary fiscal risk analysis and a preliminary review of the current budget management of natural disasters in Brazil. It benefits from the international experience of the World Bank, which has provided assistance to several countries on the design and implementation of sovereign disaster risk financing strategies (e.g. Mexico, Colombia, Peru, Indonesia, Vietnam, Philippines, Pakistan and the Caribbean island states). This experience is tailored to the extensive risk pro- file and institutional, social and economic characteristics of Brazil, as well as to the availability of relevant data. 6 Coping with Losses: Options for Disaster Risk Financing in Brazil Executive Summary Brazil is exposed to several adverse natural events, the most recurrent being of a hydro-meteorological nature, namely drought, severe rainfall and landslides. The high population density in urban areas is the main determinant of vulnerability to natural hazards in the country, largely driven by the poor land use and planning that characterizes Brazilian cities. Since 2008, extreme weather events have caused several major disasters in Brazil, leading State and Federal Governments to prioritize the strengthening of Di- saster Risk Management (DRM) practices in their agendas. While the number of recorded disasters has grown significantly, both the human and economic losses associated with these occurrences have also increased over the past five years. However, since the country is exposed to recurrent small-scale events1, whose negative financial impacts tend to be localized in specific cities or regions, the impacts of these events have been overlooked for decades. While the cumulative effects generate major economic and fiscal losses, little is known about the extent to which these disasters affect the economies and welfare of the people. To address this lack of information, the World Bank team, in partnership with State Governments, carried out a series of case studies which estimated that disaster-related costs of four major events2 alone totaled ap- proximately R$ 15.3 billion – R$ 9.4 billion in damages (direct costs) and R$ 5.9 billion in losses (indirect costs3). The case studies also suggested that fiscal impacts were significant and that the GoB (Federal Government of Brazil) plays a major role in supporting disaster response in the affected communities, since government liability associated to such events can be substantial at the local level. Within this context, the purpose of this study is to initiate a debate on the fiscal impacts of natural disasters in Brazil and to propose the next steps towards developing a broader Disaster Risk Financing and Insurance (DRFI) strategy that strengthens the financial resilience of Federal and State Governments. The main compo- nents of the study are: • A preliminary financial risk assessment of natural disasters at the national level, including assessment of private and public costs; • A review of the current budgetary process used for financing responses to natural disasters in Brazil, both at the Federal and State levels, which includes the development of series of fiscal data on disaster public budgetary resources and expenditures during the past 10 years; • A retrospective funding gap analysis that takes into account budgetary resources before and after real- locations, under different cost scenarios; • A disaster insurance market overview focusing on the role of private agricultural insurance as a risk transfers mechanism and on the use of temporary cash transfers by the GoB to deal with disaster effects among vulnerable populations in both urban and rural areas; • A discussion on the next steps that government bodies such as the Ministries of Planning, Finance and National Integration could take to develop a broader DRFI strategy in Brazil. 1 Following the GAR2013 report, an extensive risk profile is characterized by small-scale, localized, but frequent events. These events can also be spread over a large area, affecting local communities but not disrupting the country as a whole. In contrast, countries with intensive profile risk are those that might experience a severe catastrophe of national relevance, but less frequently. 2 The 2011 floods and landslides in Rio de Janeiro, the 2010 floods in Pernambuco and Alagoas and the 2008 floods in Santa Catarina. 3 Total costs in this case were estimated using information available at the time of the study and, for this reason, the final results may represent a slight underestimation of the real-life economic impact. 7 Coping with Losses: Options for Disaster Risk Financing in Brazil in Brazil is not guided by an integrated national disaster Fiscal Risk Assessment risk financing and insurance strategy. The GoB primar- ily relies instead on ex-post disaster risk financing The use of fiscal risk assessment tools can support mechanisms. Budgetary reallocation after the occur- policy makers to design a disaster risk financing rence of natural disasters has been the most common and insurance strategy better suited to the coun- budgetary practice, since annual budgeting is usually in- try’s hazard profile. By using partial historical data on sufficient to meet disaster response requirements. disaster losses, it was possible to carry out a preliminary Multi-year reserves are not used at the national fiscal risk assessment of Brazil. The total Average An- level, and some states have already established nual Losses (AAL) is US$3.9 billion (R$8.9 billion) and, their own disaster funds. In parallel to the ongo- assuming that the government´s liability amounts to ing discussions in the National Congress regarding the 30% to 40%4 of total costs, the public AAL is be- regulation of the established (but not operational) Na- tween US$1.2 billion (R$2.7 billion) and US$1.7 tional Calamity Fund (FUNCAP), the development of billion (R$3.9 billion). a national disaster risk financing and insurance strategy An event with a 10 year return period could be enough might help optimize the GoB’s financial management to generate public losses as high as US$3.7 billion (R$8.7 of disasters. billion) and an event with a 15 year return period could result in total losses as high as US$ 12.8 billion (R$29.4 billion), with public costs amounting to as much as Funding Gap Analysis US$5.1 billion (R$11.7 billion). Recovery and reconstruction gaps exacerbate the nega- tive impacts of disasters on socioeconomic outcomes Budgetary Process Review and for this reason we estimate post-disaster needs funding gaps in Brazil from 2006 to 2010. In this pe- The current financial management of natural disasters riod, the financing gaps in a given year was an av- 4 According to the case studies discussed below, public costs correspond to about 50% of total costs. However, given the omission of some elements of the private sector due to the lack of available data, a 30-40% share of public costs is assumed in the fiscal risk assessment. 8 Coping with Losses: Options for Disaster Risk Financing in Brazil erage of R$1.4 billion, or 30% of the estimated while promoting the right incentives for prevention and average disaster-related government liability. adequate conditions for resilient reconstruction. In or- der to improve the current disaster risk financing ap- These numbers are based on a series of assumptions, proach, government bodies such as the Ministries of but they corroborate anecdotal evidence from the field, Planning, Finance and National Integration may consider suggesting that improving both disaster risk financing ar- the following: rangements and damage and loss assessment systems are needed for resilient and timely reconstruction. • The Ministry of Planning may be central in the operationalization of FUNCAP by com- pleting the standard operating procedures and Disaster Insurance by formalizing the funding flow arrangements be- tween FUNCAP and local institutions, such as the Market Overview calamity funds at state level. One of the main features of the Brazilian disaster • The Ministry of Finance may coordinate an insurance markets is the high level of public sector assessment of the GoB subsidy programs for support for risk management in the agricultural natural hazard insurance, particularly for ag- sectors. In 2003, the GoB established a program of riculture (PSR). Updating these schemes after subsidies for private agricultural insurance and penetra- defining the optimal relationship between insur- tion rates have since then increased significantly. In ad- ers, banks, beneficiaries and the Government may dition, in the occurrence of a disaster, temporary cash translate into expanded coverage at reduced costs. transfers are made to farmers not covered by private insurance, with a focus on smaller producing units. • Regarding temporary cash transfers, both in the agriculture and housing sectors, better However, despite being relatively well established, understanding the contingent liability (AAL the financial management of these initiatives has and PML) for GoB cash transfer systems to sup- been challenging and available relevant information port vulnerable populations such as the PROAG- about Brazil’s hazard profile is not being used to pro- RO, Garantia Safra, Bolsa-Estiagem, and the cash mote adequate budgetary planning and pricing. In- transfers for temporary housing programs, may corporated in a broader DRFI strategy, such programs serve as the basis for increasing cost-effectiveness. could be more cost-effective. The possibility of consolidating these payments into a stand-alone risk pool managed through risk Discussing a National retention or risk transfer schemes may also be considered. DRFI Strategy • Planning and finance institutions may con- The main goal of a DRFI strategy is to increase finan- sider carrying out an improved fiscal risk as- cial response capacity while reducing the fiscal burden sessment, potentially focused at the State level generated by the government´s liabilities associated to and in this way use the risk metrics produced to in- natural disasters. The World Bank has developed a DRFI form how transfers to local governments should be framework that takes account of different layers of risk managed. Developing catastrophic risk models that a country might be exposed to, and considers the may also be the key to complementing the fiscal most appropriate financial instruments for financing re- risk analysis based exclusively on historical losses. sponses to disasters. • The Ministry of National Integration adopt- Given Brazil’s hazard profile, the low risk layer (i.e. ing an objective damage and loss assessment those associated to small-scale, but more frequent methodology may provide the basis for (i) a clear events) could be prioritized, and for this a risk retention guideline for allocating disaster response funds and strategy based on reserve funds, budgetary allocations (ii) the data collection efforts needed to keep track and contingent funds is most appropriate. However, it is of the historical material losses - inputs that are re- important to work up the details of such a risk retention quired for the appropriate design and updating of strategy in order to avoid financing gaps and ensure the an evolving national DRFI strategy. cost effective and timely use of disaster response funds, 9 Coping with Losses: Options for Disaster Risk Financing in Brazil Chapter 1. Introduction B razil is exposed to several adverse natural events, the most recur- rent being of a hydro-meteorological nature, namely drought, severe rainfall, and landslides. The high population density in urban areas is the main determinant of vulnerability to natural hazards in the country, large- ly driven by the poor land use and planning that characterizes Brazilian cities. Since 2008, extreme weather events have caused several major disrup- tions in Brazil, leading to both State and Federal Governments to priori- tize the strengthening of Disaster Risk Management (DRM) practices in their agendas. While the number of recorded disasters has grown sig- nificantly, the associated human and economic losses have also increased over the past five years. Despite Brazil´s exposure and vulnerability to adverse natural events and re- curring disasters, the impacts have not been subject of systematic, in depth, studies. Therefore little is known about the fiscal effects of natural disasters. A series of case studies by the World Bank, in partnership with State Governments, indicates that the disaster-related costs of only four ma- jor events that occurred between 2008 and 20115 totaled approximately R$15.3 billion – R$9.4 billion in damages (direct costs) and R$5.9 billion in losses (indirect costs)6. 10 Chapter 1. Introduction Coping with Losses: Options for Disaster Risk Financing in Brazil The total cost of the 2011 Rio de Janeiro landslides (R$4.78 billion), for example, was equivalent to 36.20% of the regional GDP (Gross Domestic Product). The 2008 floods in Santa Catarina occurred in an area with a high concentra- tion of assets. In Pernambuco and Alagoas, the floods mainly affected vulnerable and low-income populations, further complicating the recovery process. In the specific case of Alagoas, the impacts were estimated at 8% of that state’s GDP, illustrating the disruptive impact of disasters on state and local economies. The analysis also suggests that fiscal impacts were considerable. Public damages and losses due to flooding in Santa Catarina were estimated at approximately R$2 billion, roughly 20% of the state’s tax revenue in 2009. Similar figures were also recorded in the states of Alagoas and Pernambuco, where public damages and losses corresponded to 14% and 16% of each state’s net revenues in 2010 respectively. Finally, the state of Rio de Janeiro had to contend with public damages and losses of R$3.1 billion, equivalent to 8% of current revenues and 10% of the state’s tax revenue in 2011. However, data on fiscal impacts are still very limited and often recorded by government institutions that do not always interact with each other. As a result, neither Federal nor State governments have a comprehensive understanding of how disasters affect government budgets and jeopardize fiscal balances. The GoB instead relies primarily on ex-post disaster risk financing mechanisms. Budgetary reallocations after natural disasters occur have been the most common budgetary practice given that annual budgeting is usually insufficient to meet disaster response needs. Multi-year reserves are not used at the national level, and some states have established their own disaster funds. Budgetary execution procedures are still cumbersome, especially in the case of local govern- ments. As a result, the GoB has recurrent financing gaps for recovery and reconstruction. The current financial instruments could be further optimized to better meet financing needs in the aftermath of natural disasters. Developing a national disaster risk financing and insurance strategy would strengthen the GoB’s fis- cal resilience as well as that of local governments, closing funding gaps that amplify the effects of natural disasters on economic activity and welfare, while simultaneously promoting prevention and resilient reconstruction. Objectives of the Study In the light of the above, the purpose of this study is • A disaster insurance market overview, focusing to stimulate dialogue on the fiscal impacts of natural di- on the role of private agricultural insurance as a sasters in Brazil and to propose some options for the risk transfers mechanism and on the use of tempo- development of a broader disaster risk financing and in- rary cash transfers by the GoB to handle disaster surance strategy to strengthen the financial resilience of effects on vulnerable populations, in both urban governments at the Federal and local levels. The main and rural areas; components of the study are: • A discussion on the next steps that government • A preliminary financial risk assessment of natural bodies such as the Ministries of Planning, Finance disasters at the national level, including both pri- and National Integration could take to develop a vate and public costs; broader DRFI strategy in Brazil. • A review of the current budgetary process for This report consists of five chapters including this In- financing response to natural disasters in Brazil, at troduction. Chapter 2 provides a brief presentation of both Federal and State levels. This includes the de- Brazil’s hazard profile and a preliminary fiscal risk assess- velopment of series of fiscal data on disaster public ment. Chapter 3 presents an overview of existing post- budgetary resources and expenditures over the disaster funding mechanisms at the Federal and State lev- last 10 years; els and suggests a recovery and reconstruction financing gap analysis. A private disaster insurance markets review • A retrospective funding gap analysis that takes is presented in Chapter 4. Finally, Chapter 5 reviews the into account budgetary resources before and after main issues that Brazil should tackle to enhance its disas- reallocations, under different cost scenarios; ter risk financing and insurance strategy and suggests the next steps to be taken to stimulate this debate. 5 The 2011 floods and landslides in Rio de Janeiro, the 2010 floods in Pernambuco and Alagoas and the 2008 floods in Santa Catarina. 6 Total costs in this case were estimated using information available at the time of the study and, for this reason, the final results may represent a slight underestimation of the real-life economic impact. 11 Coping with Losses: Options for Disaster Risk Financing in Brazil Chapter 2. Financial Disaster Risk Assessment in Brazil R ecently released data show that Brazil is exposed to recurrent and small-scale disasters, localized in cities or regions, deemed with exten- sive risks. Such a “hidden risks” risk profile is one of the main reasons for the impacts of these frequent events to have been overlooked for decades. Little has been known to date about the extent to which disasters affect the economy and welfare. The World Bank assessed four recent extreme adverse natural events in Brazil and estimates US$6.6 billion (R$15.3 billion) in damage and losses, most of it concentrated on the housing sector and on the poorest and most vulnerable populations. The use of fiscal risk assess- ment tools can support policy makers and planners to design financial plans better suited to a country’s hazard profile. Using partial historical data on di- saster losses it was possible to carry out a preliminary fiscal risk assessment of Brazil which suggests that the country may expect annual losses in excess of US$3.9 billion. The total Average Annual Loss (AAL) is US$3.9 billion (R$8.9 billion) and, under the assumption that the government liability cor- responds to 30% to 40% of total costs, the public AAL is between US$1.2 billion (R$2.8 billion) and US$1.7 billion (R$3.9 billion). An event with a 5 year period of return could be enough to generate public losses as high as US$ 1.9 billion (R$4.5 billion). 12 Chapter 2. Financial Disaster Risk Assessment in Brazil Coping with Losses: Options for Disaster Risk Financing in Brazil 2.1 Exposure to natural hazards in Brazil H igh magnitude earthquakes, tsunamis and hurricanes are not part of Brazilian history. Floods, landslides and droughts are routine. Despite the known fact that Total2.1 Figure Total Number Number of Recorded of Recorded Disasters in Brazil by Year Disasters in Brazil by Year small-scale and highly frequent disasters can have a nega- tive impact on many Brazilian cities, it was only in 2011, 3000 when the Federal Government launched the Brazilian Natural Disaster Atlas (MI, 2012), that the country’s haz- ard profile could be subjected to an in-depth analysis. Based on this publication, figure 2.1 points to an upward 2000 trend in the number of officially recorded disasters. This could be a result of a higher frequency of adverse natural events and/or of an improvement in recording systems. From the total official number of recorded disasters, 1000 more than half are due to droughts, a type of disaster that affects mainly the Northeastern and the Southern regions of the country. This hazard is associated to only 10% of the people killed, but responds to 50% of the affected population. Damage and losses caused by droughts are particularly hard to assess, but several 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 studies have found evidence of major impacts on agri- cultural and related sectors, as well as on health and Figure 2. Total Number of Recorded Disasters in Brazil by Year Source: Atlas Brasileiro de Desastres Naturais, CEPED-UFSC education outcomes. The latter are important channels through which natural adverse events jeopardize the long-term development of affected communities. Figure 2.2 Spatial Distribution of Natural Disasters in Brazil by Region Flash floods are also frequent in Brazil, especially in the Southern, Southeastern and Northeastern regions. The fatalities associated with floods correspond to 43% of the total number of people killed by disasters, but the number of people made homeless by floods is also alarm- ing. Since 1991, more than 1.3 million people have had to leave their homes in the aftermath of disasters in Brazil. Also associated with heavy rain, the number of offi- cially recorded landslides has increased more than 20 times since the 1990s. This kind of event is usually as- sociated with costly damage to private properties and public infrastructure, which can be worse if a disaster hit areas with a high concentration of high value assets. In short, droughts, floods, and landslides are the most frequent adverse natural events in Brazil and, while the occurrence of catastrophic events is not of significant importance to the country’s hazard profile (unless there is a string of such events in short period of time), the Drought Cyclone and High Speed Winds Frost River Erosion Flash Flood Tornado Bush Fire Linear Erosion exposure to major risks jeopardizes local development Flood Hail Mass Movement Marine Erosion and imposes a heavy burden on the poorest and more vulnerable populations, even though this burden might Source: Atlas Brasileiro de Desastres Naturais, CEPED - UFSC, 2012 seem immaterial in the countrys overall economy. 13 Coping with Losses: Options for Disaster Risk Financing in Brazil 2.2 Looking for the unaccounted risks: damage and losses from four recent major disasters A common problem to countries with such an extensive risk profile7 is that it can lead governments to ignore Disaster Risk Management (DRM) practices in their national agendas. As a result, assessment systems and data management are not incorporated in disaster response policies. This lack of basic infrastructure for developing private risk mar- kets and appropriate planning tools to support sovereign risk financing strategies severely undermines the financial resilience of public and private agents, causing reconstruction delays and increasing the negative effects of disasters on economic development. Figure 2.3. Damage and Loss Assessment of Four Major Recent Events Damages and losses Damages by sectors Housing 38% Transport Loss Industry Trade Agriculture Education Health Environment Sanitation Energy Tourism 62% Comunications Damages 0 500 1000 1500 2000 2500 3000 3500 R$ (million) Public and Private sectors Losses by sectors Housing 48% Transport Private Industry Trade Agriculture Education Health Environment Sanitation Energy Tourism 52% Comunications Public 0 500 1000 1500 2000 2500 3000 3500 R$ (million) Source: World Bank Estimates 7 Following the GAR2013 report, an extensive risk profile refers to that that is characterized by less severe, but frequent events. These events can also be spread over a large area, affecting local communities but not disrupting the country as a whole. In contrast, countries with an intensive profile risk are those that might experience severe, but less frequent catastrophes of national relevance. 14 Coping with Losses: Options for Disaster Risk Financing in Brazil In Brazil, although the primary data sources on which The abovementioned events are not an exhaustive list the Brazilian Natural Disasters Atlas was compiled in- of the disasters of the past four years. Nonetheless, clude information on material damage and losses8, the they are representative of the economic and human damage assessment systems used before 2011 put losses caused by natural hazards in Brazil. They were substantial difficulties for disaster data management. A also chosen for complete Damage and Loss Assessment comprehensive historical damage and losses dataset as- (DaLA) studies8 given the relative availability of infor- sociated to major risks is not therefore currently avail- mation related to these events. According to the DaLA able. The absence of a workable dataset constitutes a studies conducted in all four of the affected states, disas- challenge for all kinds of decision makers: public policy ter-related costs totaled approximately R$15.3 billion – makers, entrepreneurs, families, etc. R$9.4 billion in damages (direct costs) and R$ 5.9 billion in losses (indirect costs)9. Within this context, the World Bank, in partnership with the Ministry of National Integration (MI), ECLAC (Eco- Interestingly, total costs were evenly distributed between nomic Commission for Latin America and the Carib- the public and private sectors (Figure 2.3). Property of bean), and the Brazilian State Finance and Planning Sec- damage and losses were assigned on a case-by-case basis, retariats, assessed damages and losses from four major depending on the sector and state under assessment; pri- disasters that beset Brazil since 2008. The case studies vate sector losses and damages were harder to estimate, aim to assess damage and losses from a sector perspec- due to the lack of data. Overall, however, the studies sug- tive, in accordance with ECLAC’s Damage and Loss gest that in Brazil the public sector tends to absorb a large Assessment (DaLA) methodology, identifying the main share of private sector costs during disaster relief and the channels of impact and outlining an agenda on which an reconstruction process. effective and efficient disaster risk management (DRM) strategy can be built. Beyond the government’s explicit liabilities (such as infrastructure repair), post-disaster relief must also ad- In January 2011, the state of Rio de Janeiro experienced dress liabilities beyond the concerns of the DaLA meth- heavy rainfall, eventually culminating in a series of flash odology. For example, the government has a moral re- floods and landslides. Resulting in around 1000 causali- sponsibility in post-disaster scenarios to support home ties in 7 municipalities, the event was quickly deemed the reconstruction for poor families. Studies also show that worst disaster in Brazilian history. A few months earlier, the public sector will frequently offer subsidized credit in June 2010, the states of Alagoas and Pernambuco suf- to local businesses, industry and agriculture as a means fered severe flash flooding during the worst rainy season of restoring a degree of normality to traumatized com- of the past 20 years. In late 2008, floods and landslides munities. The task of effectively limiting the private indi- killed 110 people in the state of Santa Catarina, and dam- rect economic impacts of disaster, in other words, rests aged the Port of Itajaí and the nearby Bolivia-Brazil gas pipeline. Table 2.1. Disasters and economic losses: four major disasters in Brazil Year Event Number of Victims Total Cost (R$ bi) Total Losses (% State GDP) 2008 Floods and Landslides in Santa Catarina 110 5.32 2.67% 2010 Floods in Pernambuco 20 3.37 4.30% 2010 Floods in Alagoas 36 1.85 8.72% 2011 Floods and Landslides in Rio de Janeiro 1,000 (approx. ) * 4.78 1.35% (Região Serrana) * Given that there remain missing persons in the Região Serrana, these individuals cannot be counted as victims. 8 In the Insurance sector, damages are often called “physical damage”, i.e. the value of lost physical assets, while “losses” most closely corresponds to financial losses due to the natural event, such as the interruption of businesses and the need to secure emergency funds. Following ECLAC’s Damage and Loss Assessment methodology, the interpretation is the same. The term “damage” refers to direct costs such as the loss of assets that have been partially or totally destroyed. The term “loss” refers to indirect losses caused by disasters. Examples of indirect losses are the income lost due to equipment that that has been destroyed and requires replacing, reduced agricultural output resulting from flooded land, and so on. 9 Total costs in this case were estimated using information available at the time of the study and, for this reason, the final results may represent a slight underestimation of real-life economic impact. 15 Coping with Losses: Options for Disaster Risk Financing in Brazil largely with government. According to the studies below, the 2008 Santa Catarina floods and landslides generated the highest estimated costs (R$ 5.32 billion10), while the assessment conducted for the Rio de Janeiro event in 2011 potentially underestimated its total impact by omitting a handful of relevant sectors due to a shortage of data. The authors of this report suspect that if this information had been included, Rio de Janeiro may have been – in terms of human and economic losses –the most costly and damaging disaster. On one hand, the area affected in Rio de Janeiro was geographically smaller than those of the other events and the costs of re-building the region’s physical infrastructure was significantly higher. Nevertheless, the total cost of the disaster (R$ 4.78 billion) exerted only a small impact on the overall state GDP – although its impact on the Região Serrana was severe, equivalent to 36.20% of the regional GDP (when one considers that the Rio disaster affected 7 of the Serrana’s 11 municipalities, however, this is perhaps unsurprising). On the other hand, the floods in Santa Catarina occurred in an area with a high concentration of valuable assets (ex. the Port of Itajaí and a variety of federal and state highways). In Pernambuco and Alagoas, the total costs from flash floods were lower than those estimated in Santa Catarina, but the impacts mainly affected vulnerable and low-income populations, further complicating the recovery process. In the case of Alagoas specifically, the impacts were estimated at 8.7% of the state’s GDP , illustrating the disruptive impact of disasters such as this on state and local economies. Table 2.2 presents a comparison of the four events in which total losses vis-à-vis state GDP serve to illustrate the economic impact of each disaster. Table 2.2 shows that the government liability generated by these events is too high compared to the budgetary alloca- tions to the Civil Defense related programs at those states. This supports the claim that local governments lack the capacity to respond on their own and that intervention by the Federal Government was of paramount importance. Table 2.2. Recent damage and loss assessment for selected Brazilian states Year Event Government State Civil Defense Liability (R$ bi) Initial Budgetary Al- location* (R$ bi) 2008 Floods and Landslides in Santa Catarina 2.3 0.03 2010 Floods in Pernambuco 2 0.23 2010 Floods in Alagoas 0.605 0.05 2011 Floods and Landslides in Rio de Janeiro (Região Serrana) 3.2 0.76 *Santa Catarina: 2008 and 2009; Alagoas and Pernambuco: 2010 and 2011; Rio de Janeiro: 2011 e 2012. While one could argue that in the aftermath of a disas- (roughly 50,000 low-income homes were ruined), high ter the state governments could reallocate budgetary losses stemmed from the very high costs of retrofitting resources and are not limited to the resources provi- or replacing existing infrastructure to reduce communi- sioned in their annual budgets, figure 2.4 shows that ty vulnerability. For example, in the northeast, dams are the government liability corresponding to each of these often engineers’ main recourse for reducing flooding, events was equivalent to a relevant share of the states’ while recovery efforts in Rio de Janeiro concentrate on financial revenues and that the realignment of resources soil nailing and hill stabilization to prevent future land- towards disaster response would imply major disrup- slides. Regardless of the approach, both are costly and tions in their regular systems. require specific and highly technical engineering-based interventions. As for the impacts by sector, housing – during all four events – was by far the most affected. Beyond those After housing, the transport sector was the second dwellings that were partially or totally destroyed most affected by direct impacts, provoking lasting physi- 10 Value adjusted for 2011 prices. 16 Coping with Losses: Options for Disaster Risk Financing in Brazil Figure 2.4. Damage and losses versus state fiscal revenues* Figure 1.6 Damage and losses versus state fiscal revenues* 70% 60% 50% 40% 30% 26,52% 20,43% 21,15% 20% 10% 9,71% 0% Government Liability/Total State Revenues (%) 70% 61,61% 60% 50% 47,94% 40% 30% 28,01% 20% 16,98% 10% 0% Government Liability/Total Transfers from Federal Government (%) *Fiscal data refer to the relevant year given the date of the ocurrance of the disaster in each state. Source: World Bank estimates based on official figures and National Treasury Secretariat. cal damage to high value network assets and indirect after a disaster. Each municipality must submit a damage costs that caused significant travel delays, impacting lo- and loss assessment (based on official forms) within 5 cal economic activities in the region. Note in Figure 2.3 days after the occurrence of the event. Before 2012, that the estimated proportion of damages and losses hard copies of such forms had to be submitted by the for the transport sector is highly unbalanced. (Also note affected municipalities to the State and Federal Govern- that this data may be partially incomplete: exact figures ments. These files were recently digitalized and since on indirect impacts on the transport sector are difficult 2012 the system was upgraded to an online platform, to obtain and rarely identified in research). the Integrated System on Disaster Information (S2ID), and the AVADANS (Damage Assessment Forms) was In short, well over a third of total estimated costs stem replaced by a simpler version (details of the past and from the housing and transport sectors. Such findings current damage and losses assessment systems are pre- underscore the urgent need to target both sectors sented in annex 5). through use of a practiced, holistic DRM approach for reducing community vulnerabilities and potential losses These forms are the primary source of information of in case of disaster. most historical datasets on the occurrence and costs of disasters in Brazil, but their role in the process of releas- 2.3 Damage assessment systems ing emergency funds has implications on the quality of the data available. Firstly, the fact that the damage and and data quality issues loss assessments must be carried out within five days, and during the emergency phase, makes it very hard for local authorities to produce accurate estimates, since In Brazil, conducting damage and losses assessments is carrying out such assessments while most of financial part of the broader list of procedures that affected cities and human resources are devoted to the immediate have to follow in order to receive funding from the GoB 17 Coping with Losses: Options for Disaster Risk Financing in Brazil response and when access to many affected locations is not an option. Insurance penetration among low-in- might be restricted, imposes major constraints on the come populations is limited and as a result the quality of quality of the analysis. information available from the insurers concerning the economic impacts of disasters is still limited. As a result, consolidating the historical data on material and human losses has been a difficult task, since inac- Simplifying the procedures and requirements to be met curacies and inconsistencies have been frequently found by the affected communities is a very positive step. in the data. To date, only the information on the human However, detailed information about the damage and toll has been consolidated and detailed official data on losses in both public and private sectors are still needed material losses are not available. for appropriate reconstruction planning and for devel- oping an adequate DRFI strategy. Further adjustments To address the abovementioned difficulties faced by the on the damage assessment framework are therefore affected cities, a simplified version of the damage and still needed to reconcile the different features of the dif- loss assessment form was adopted in 2012. This means ferent post-disaster phases11, and the data requirements that from 2012 onwards the primary sources will not be for the development of improved DRFI instruments. as complete as in the original AVADAN format. More- over, each municipality uses its own valuation criteria for As a next step, consolidating the existing data on mate- affected assets, and this remains an issue under the new rial damage and making it available to the general public, format. Another concern is the fact that these forms as was done with the human losses data, would allow are mandatory only for the cities requesting financial for improved fiscal and economic risk assessments. support from the GoB. This means that the national da- The Ministry of National Integration has already made tasets on disasters miss data from the richer states that great progress collecting the existing data, and the final usually do not request federal financial support. consolidation of historical material losses data would be useful to many public and private stakeholders. This ex- At the same time, using data from the insurance sector, ercise would also inform about the further adjustments typically an important resource for disaster planning, needed in the current damage and loss assessments protocols, thus enhancing data quality in the future. 2.4 Preliminary Fiscal Risk Assessment This section presents a preliminary assessment of the vulnerability to disaster risk, even as the result of fre- fiscal disaster risk profile of Brazil which is built on actu- quent and recurrent events (GAR, 2013). arial analyses of available historical disaster impact data and on findings from the four abovementioned case In order to assess the annual probability of occurrence studies. The risk profile includes, at the aggregate level (return period measure) of potential losses from events of public and private sectors, metrics about Loss Ex- of a specific frequency and severity, an estimated LEC ceedance Curves and a corresponding Average Annual for Brazil12 is presented in Figure 2.5 below. Full de- Loss (AAL). tails of the analysis are presented in Annex 3. Based on the following results it is possible to assess what is These risk metrics have been widely used as inputs for the probability that, in a given year, disaster losses in the design of disaster risk financing strategies. For in- Brazil amount to, for instance, US$5 billion. According stance, based on a PML analysis, Colombia has found to Table 2.3, the probability that disaster losses add up that natural hazards and corresponding disasters are to US$4.8 billion or more in a given year is 20%. This the second most relevant source of contingent liabili- means that Brazil is expected to suffer disaster losses ties (after legal contingent liabilities) and has decided to amounting to at least US$4.8 billion every 5 years. Un- shape the country´s disaster risk financing strategy ac- der the hypothesis that public costs account for 40% cordingly. Similarly, based on a Loss Exceedance Curves of the total losses13, even a relatively likely event, with (LEC) analysis, Honduras was able to assess its financial a 5 year period of return, has a public PML of at least US$1.9 billion (R$4.5 billion). 11 Emergency response, recovery and reconstruction. 12 Cummins and Mahul (2009) define the LEC as the amount that may be equaled or exceeded with a specific probability. The Loss Exceedance curves show that the historical maximum annual loss observed in the last 50 years is US$25.8 billion. With the AAL estimation, we can see that the value obtained using the fitted EM_DAT (EM-DAT: The OFDA/CRED International Disaster Database – www.emdat.be – Université catholique de Louvain – Brussels – Belgium), for a period of return equal to 50 years is 7% higher, but for the mixed data (the uses both EM-DAT and extrapolation from States) it is lower by 11%. Again, a conservative estimation obtained with the EM-DAT database is preferred. 13 According to the case studies presented above, public costs correspond to about 50% of total costs. However, given the omission of many private sector due to the lack of available data, a 30-40% share of public costs is assumed in the fiscal risk assessment. 18 Coping with Losses: Options for Disaster Risk Financing in Brazil Figure 2.5. Loss Exceedance Curves 60 50 40 US$ Billion 30 20 10 0 0 50 100 150 200 (Years) Period of Return - Years Actuarial Method 1 (EM_Act) Actuarial Method 2 (EM_Pearson) Actuarial Method 3 (Low_Pearson) Similarly, the LEC estimates suggest a Probable Maximum Loss associated to a 10 year period of return (i.e. annual probability of 10%) event in Brazil is in excess of US$9.4 billion (R$21.6 billion). Assuming government liability cor- responds to 40% of the total costs, then in a given year there is a 10% probability that disaster related public costs amount to at least US$3.7 billion (R$8.5 billion). The AAL (Average Annual Loss) was estimated at US$3.9 billion (R$8.7 billion). If the government liability corre- sponds to 40% of the total costs, that means that in a given year the Average Expected Public Loss is US$ 1.7 billion (R$3.6 billion). These values are comparable with the most recent disaster related budgetary resources (final alloca- tions, at the federal level) shown in Chapter 3. However, these estimations are based on incomplete data, suggesting that in reality the public AAL might be even higher and, therefore, that response funding gaps might be an issue to be addressed in Brazil. Table 2.3. Probable Maximum Loss estimates (US$ Million) Period of Return Probability of event Total PML Public PML in a given year (US$ Million) (US$ Million) 5 years 20% 4,857 1,943 10 years 10% 9,397 3,759 25 years 4% 18,201 7,280 50 years 2% 27,720 11,088 100 years 1% 40,639 16,256 200 years 0.5% 58,087 23,235 19 Coping with Losses: Options for Disaster Risk Financing in Brazil Chapter 3. Fiscal Management of Natural Disasters C urrently, Brazil heavily relies on ex-post sources of funding for disaster response. Issuing extraordinary credits after the occurrence of natural disasters has been the most common budgetary practice, since annual bud- geting is usually insufficient to meet the disaster response needs. Multi-year reserves are not used at the federal level, while some states have already regulated their own disaster funds. Regulating the established (but not op- erational) national calamity fund (FUNCAP) is an ongoing discussion in the National Congress. Evidence indicates recurrent financing gaps for recovery and reconstruction, suggesting that there is room for improving the financial schemes being used. Advancing a national disaster risk financing and insur- ance strategy could strengthen the fiscal resilience of both Federal and sub- national governments, closing funding gaps that amplify the effects of natural disasters on economic activity and welfare. 20 Coping with Losses: Options for Disaster Risk Financing in Brazil This chapter aims to discuss the adequacy and efficiency The discussion presented below has as reference the Chapter 3. Fiscal Management of Natural Disasters of the current fiscal management of the natural disasters World Bank framework for disaster risk financing and framework. First, a description of disaster risk financing insurance, according to which the various sources of arrangements currently in use and under discussion in financing in the aftermath of a disaster can be catego- Brazil, with emphasis on the changes recently proposed rized as ex-post and ex-ante instruments. The ex-post for the FUNCAP (Special Fund for Public Calamity) is instruments are those that do not require planning in presented. A brief overview of the financial instruments advance, such as in-year budget reallocations, donor as- used by local governments is also shown, as well as a sistance and credit. On the other hand, ex-ante financ- preliminary retrospective public funding gap analysis in ing schemes require planning in advance and include Brazil. contingent credit, multi-year reserves and risk transfers mechanisms (World Bank, 2011a). 3.1 Federal Government’s ex-post disaster risk financing practices and arrangements This section presents the main ex-post sources of funding currently used in Brazil and discusses the excessive reliance on urgent budgetary reallocations in the aftermath of a disaster. In-year budget reallocation Regarding the budgetary allocations, the first relevant the case of those of an exceptional nature, the so-called aspect to be considered is the initial allocation defined créditos extraordinários (extraordinary budget credit), by law. This allocation is an indication of how much that are used to adjust the Annual Budget Law (LOA) money the government expects to earmark to disaster in unforeseen and urgent circumstances, such as those response. resulting from public calamity situations. In the case of DRM, these in-year budget reallocations make the However, in the specific case of disasters, it is important majority of the budget due to the difficulty in precisely to consider not only the initial budget allocation, but anticipating an ‘emergency’ event and the resources also the in-year budget reallocations. This is especially needed. Prior to 2012, disaster related initial allocations Figure 3.1 Disaster related initial and final budgetary allocations 7.000 0,7 58% 58% 6.000 0,6 49% 5.000 46% 0,5 40% R$ Million 4.000 36% 0,4 3.000 25% 0,3 2.000 18% 0,2 13% 7% 1.000 4% 7% 0,1 0 0 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 Initial Allocation Final Allocation Initial Allocation/Final Allocation Source: author´s compilation using original data from SigaBrasil. 21 Coping with Losses: Options for Disaster Risk Financing in Brazil were so low that virtually all the major and many minor Given the amounts requested by the MI, the Ministry events had to be financed through in-year budget real- of Planning determines the sources of the funds, which location instruments such as extraordinary credits. are usually composed of financial surplus revenues from previous years rather than new debt. In the absence of The 2012 initial allocation indicates a trend towards surpluses, the revocation of other allocations can be a strategy more focused on planned financing mecha- suggested. At present, the governing body that would nisms, as shown in figure 3.1, however the need for have its allocations revoked must agree to the proposal, additional resources was such that the initial budget but the in-year realignment of investment budgets is accounted for only about 36% of the total budgetary rare and unlikely. resources for disasters. The PM is valid for three months, extendable for an ad- To issue extraordinary budget credits, the Ministry of ditional three if not voted into Law. After six months, if National Integration (MI) submits a request to the Min- it is not passed into Law, the PM is canceled. During its istry of Planning (MPOG) for additional resources based passage into Law, the terms of the PM can be modified. on its damage, losses and needs estimates. The MI re- If the PM is canceled, the blocked resources remain as quests funding for emergency services and assistance, as a financial surplus. well as reconstruction, but avoids including preventive measures in the extraordinary budget request, which It is important to note that a key benefit of this whole is enacted through a Medida Provisória or Provisional process is that it takes place within a few days, ensuring Measure (PM), which is an executive order with imme- that funds for immediate response and relief are avail- diate effect, but which has to be approved by Congress able in a timely manner. However, the fact that relief, within 180 days. The MI also defines the beneficiaries recovery and reconstruction resources are usually re- of such resources and consolidates a proposal that is leased through the same process implies that recon- passed through the Ministry of Finance, Ministry of Plan- struction activities cannot be planned accordingly. Thus, ning and the Chief of Staff of the President of the Re- following a disaster, the Federal Government must re- public. The proposal, based on the estimates provided spond to the immediate aftermath in a matter of days by the MI (based on technical criteria), set forth: the and at the same time prepare a reconstruction financial amount of extraordinary budget credits, the distribution plan under far from ideal circumstances while a broad between the beneficiaries and the amounts that can be analysis of the real impacts from the disaster still has to used in between current and capital expenses. be conducted. Figure 3.2 Disaster related final budgetary allocations and disbursement 7.000 86% 0,9 80% 75% 0,8 6.000 72% 70% 0,7 5.000 59% 60% 53% 57% 0,6 50% R$ Million 4.000 45% 0,5 3.000 32% 0,4 0,3 2.000 0,2 1.000 0,1 0 0 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 Final Allocation Disbursement Disbursement/Final Allocation Source: author´s compilation using original data from SigaBrasil. 22 Coping with Losses: Options for Disaster Risk Financing in Brazil Moreover, as the budgetary execution is carried out by both national and local executive institutions (e.g. while the reconstruction of federal roads is managed by the Ministry of Transport, reconstruction of schools is nor- mally executed by the state education secretariat), dis- bursement, monitoring and accountability issues have an additional dimension, which is to accommodate this multi-level feature. As for disbursement, following the official recognition of a disaster, special arrangements are made to speed up funding flows. These procedures have already been simplified in recent years, but local governments in particular still find it cumbersome to satisfy all the le- gal and technical requirements for project approval. The execution rate has shown significant variation over has been used over the years: from property tax exemp- the years, but with no clear trend from 2001 to 2012 tions to benefits awarded to the industrial, agricultural (Figure 3.2). According to the data below, in 7 of the and commercial sectors, but an analysis on the impacts of 12 years considered the execution rates were between such measures has not been done. 45% and 60% of the final allocations, which suggests that improving budget execution could lead to more International assistance effective responses to disasters. Moreover, given that most of these resources correspond to reallocations, it From a national perspective, international donor fi- is necessary to investigate whether low budget execu- nancial assistance is not particularly relevant in Brazil. tion is also determined by sub-optimal distribution of Recovery and reconstruction costs are usually much the resources among the affected sectors/regions. higher than the international assistance inflows, but at the local level this source of funding is more substan- Monitoring the use of disaster funds is also particularly tive. For example, following the 2008 Santa Catarina challenging when execution happens at the state level floods and landslides donor assistance (not exclusively and, in order to tackle this issue, the Ministry of Nation- international) was sufficient to fund the temporary cash al Integration has implemented the Civil Defense Pay- transfers program to eligible households. International ment card, which can be used only at pre-authorized donor assistance also supported the reconstruction suppliers and has already proven to induce more cost- of low-income housing in that state. However, this is effective spending. However, while successful in track- not the main source of funding: local governments rely ing spending during the emergency stage, the card can- more on assistance from the Federal Government than not be used as a monitoring device for reconstruction. on international donations. In this context, the establishment of a facility especially designed to manage funding with such different charac- Other ex-post disaster teristics could allow the Federal Government to better plan and execute disaster related policies, while opti- funding resources mizing collaboration with local institutions. Depending on the sectors most affected, a diverse range of funding resources can be used to support recovery Taxation and reconstruction. For example, extended loan repay- ments and subsidized loans have been used to support At the federal level specific taxation is not currently used farmers. Small business recovery programs and special to raise revenues for disaster prevention, preparedness loans are also among the measures taken to support and response/recovery/reconstruction. On the contrary, economic activity recovery in some affected municipali- following a disaster the national and local governments ties. Even investments in marketing to support the tour- usually adopt tax exemptions to support recovery, gen- ism sector were made by the government following the erating a reduction of the revenues along with the new 2008 event in Santa Catarina (the floods occurred just expenses requirements. A variety of these exemptions before the peak tourist season, raising concerns about 23 Coping with Losses: Options for Disaster Risk Financing in Brazil tourism sector performance in 2009). 3.2 Federal government’s ex-ante disaster risk financing practices and arrangements This section discusses the use of ex-ante sources of funding in Brazil. Fiscal data indicates that annual budgeting pro- cedures could be improved to better accommodate disaster related public expenditure. The ongoing discussion about the regulation of a national disaster fund suggests that the need for a multi-year platform for sovereign disaster risk financing is recognized. Annual budgeting Multi-year reserves In Brazil, whether or not there is a disaster-specific bud- At the federal level, a multi-year reserve mechanism get appropriation depends on the multi-year plan (PPA) has not been established to date. The FUNCAP (Special established by the Ministry of Planning every four years. Calamity Fund), created in 1969, is the proposed multi- Since 2004, the approved guidelines have included pro- year facility, but since its establishment it has not been grams specifically designed to track disaster prevention operational given the lack of appropriate regulation. and response policies, while prior to that only general Major changes are currently under debate, but have not Civil Defense appropriations were included in the multi- yet been approved. See box 2.1 for details. year plans. Nevertheless, even though the recent years budgetary guidelines are more precise in allocating re- Other than the non-operational FUNCAP , if the ex- sources for disaster prevention and response, relevant traordinary budget credits are issued in the last four disaster related allocations are also made through non months of the year they can be carried over the next fis- disaster-specific lines. Any analysis of available and used cal year under special budgetary procedures. The PPA public funds must therefore take into account additional could serve as multi-year budget tool, ensuring resourc- appropriations. es in the medium term for the needs of the line minis- tries. However the budget practices have undermined In order to address this issue, details of each multi-year the credibility of the PPA as a medium term resource plan were taken into consideration and a more compre- envelope, which explains why several line agencies have hensive methodology for public resources and expendi- resorted to short term stop-gap methods such as the ture tracking (at the federal level) was developed (for so-called fundos financeiros. details see appendix 3). The line ministries involved are the Ministries of Transport, National Integration, Justice, Contingent loans Cities, Defense, and Science, Technology and Innova- tion. Contingent loans are not currently used in Brazil. The Federal Government retains risk based on the premise These adjusted measures of disaster related budgetary that the costs of natural disasters can be accommodated allocations corroborate the change in the mindset after in the budget either through annual or in-year budget- the 2011 Rio de Janeiro State flash floods and landslides: ary reallocations, always in the expectation of regular not only the focus has changed from response to pre- tax revenues or other financial inflows, without having vention and preparedness, but also the traditionally low to rely on contingency credit operations. initial budget allocations have significantly increased in 2012, providing evidence that the Federal Government has recognized the importance of a disaster risk financ- Risk transfer ing strategy in a more comprehensive disaster risk re- duction (DRR) agenda. Traditional insurance of public assets and parametric sovereign insurance has not been developed for Bra- In accordance with Brazil´s legal framework, budgetary zil, and the Federal Government’s strategy to date has provisions last for one fiscal year, except for committed been to retain risk by self-insuring its assets. ‘Catastro- resources that are carried over to the next budget year, phe bonds’14 have not been issued in Brazil either. the so- called Restos a Pagar or carry-overs. 14 “High-yielding, insurance-linked security providing for payment of interest and/or principal to be suspended or cancelled in the event of a specified catastrophe, such as an earthquake” (Cummins, D.; Mahul, O., 2009) 24 Coping with Losses: Options for Disaster Risk Financing in Brazil On the other hand, housing projects financed by the promoting development in Brazil through financial in- Minha Casa, Minha Vida program are covered against novations. damage caused by natural disaster as part of the broader insurance component of this program, including insur- Any insurance or risk transfer mechanism for the gov- ance against household income fluctuations or default ernment should seek to go beyond insuring the physi- caused by death. Damage-related claims are very low cal assets of the GoB and also include new instruments compared to the other events covered by insurance. such as parametric solutions to hedge a portion of the economic / fiscal exposure to disasters. Details of this In addition, since 2008 a set of policies and regulations and related topics (such agricultural insurance policies) for property microinsurance is back on the Superinten- are part of the insurance market review in chapter 4. dence of Private Insurance (SUSEP)’s agenda, aimed at Box 3.1 The FUNCAP as it is Brief history: The FUNCAP (Fundo Especial para Calamidades Públicas) was created by Decree 0590 (in 1969), with resources originally from the Federal Government budget, donations, available funds allocated to civil defense actions and other ad hoc sources of funding. Sources of funding: FUNCAP is currently a fund divided into quotas which can be bought by states and mu- nicipalities via a voluntary transfer of resources. It is structured in the form of a matching grant fund, in which for each Brazilian real (BRL) transferred by a state or municipality to the fund, the Federal Government must transfer three BRL. In the event of a disaster, subnational governments can withdraw resources from it only up to the extent of their quotas. Furthermore, the Federal Government is entitled to make voluntary transfers to FUNCAP in order to encour- age participation by local governments. However, FUNCAP has not received any financial resources since 1995. Goals: FUNCAP must finance the reconstruction of areas that have been affected by natural disasters. In exceptional cases, the resources can be used to finance immediate rescue operations, aid to victims and the reestablishment of essential services. Use of funds: if no natural disaster occurs, the invested quota may be withdrawn after two years have elapsed from the time of the investment. In case of no disaster, the local government is not entitled to the correspond- ing contributions made by the Federal Government. The chart below explains this mechanism: YEAR 0 YEAR 1 YEAR 2 Municipailty Federal government Natural disaster Municipality transfers R$ 1,00 to transfers R$ 3,00 to happens withdraws R$ 4,00 FUNCAP FUNCAP Natural disaster Natural disaster Municipality happens does not happen withdraws R$ 1,00 Municipality Municipality withdraws R$ 4,00 maintains its quotas 25 Coping with Losses: Options for Disaster Risk Financing in Brazil Box 3.2 Proposals for a new FUNCAP (law 3.084/2012, with amendments) Expansion of funding sources: Regarding funding sources, the main proposals are to include a 2.50% share of federal lottery receipts* (from the Caixa Econômica Federal) and a Federal Government investment of R$ 5,00 in FUNCAP for each R$ 1,00 obtained in this way from lottery cash. While these funds would not purchase additional FUNCAP quotas, they would add value to the existing municipal and state quotas. As an illustration, given the federal lottery revenues in the past years, FUNCAP would have had lottery rev- enues according to the table below: Table B.1 FUNCAP expected lottery revenue Year R$ Million 2008 143.50 2009 184.00 2010 220.25 2011 243.25 2012 262.25 2013* 291.93 Source: author’s elaboration. (*)Forecast Extension of FUNCAP’s goals: The funds may be withdrawn to finance rescue operations, reestablish essential services and provide aid to victims. The original goal (reconstruction of natural disaster areas) is expanded. Article 8 of Law 12.340 (2010) would include civil defense preparedness and financing for disaster prevention as part of FUNCAP’s objectives, subject to certain restrictions: the resources invested by the GoB must be allocated to prevention and preparedness and at least 50% of the resources withdrawn must be allocated to (i) identifying risk areas, (ii) re-greening risk areas, (iii) implementing urban drainage systems, (iv) hydrometeoro- logical monitoring, (v) implementing the Civil Defense Contingency Plan, and (vi) training civil defense agents. * Currently a share of the federal lotteries revenues is distributed to a series of culture, sports, education, health and security programs. 26 Coping with Losses: Options for Disaster Risk Financing in Brazil 3.3 Local governments’ disaster risk financing practices and arrangements Annual budgeting At the state level the same budgetary process apply, meaning that specific disaster related budget appropriations are defined within the context of each state’s multi-annual plan. In terms of gathering fiscal data at the state level, given the lack of access to detailed information, the best approach is to focus on the budgeted and executed expenses recorded in the more general Civil Defense government sub-function. An interesting finding is that at the local level disaster related (Civil Defense) initial budget allocations have been os- cillating around higher (compared to the GoB’s numbers) levels since 2006 (for details, see appendix 2). However, relative to likely needs, budgetary provisions are very limited at the state level as well, which is not surprising given the response capacity of local governments. Looking at the expenditures in each of the states, there are some key points to be emphasized. Considering the period between 2006 and 2012, three of the states respond for more than 50% of the entire amount allocated (initial and updated), committed and paid off in the Civil Defense sub-function, to all the states together. Rio de Janeiro allocates approximately 29% of the paid-off expenditure in the period and Pernambuco and Paraná allo- cate 20% and 10% respectively. Moreover, it is important to stress that the initial allocation of Pernambuco has been considerably lower than the final allocation over the past few years. This indicates that this state allocates resources for dealing with disasters essentially by in-year budget reallocation. 27 Coping with Losses: Options for Disaster Risk Financing in Brazil Table 3.1. Total of the States’ expenditure with the sub-function ‘Civil Defense’ for the period 2006-2012* (including intra-budgetary expenditure, R$ million) Year Initial Allocation Final Allocation Committed Verification* RAPs (carry-over) 2006 542,13 574,80 421,39 415,74 - 2007 658,25 791,80 580,28 507,10 71,27 2008 745,37 1.092,44 865,58 763,71 100,25 2009 1.026,18 1.462,59 1.256,59 1.141,55 75,57 2010 874,28 2.028,22 1.556,18 1.298,10 238,06 2011 1.221,45 2.168,93 1.405,46 1.281,45 116,34 2012 1.306,05 1.970,79 1.359,15 1.159,88 186,42 TOTAL 6.373,69 10.089,57 7.444,62 6.567,52 787,91 Source: Budget Execution Summary Report – SISTN/CEF. The Civil Defence sub-function does not include security spending, which are listed under a separate sub-function. * Intermediate step between commitment and financial disbursement that refers to the verifica- tion that goods or services were provided accordingly. Only after liquidation the financial payments can be made. A second key piece of evidence to note is the large amount registered as carry-overs of Civil Defense expenditure commitments, henceforth RAP (Restos a Pagar)15, in some of the states. Alagoas is alone responsible for almost 34% of the total RAP in the Civil Defence sub-function in the states, which indicates this state faces difficulties in executing and paying expenditures related to disasters. After Alagoas come the states of São Paulo (16%), Minas Gerais (15%), Rio de Janeiro (11%) and Paraná (9%). Altogether these states account for 84% of the total RAP registered between 2006 and 2012. It is also important to notice that there is a lack of budget data in the State of Rio Grande do Sul in 2008, where it did not record any amount for Civil Defence in any of the sources examined. 15 In Brazil, following the initial allocation of resources (and reallocation when required), the budget execution process includes the commitment, the verification and the financial disbursement stages. Only after verifying the adequate provision of goods or services the Government proceeds with the financial disbursement. 28 Coping with Losses: Options for Disaster Risk Financing in Brazil funds for reconstruction, but this policy was rejected Multi-year reserves by the population and could not be implemented suc- cessfully. As happens with the Federal Government, tax At the local level, disaster (calamity) funds have been exemptions to specific sectors, on the other hand, are used to finance the costs of natural disasters, and with often used as an instrument to support recovery. these facilities states and municipalities are able to carry over resources for the following fiscal years and, therefore, manage multi-year reserves. However, even Contingent loans where these local funds are operational, resources al- located are usually very limited compared to recovery At the state level, contingent credit operations are not and reconstruction needs. Another interesting feature being used. DRM components of broader World Bank of local disaster funds is that some of them exist with loans are usually focused on preventive measures, revenues linked to specific taxes. Such a difference in equipment acquisition and technical assistance, not in- the approach taken by the Federal and local Govern- cluding contingent credit products. Following the 2011 ments suggests that where the impacts of disasters are floods and landslides in Rio de Janeiro State, a US$ 485 actually felt, the need for improved disaster risk financ- million Development Policy Loan (DPL) was approved ing instruments has been identified and is already lead- and disbursed to finance reconstruction efforts in the ing to institutional change. See annex 3 for details on mountainous hinterland. However, this project was al- the established local disaster funds. ready under preparation as the result of the floods and landslides that hit the state before 2011 (the region is af- Taxation fected annually by this type of event), so that by January 2011 a series of necessary procedures was already in Following a natural adverse event, tax increases are un- place. In the absence of this project preparation, the Rio common at the local level as well. Following the 2008 de Janeiro state government could have faced additional floods and landslides in Santa Catarina, the state govern- difficulties in accessing liquidity. ment attempted to temporarily increase taxes to raise 3.4 Funding gap analysis Recovery and reconstruction gaps exacerbate the nega- tive impacts of disasters on socioeconomic outcomes and for this reason we estimate post-disaster needs funding gaps in Brazil from 2006 to 2010. In this period, the financing gaps in a given year was, on average, of R$ 1.4 billion, or 30% of the estimated average disaster related government liability. These numbers are pre- liminary and based on a series of assumptions (see Box 2.3), but they corroborate anecdotal evidence from the field, suggesting that improving both disaster risk financ- ing arrangements and damage and loss assessment sys- tems are needed for resilient and timely reconstruction. Given that delays in recovery and reconstruction exac- erbate the negative impacts of disasters on socioeco- nomic outcomes, one issue that arises is whether the financing instruments currently employed provide the affected cities with the necessary response resources in a timely and cost effectively manner. Anecdotal evidence from the field suggests that recov- ery and reconstruction funding gaps are major and have significant implications in terms of welfare in Brazil. In 29 Coping with Losses: Options for Disaster Risk Financing in Brazil order to evaluate whether the budgetary resources are caused by disasters in Brazil, the following analysis must sufficient to fund recovery and reconstruction costs we be considered in light of important caveats and simplify- carry out a preliminary retrospective funding gap analy- ing assumptions. First, a time series of aggregate gov- sis in this section. ernment liability is built based on partial information and on a set of assumptions that can be adjusted as addi- More specifically, for the purposes of this study we de- tional data becomes available. fine a funding gap as the difference between the bud- getary resources available to finance disaster response Second, the disaster related fiscal dataset was built in (from both federal and state budgets) and the estimated accordance with a detailed budgetary process review. government liability due to natural disasters in a given The employed methodology was meant to be as com- year. We do not make explicit considerations about prehensive as possible, but given the lack of better emergency relief funding gaps based on the premise tracking systems some approximations had to be made, that this kind of financing need is met by the local and especially at the state level. The details on the estima- Federal Governments. tion hypotheses and procedures are discussed in An- nex 2 and the basic assumptions summarized in box 3.3 Given the lack of an appropriate budgetary tracking sys- below. tem and even of historical data on damage and losses Box 3.3 Funding Gap Analysis: Underlying Assumptions The government Liability Estimates are based on actual and simulated data. Available data are the annual number of affected people in Brazil, and historical data on the number of damaged and destroyed dwellings in Paraná State, based on which the historical number of damaged and destroyed dwellings in Brazil is simulated. Other parameters were assumed as well: based on the 2010 dwellings/public schools ratio and on the dwell- ings/public health are defined to approximate the number of affected schools and health centers. The following three scenarios are defined in order to estimate funding gaps under low, medium and high costs assumptions. For each affected dwelling and person a multiplier is associated to the cost assumptions detailed below in order to generate three series of annual government liability (at current values) in the social sectors (housing, education and health). Table B.2. Unit Cost Assumptions Type of Asset Unit Costs to the Public Sector - (R$ 1.00, at 2012 values) Low Costs Medium Costs High Costs Destroyed Dwellings 40,114.29 40,114.29 40,114.29 Damaged Dwellings 2,005.71 4,011.43 8,022.86 Affected Schools 25,000.00 50,000.00 100,000.00 Affected Health Center 12,500.00 25,000.00 50,000.00 Affected Person 250.00 500.00 750.00 These assumptions capture part of the public damage and losses in the social sectors (housing, health and education). The high fiscal burden associated to the recovery and reconstruction of the infrastructure sectors is not taken into account because average costs are very poor approximations for the true parameters given the high het- erogeneity of the variables underlying the distribution of public infrastructure recovery costs. 30 Coping with Losses: Options for Disaster Risk Financing in Brazil Since in-year budgetary reallocations (through budgetary reallocations) are the main source of funding for post-disas- ter recovery and reconstruction needs, we analyze funding gaps with and without extraordinary credits. As shown in figure 3.3, compared to the disaster related initial budgetary allocations the financing gaps are significant even under conservative assumptions. Resources within the annual budget account for 60% or less of total estimated government liability under the low cost assumptions, 33% or less under the medium costs assumptions and 21% or less under the high costs assumptions. Figure 3.3. Funding Gap Analysis: Before in-year Reallocations 1.500 -500 R$ Current Million -2.500 -4.500 -6.500 -8.500 2006 2007 2008 2009 2010 Federal and State Budget Resources before Reallocations Recovery and Reconstruction Funding Gap (Low) before Rellocations Recovery and Reconstruction Funding Gap (Medium) before Rellocations Recovery and Reconstruction Funding Gap (High) before Rellocations Source: author´s elaboration with data from SigaBrasil and Civil Defense. When taking into account budgetary reallocations that are commonly issued following more severe disasters, the funding gap for recovery and reconstruction is smaller but still considerable: figure 3.4 shows that under the low cost assumptions funding gaps have not been observed since 2007, and medium or high costs assumptions produced fund- ing gaps estimations, for example, as high R$ 5.3 billion in 2009. Figure 3.4 Funding Gap Analysis: After in-year Reallocations 6.000 4.000 R$ Current Million 2.000 0 -2.000 -4.000 -6.000 2006 2007 2008 2009 2010 Federal and State Budget Resources afterReallocations Recovery and Reconstruction Funding Gap (Low) afterRellocations Recovery and Reconstruction Funding Gap (Medium) afterRellocations Recovery and Reconstruction Funding Gap (High) after Rellocations Source: author´s elaboration with data from SigaBrasil and Civil Defense. Even when considering the budgetary reallocations following major events, moderate assumptions suffice to indicate that most of the time the resources made available through these budget lines are not enough to finance full recovery and reconstruction in the affected communities. See table 3.2 for a detailed exercise. 31 Coping with Losses: Options for Disaster Risk Financing in Brazil Table 3.2. Budgetary Provisions as a % of Estimated Government Liability Allocated Resources/Estimated Government Liability (%) Before In-year Reallocations After In-year Reallocations Year Low Costs Medium Costs High Costs Low Costs Medium Costs High Costs 2006 61.28% 32.44% 21.08% 87.95% 46.56% 30.25% 2007 55.83% 29.40% 18.98% 110.86% 58.37% 37.69% 2008 47.51% 25.48% 16.22% 157.42% 84.42% 53.76% 2009 42.95% 23.42% 14.92% 124.01% 67.64% 43.09% 2010 38.46% 20.64% 13.35% 202.68% 108.79% 70.35% Source: Author’s compilation with SigaBrasil and Civil Defense data. It is of major importance to point out what is not be- cal indicators was built with the goal of avoiding under- ing taken into account in the above exercise. First, we estimating the public resources available for disaster re- are not explicit about including the costs of emergency sponse. However, given that they are an approximation, and relief operations, which can for example be high the numbers obtained might differ from the resources during floods and mass movements. Additionally, these actually raised. estimates do not include a proxy for the impacts on the transport sector, which is a major component of public With these considerations in mind, the key messages damages according to the case studies presented earlier, from the analysis above are: or for the impacts on any other public infrastructure. Moreover, the costs of urgently required vulnerability • Even based on conservative assumptions and reduction and adaptation measures following a disaster, taking into account extraordinary credits made as well as the impacts on the economic sectors are not available during the fiscal year, there is evidence of considered. Therefore, the analysis above cannot be seen actual government liability being regularly higher as a definitive figure, but instead as a preliminary exercise than the public resources available to the affected in which we sketch lower bounds based on different as- cities and states during the past years. sumptions for the government liability and funding gaps from 2006 to 2010 regarding the social sectors only. • The determinants of these funding gaps have to be further studied since it is not clear to what ex- Besides that, ideally we would carry out a dynamic tent the lack of available resources reflects institu- funding gap analysis, meaning that we would estimate tional, technical and/or financial constraints. How- recovery financing gaps separately from reconstruction ever, the unstructured budgetary process suggests financing gaps. The inclusion of this time dimension is that the establishment of a multi-year platform important because both liquidity and financing needs such as the FUNCAP could be used to formalize to differ substantially depending on the post-disaster funding flow arrangements to different line minis- phase. However, the nature of the budgetary process tries and state level calamity funds. in Brazil is such that the most detailed accessible budget lines are still too general to allow for these dynamic con- • Moreover, the inclusion of missing data would siderations. Moreover, given the recent changes noticed more than likely reveal a worse situation that is con- in 2012 it would be important to expand this analysis sistent with anecdotal evidence from the field. This beyond more recent periods. However, the absence of preliminary assessment was based only on several available data on the occurrence of disasters since 2011 assumptions and simulated data, and the availability rules out such expansion at the moment. of improved primary sources of information would significantly enrich an analysis such as the one pre- Finally, given the fact that reconstruction programs sented here. The availability of improved data man- could be financed through the realignment of invest- agement systems, as suggested above, could further ments plans, it is possible that some resources have not enrich these kinds of analysis and improve the accu- been tracked. The methodology developed for the fis- racy and timeliness of funding gap monitoring. 32 Coping with Losses: Options for Disaster Risk Financing in Brazil 33 Coping with Losses: Options for Disaster Risk Financing in Brazil Chapter 4. Insurance Market Overview B razil has the largest non-life insurance market in Latin America, but penetration rates are still low (1.08% of GDP) compared to developed economies (3.6% of GDP). Agricultural insurance penetration rates have increased since 2003, when a government subsidies program (PSR) was launched. Disaster property insurance is usually covered by comprehensive policies and a framework for disaster microinsurance still awaits approval. Several of the GoB’s temporary cash transfer programs are used in the af- termath of disaster, and like the PSR, the development of a broader DRFI strategy could make these programs more cost-effective. 34 Coping with Losses: Options for Disaster Risk Financing in Brazil This chapter overviews the most relevant catastrophe insurance market segments in Brazil, namely agricultural, prop- Chapter 4. Insurance Market Overview erty and microinsurance. In general, catastrophe related insurance markets are relatively undeveloped. Agricultural insurance is the most developed segment, sustained by a GoB subsidies program and enforcement policies linking insurance to credit. In addition, cash transfers programs play the role of microinsurance both in the property and agricultural sectors. These programs have become more important over recent years, but the establishment of a comprehensive DRFI framework could make them more cost-effective. 4.1 Non-life insurance market overview Figure 4.1 Insurance Penetration Rates (% of GDP) 5,1 3,6 2,6 2,7 2 1,7 1,3 1,5 1,4 1,1 1,1 1,1 1 1 0,6 0,1 Industrial Emerging Emerging Latin Eastern Africa Middle Brazil Markets Markets Asia America Europe East Non-life Life Source: Zurich Seguros and Swiss RE, 2011. With a share of over 20% of the Latin American non-life The catastrophe insurance share of the non-life sub-sec- insurance premiums (around US$23 billion in 2012), tor is also low. Since 2008, non-life insurance penetra- Brazil is the largest market in the region, followed by tion has been around 1.00% of GDP , while the catastro- Mexico and Puerto Rico, with US$10 billion each. phe insurance penetration (considering rural, property and microinsurance) has been less than 0.2% of GDP16. Compared overall with Latin America, Brazil´s average penetration rates are fairly strong, but still much lower Given these low penetration rates and the recent eco- than those in developed economies (Figure 4.1). nomic and regulatory developments, private insurers still see great potential for a continued expansion of the non-life insurance market in the country. 16 It is important to note, that in the property sector in particular coverage against catastrophic events is usually included in more general, compre- hensive insurance policies, unless explicitly stated otherwise. For the purpose of this study, comprehensive insurance policies against general damage are considered as catastrophe insurance as well. The auto sub-sector is not considered to be catastrophe insurance in the above calculations, but auto policies might cover damage caused by natural disasters as well (depending on the details of each contract). Therefore, it is not possible to present a complete assessment of the market size and penetration rates of catastrophe insurance in Brazil. 35 Coping with Losses: Options for Disaster Risk Financing in Brazil Figure 4.2 Non-life and catastrope insurance: premiums (R$ million) and penetration rates (% of GDP) 60.000 1,40% 50.000 1,20% 40.000 1,00% 30.000 0,80% 0,60% 20.000 0,40% 10.000 0,20% 0 0,00% 2008 2009 2010 2011 2012 Non-life insurance premium Cat-related insurance premium Cat-related insurance penetration Non-life insurance penetration Source: Susep, Axco and IBGE. Regarding the market structure, in recent years a number of insurers have entered the market, at the same time as a trend towards consolidation among major insurers. (Table 4.1). In 2011, 116 insurance companies were operating in Brazil, but about 60% of the total non-life premiums were concentrated on the 7 largest insurance groups (Axco, 2013). Table 4.1. Leading non-life insurance groups: market share Insurance Group Non-life insurance market share BB-MAPFRE 14.30% Itaú Unibanco 14.10% Porto Seguro 11.40% Bradesco 10.50% Santander 3.80% Caixa 3.50% SulAmerica 2.20% Source: Axco. Another important recent trend has been the opening of the Brazilian reinsurance17 market. The monopoly of the IRB (Brazilian Reinsurance Institute) was terminated in 2007, and international reinsurers were allowed into the Brazil- ian market, although under special conditions, i.e. a reserve for local reinsurance companies. This characterizes the Brazilian reinsurance sector as a not entirely free market. As a result, over the past years the number of domestic reinsurers (which might be owned by international groups) has increased to 10, driving the market share of the IRB from 100% before the opening of the market to about 30% in 2012 (A.M. Best, 2013). This might have significant implications, for instance, on the agricultural sub-sector given that the country is highly exposed to systemic risks and reinsurance cessions in this sector can be as high as 80% (World Bank, 2010). 17 Reinsurance is an operation by which insurers transfer risk to reinsurer(s). If an insurer is obliged to pay compensation to its clients, the reinsurer would pay for the insurer losses (according to the terms of the reinsurance policy), therefore protecting the insurer from major financial losses. 36 Coping with Losses: Options for Disaster Risk Financing in Brazil 4.2 Agricultural Insurance This section aims at initiating a discussion on how agricultural insurance schemes currently available in Brazil could be further developed to improve the Federal Government´s fiscal management of both ex-ante and ex-post programs designed to support farmers exposed to natural hazards and to enhance private insurers´ ability to compete in the insurance markets. full agribusiness supply chain, the agribusiness share of Importance of the agricultural sector GDP was 21% in 2010, while the agriculture value-add- ed as a percentage of the national GDP was approxi- Among the most important agricultural producers in mately 4.5% during the same period. the world, Brazil has a strong tradition in the agricultural sector and a long history of droughts, floods and hail- Public sector support to agricultural insurance in Brazil storms. In recent years many policies and institutional is substantial when compared to other Latin American changes have been introduced to improve risk manage- countries, but policies involving public subsidies and ment in the sector, especially in states highly vulner- institutional frameworks can still be better tailored to able to droughts, such as those in the northeastern and achieve the government´s goal of increasing farmers´ southern regions. resilience to natural disasters while preserving the fiscal balance of the governmental support programs. This is not surprising given the importance of the agri- culture sector to the economy. When considering the Figure 4.3 Agriculture GDP as % os Total GDP 25% 24% 23% 22% 22% 22% 21% 21% 6,39% 5,93% 4,90% 4,71% 4,78% 5,03% 4,85% 4,54% 2003 2004 2005 2006 2007 2008 2009 2010 Agribusiness GDP as % os Total GDP Agriculture Value Added as % of GDP Source: IBGE and Esalq. 37 Coping with Losses: Options for Disaster Risk Financing in Brazil Box 4.1. Agriculture insurance: products available In Brazil, crop, livestock, aquaculture and forestry insurance are available both with and without guarantee from an agriculture insurance stabilizing fund (FESR, described below). According to SUSEP (Superintendence of Private Insurance), in addition to protecting production, farmers can also insure assets and inputs against damage. This type of insurance can be specifically tailored depending on whether or not such assets are col- lateral to rural credit operations. A specific life insurance is also available for rural producers with financial liabilities and some receivables can also be insured against default. Multi-peril crop insurance (MPCI) and indemnity-based (named peril) crop insurance are the most common types of agricultural insurance in Brazil. The first type establishes an insured yield (given by a chosen measure) and if actual yields are below the insured level, an indemnity is paid to cover the losses. The second type offers cover against a specific hazard, and claims are based on the extension of the damage in the field. Area-yield index-based crop insurance, which offers coverage according to the average yield of a larger area that includes more than one insured party, was used in Rio Grande do Sul State by maize farmers, but the pro- gram was discontinued after 2009 (World Bank, 2010). Livestock, aquaculture and forestry insurance is also available and eligible for subsidies as well, but penetration rates are estimated at much lower levels than those for crop insurance products (World Bank, 2010). Crop insurance and credit related operations constitute the bulk of the agricultural insurance market in Brazil. However, the relative share of credit linked insurance operations has decreased since 2003. This is partially related to the public subsidies program launched in 2003 which significantly reduced the cost of insurance premiums to farmers (Figure B.4.1). Figure B 4.1. Rural Insurance Products 100% 1400 90% 1200 80% 70% 1000 60% 800 50% 600 40% 30% 400 20% 200 10% 0% 0 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 Other Rural Insurance Credit related Insurance (% of Total) Crop Insurance (% of Total) Total Rural Insurance (R$ Million) Source: Susep. *Partial data.. Mutual insurance schemes, arrangements in which the insurers are also the policyholders, have also been suc- cessfully developed in Brazil. In Rio Grande do Sul state, for example, AFUBRA (Associação dos Fumicultores do Brasil) protects farmers against hail and windstorm hazards. Other groups such as the Batavo, Coamo and Irga Cooperative (rice producers association) are also among the main mutual insurance groups (Buainain and Vieira, 2011). These small-scale arrangements can be used to protect farmers against more severe, but less frequent and non-systemic hazards such as hail in southern Brazil (World Bank, 2010). 38 Coping with Losses: Options for Disaster Risk Financing in Brazil Penetration of agricultural insurance Figure 4.4 Rural Insurance Penetration The fivefold increase in total agricultural insurance pre- 0,70% miums was accompanied by a corresponding upswing in 0,60% penetrations rates, which increased from under 0.2% in 2003 to 0.6% in 2012 (taking into account total rural 0,50% insurance premiums over agricultural valued added). 0,40% This increase in penetration rates is significant: in 2007, 0,30% when agricultural insurance penetration rates in Bra- 0,20% zil were about 0.3%, the country was comparable to 0,10% Panama, the Windward Islands and Paraguay. By then, more advanced markets such as Mexico and Chile had 0,00% agricultural insurance premiums corresponding to 0.6% 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 of their agricultural GDP, while the remaining countries Rural insurance penetration in the region had penetration rates of less than 0.1% of Rural insurance penetration (Agrobusiness) agricultural GDP (World Bank, 2010). However, it is important to note that despite recent improved performance, penetration rates are still far from those observed in high income countries (Figure 4,5). Despite being a relatively developed market, agricultural insurance penetration varies from region to region in Brazil, with the southeastern and central-southern parts of the country having much higher levels of agricultural insurance penetration than the northeastern states characterized by dry-land mixed-farming systems (World Bank, 2010). Figure 4.5. Comparative Agricultural Insurance Penetration 2,3% 0,9% 0,3% 0,3% 0,2% 0,0% High All Countries Upper Middle- Brazil Lower Middle- Low Income Income Income Income Countries Countries Countries Countries Source: Stutlev and Mahul, 2010. 39 Coping with Losses: Options for Disaster Risk Financing in Brazil Box 4.2. Market Concentration As for the number of insurers operating in the market, while Uruguay, Paraguay, and Chile had in the recent years four insurers operating in the agricultural sector, and concentration was absolute in Costa Rica, the Do- minican Republic, Guatemala, and the West Indies - each with only one insurance company offering agricultural insurance (World Bank, 2010) - Brazil seems to have a market structure more similar to the Argentinian one (which had over 27 agricultural insurers in 2007). Brazil had about 25 groups offering agricultural insurance in 2012 (SUSEP). However, most groups with dif- ferent names run joint operations and the largest groups concentrate a very high share of the market (P&C, 2013). In other words, despite an apparent competitive structure, the market is actually concentrated on the few players with better distribution channels and agricultural credit operations. In Brazil, even though insurance brokers are highly specialized in agricultural insurance delivery, linking insur- ance to rural credit and therefore using bank branches as distribution channels has been an important strategy to increase coverage and make the subsidies program effective (World Bank, 2010 and Adami and Ozaki, 2012). This suggests that delivery costs might be an important driver of market concentration. Government support programs Along with Mexico, Brazil provides a high level of public Launched in 2003, the PSR is a leading government pro- sector support to risk management in the agricultural gram to support agricultural risk management in Brazil. sectors. Indeed, three of the government programs Since 2003, budgetary resources allocated to the pro- available in Brazil (PROAGRO, SEAF and Fundo Ga- gram and the amount of subsidies has increased signifi- rantia Safra) are the only ones which follow a full gov- cantly, especially after insurance was linked to major ernment-intervened model in Latin America (World rural credit programs (table 4.2). Bank, 2010). Furthermore, in 2003 Brazil established a subsidies program for private agricultural insurance, since then penetration rates increased significantly. The support provided to the private insurance market also includes a stabilization fund (FESR) and a public-private partnership insurance facility. This catastrophe fund (not the FUNCAP) was approved in 2010 but still requires further regulation. a) Government support programs to private in- surance: PSR (Programa de Subvenção ao Prêmio do Seguro Rural): The PSR subsidizes private agricultural insurance bought from authorised insurers. The national program is fi- nanced by the Ministry of Agriculture and the benefit can be complemented by local (state) programs where available. The share that is subsidized varies according to the food crop and the season. The insurance premium subsidy ranges from 30% to 70%, with a cap of R$ 96 thousand per crop insured. The PRONAMP (Programa Nacional de Apoio ao Médio Produtor Rural) program can add a 20% subsidy to medium-sized producing units, organic crops and selected crops in priority regions. 40 Coping with Losses: Options for Disaster Risk Financing in Brazil Table 4.2. PSR: Evolution of Main Performance Indicators 2005-2012 2005 2006 2007 2008 2009 2010 2011 2012 Market Demand (R$ million) - 61 114 158 272 460 526 717 Approved Budget (R$ million) 10 61 100 160 272 328 406 329 Available (R$ million) 10 61 100 160 172 238 255 329 Total Subsidies (R$ million) 2.3 31 61 158 260 198 254 318 Insured Capital (R$ million) 127 2,869 2,706 7,209 9,684 6,542 7,339 8,782 Insured Area (millions of hc) 0.07 1.56 2.28 4.76 6.67 4.79 5.58 5.24 Insured farmers (unit) 849 16,653 27,846 43,642 56,306 38,211 40,109 43,538 Average subsidies per 2,709 1,867 2,187 3,608 4,610 5,189 6,319 7,307 beneficiary (R$ 1.00) Source: MAPA. However, from 2009 onwards the high demand for fore reduce risks for the private insurers, possibly lead- subsidized agricultural insurance and the lower than ex- ing in addition to improved pricing schemes. pected annual budgetary provision (as projected by the triennial program plans18) have imposed financial diffi- FESR (Fundo de Estabilidade do Seguro Rural): de- culties on the program (Adami and Ozaki, 2012). signed to act as a stabilizing institution, the Fund can provide insurers with financial support if the claims cor- According to Adami and Ozaki (2012), since 2009 the respond to 100% to 150%, or more than 250% of the PSR has delayed due payments to private insurers and premiums. Insurance claims of more than 150% but less the triennial plans are no longer totally relied on for the than 250% of the premium must be covered by reinsur- resources made annually available to the program. In ance contracts since they are not covered by the FESR. 2010, R$ 90 million due in subsidies were not paid to the private insurers. In 2011, the transfers due to the Sources of funding are the budgetary provisions from latter were as high as R$ 163 million. These liabilities the Federal Government in the event of deficit and shrank the resources actually available to the market in profits exceeding technical profitability caps imposed the following years. The budgetary provisions were be- on rural insurance operations. The corresponding Legal low expectations at the same time. framework is the Decree-law nº 73, from November, 21st, 1966. Despite being the major program responsible for in- creasing the penetration of agricultural insurance in Catastrophe Fund: Approved in 2010, the Catastro- Brazil, the financial management of the PSR has been phe Fund is not yet operational due to the lack of regu- a challenging task. Such instability creates additional lation. The idea is to replace the FESR in a way that uncertainty for private insurers, potentially impacting allows for the private sector to participate in the fund. the provision of agricultural insurance for farmers (both The initial proposal indicates an initial investment of R$2 through prices and quantities). billion from Federal Government budgetary resources and R$ 2 billion from bond issues. Policies with the potential to reduce insurance prices, such as improved risk assessment tools and distribution Legal framework: Complementary Law No. 137, Au- channels, and a better regulatory framework for insur- gust 26th, 2010. ers and reinsurers, could reduce the costs of the pro- gram. Moreover, more transparent financial schemes could strengthen the programs reputation and there- 18 The triennial PSR plan gives the guidelines for the agricultural insurance subsidies program and presents estimates of the available funds during the corresponding three years. 41 Coping with Losses: Options for Disaster Risk Financing in Brazil b) Programs fully intervened19 technologies to monitor and forecast cultivated areas and yields. Published reports are available to the indus- PROAGRO (Brazilian Guarantee Program): target- try as a whole. Within the scope of this project, policies ing small and middle sized producing units, this program to improve risk assessment and pricing in the insurance exempts farmers from paying specific financial obligations sector could be further developed. The most recent in the event of natural disasters that reduce farmers pay- version of the published report refers to 2010; more ment capability. up-to-date information about the continuity of the pro- gram is not available. Sources of funding: budgetary provision from the Fed- eral Government, rural producers’ contributions and The agriculture sub-sector is relatively more advanced revenues from financial surpluses. than other areas in terms of data management. Accord- ing to Silva (2011), the introduction of the Climate Ag- The corresponding legal framework is Law 5.969/1973, riculture Risk Mapping Program has changed the focus Law 8.171/1991, Decree 175/1991, NMC (National of policies such as the PROAGRO, which now fosters Monetary Council) Rural Credit Manual (MCR-16). technological innovation and improved risk manage- ment among farmers. According to Assad (2004), the “Proagro Mais”/SEAF (Insurance for Family Agri- program has saved the PROAGRO program up to R$ culture): public insurance facility small sized producer 150 million since the use of agricultural risk maps leads units operating under the PRONAF (Programa Nacional to lower claims being made. de Fortalecimento Agricultura Familiar). The program covers financial liabilities plus a share of the expected revenues in the event of a disaster not occurring. The Optimizing the Government of brazil corresponding legal framework is the NMC (National Monetary Council) Resolution nº 4.186, January 31st, programs through a DRFI strategy 2013. While the number of public policies that support risk Programa Garantia Safra: a welfare program for small management in agriculture shows that the relevance of farmers located in certain specific areas of the north- disaster risk financing in the sector is recognized as a eastern and southeastern regions that pays a fixed matter of public interest, it is important to note that the amount (currently R$ 850) to partially cover losses (if at absence of a coordinated approach to DRFI undermines least 50% of eligible crops are lost) caused by droughts the performance of such programs. or floods in participating municipalities. For example, while producers covered by the above- Bolsa Estiagem: for the farmers not covered by the mentioned fully intervened programs have to comply Garantia Safra, the Bolsa Estiagem is a cash transfer pro- with the Agriculture Risk Mapping recommendations, gram for agricultural producers with monthly income of neither PROAGRO nor Garantia Safra use the informa- up to two minimum wages that are affected by natural tion available to determine pricing or define their risk disasters, including droughts. The program transfers a financing strategy. Instead, financial management of such fixed amount (currently R$ 400) in five installments. programs are based on weak assumptions and parame- ters, such as indemnity payments and contributions that c) Technical Assistance are often politically determined. This results in deficits (related to what is raised with the farmers’ contribu- Climate Agriculture Risk Mapping: a technical assis- tions) financed by the GoB. tance program run by the Ministry of Agriculture that annually maps crops suitable for each region, taking into Private insurers, government institutions and the pri- consideration climate and soil features of the studied ar- vate agribusiness sector moreover complain that al- eas. The program monitors 40 different types of crops though the coverage of PROAGRO and Garantia Safra is in 24 Brazilian states. In order to be considered for the increasing, the actual products and coverage are failing PROAGRO and other public insurance and credit pro- to address producer needs and representing the risks grams producers have to meet with the recommenda- faced by farmers. While the financial performance of tions of the annual risk assessment. these programs could certainly be improved, they could also better focus on the real needs of their customers. GeoSafras: CONAB (Companhia Nacional de Abas- tecimento) uses satellite imagery and other advanced 19 Fully intervened programs, also referred to as pseudo-insurance programs, are those fully funded and managed by the federal government. These programs can involve contributions from the beneficiaries but do not follow regular insurance practices such as for example the use of actu- arially fair prices. They are welfare programs that transfer income to farmers affected by disaster and for this reason are sometimes referred to as “insurance”. 42 Coping with Losses: Options for Disaster Risk Financing in Brazil The design of these programs, while currently mitigat- presented serious disbursement issues over recent ing the financial impact of catastrophes on small farm- years, creating uncertainty for both farmers and private ers, also creates a contingent liability for the govern- insurers. ment. As experience with flood insurance has shown in the United States, this fiscal exposure, if left unchecked In conclusion, while the need for disaster risk financ- or not reinsured, can grow significantly over time. ing in the agricultural sector has been recognized by the GoB, the implementation of a broader DRFI strategy With regard to private agriculture insurance, no clear involving the use of an adequate risk layering approach strategy exists for engaging private insurers and, as dis- and a stronger partnership with the private insurers cussed above, the only program (PSR) with this role could significantly improve the effectiveness and effi- cannot be fully relied upon. The PSR has for example ciency of public resources allocated to these programs. 4.3 property insurance, microinsurance and temporary cash transfers property insurance market While property insurance is the main component of the non-life segment, coverage against natural hazards is usually arranged through comprehensive policies, i.e., policies offering broad coverage against unspecified peril20. One problem associated to this feature is that, since “catastrophe” risk is covered by existing property policies, the current data collected by insurance companies is unlikely to facilitate modeling of catastrophic perils since the premi- ums and losses attributed to catastrophes alone are not separated. Figure 4.6 Catastrophe Insurance Premium by Sub-Sector 6.000 5.000 4.000 3.071 3.000 (R$ million) 2.284 2.000 1.000 784 0 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013* Total Rural Total Property Total Cat Insurance y. Source: SUSEP. *Jan-August onl 20 In figure 4.6, these ‘’all-risks” policies are considered as catastrophe-related property insurance, although their coverage goes beyond that (prop- erty named peril insurance policies are not offered, except for specific engineering and operational risks at e.g. industrial facilities). 43 Coping with Losses: Options for Disaster Risk Financing in Brazil From a more critical public policy viewpoint, this also segment given that, apart from residential insurance, a raises the question of whether local insurance compa- whole range of businesses vulnerable to natural hazards nies are properly pricing for catastrophe risk and setting (from small businesses and service providers to large aside sufficient reserves to pay for future catastrophe- utility companies) are not currently covered. related claims. At the moment natural hazards are gen- erally not explicitly taken into account in the property One of the main bottlenecks pointed out by the insur- insurance market and, within this framework improve- ance industry is the lack of access to relevant informa- ment of the design, pricing and delivery of such prod- tion such as historical disaster losses, current risk maps, ucts is especially challenging. hazard models and climate monitoring tools. Without ac- curate data, adequate pricing in the property insurance Notwithstanding the many problems, private local and sector is not possible and, as a result, conservative pricing international insurers see great potential in this market methods are likely to drive prices up to prohibitive levels. Table 4.3. Estimated outreach of microinsurance: millions of risks covered 2006 2011 Asia 66 350 to 400 Latin America 8 45 to 50 Africa 4-5 18 to 24 Total 78 <500 Source: Munich Re Foundation, 2012 44 Coping with Losses: Options for Disaster Risk Financing in Brazil As for residential insurance, another key issue is that couraging homeowners, farmers etc to engage in risk many low-income potential customers do not possess reduction activities despite being insured. the property titles required for insurance coverage. Given that property insurance is not mandatory in Brazil, In Brazil, up to 2012, microinsurance policies were of- insurers find it difficult to diversify their portfolios. In gen- fered by the few institutions authorized to operate in eral it is only high-risk clients who tend to buy insurance. this segment by SUSEP and the lack of a broader and specific range of rules and regulations precluded other insurance companies from investing in the microinsur- microinsurance ance segment. The microinsurance market expansion is a global trend To fill this gap, SUSEP recently approved a regulatory (Table 4.3) driven mainly by government support and, framework that is expected to open up opportunities while disaster microinsurance could benefit from this for the microinsurance market. The process was given general trend, the role of catastrophe related microin- an additional boost in 2008 when a public-private com- surance in a broader disaster DRFI is still being estab- mission was established to assess the needs of the mi- lished. In developing countries the first pilot programs croinsurance sector. and projects are still being formulated and evaluated by governments and private insurers. As a result, SUSEP published Resolution 244 in 2011. This was subsequently regulated by complementary As with microcredit, developing and consolidating mi- norms in June and September, 2012. In August 2013 the croinsurance programs requires a tailored design to final version of the Microinsurance draft law (3.266/08) tackle key issues such as how to overcome the diffi- that addresses tax exemption issues was approved by culties involved in accessing and evaluating the target the commission responsible for the project and it is now population. A further problem is how to set about en- ready to be submitted to the National Congress (See table 4.4 for details). 45 Coping with Losses: Options for Disaster Risk Financing in Brazil Table 4.4. Recent regulatory framework for microinsurance Date Normative Description 2008 Law Proposition 3,266 Rules tax exemption on microinsurance operations. 2011 CNSP Resolution 244 Overall definitions and guidelines focused on product, market and prudential regulations. Regulates special microinsurance delivery channels. 2012 Susep Circular 439 Establishes requirements and standards to be met by entering micro- insurance companies. 2012 Susep Circular 440 Establishes the general features of microinsurance policies, including marketing policies through remote delivery channels. 2012 Susep Circular 441 Establishes the overall guidelines for the marketing of microinsurance policies by financial institutions and their intermediates. 2012 Susep Circular 442 Regulates microinsurance correspondents. 2012 Susep Circular 443 Regulates the associated use of capitalization-structured products. 2012 CNSP Resolution 262 Establishes technical provisions and collateral assets requirements. 2012 CNSP Resolution 263 Establishes minimum capital requirements. Source: SUSEP. The above regulatory framework explicitly addresses Estiagem is a program available for supporting small non-life catastrophe microinsurance. Property, named farmers suffering losses due to natural hazards that are peril, catastrophe microinsurance can be offered by pri- not covered by the PROAGRO or the Garantia Safra vate insurers under these guidelines: residential build- (previously explained). ings and their contents, as well as commercial build- ings and equipment of small businesses can be insured Many features of these programs (compensations, dura- against wind storms, hurricanes, cyclones, tornadoes, tion, sources of funding and other conditions) vary con- hail, landslides and floods. However, to date these siderably depending on the event and the lack of coor- products have not been authorized and it is not clear dination between them. This could result in high costs whether the industry will offer comprehensive and/or for the GoB since different line ministries and institutions named-peril policies. Microinsurance against income might be targeting the same areas and acting without ex- losses, unemployment and rent liabilities are also de- ploring synergies and economies of scale. At present, it fined in the new regulatory framework, but it is not is difficult to access the financial data on these transfers clear if natural disasters could be named as the events since they are managed by different institutions. causing income loss insured by these policies. The current arrangements for such transfers make it harder for the GoB to optimize the financial manage- temporary cash transfers ment of these programs and create the right incentives for disseminating prevention. Therefore, an exten- Frequently used by the GoB, temporary cash transfers sion of the AAL and PML analysis above to this specific have been used to provide assistance to vulnerable spending, both at the national and subnational levels, populations in the absence of adequate property, agri- would be a concrete step towards an improved DRFI culture and microinsurance. For example, cash transfers strategy. Based on the resulting risk metrics, the line have been made to families that had their dwellings to- ministries and state secretariats involved could in due tally or partially destroyed in the 2008 Santa Catarina course enhance their financial performance. Moreover, floods and in the 2011 Rio de Janeiro landslides. In some this analysis could be used by the Ministry of Planning to circumstances these benefits were extended until the explore the possibilities for adopting a centrally-man- affected families were resettled. In addition the Bolsa aged risk pool as a basis for a broader DRFI strategy. 46 Coping with Losses: Options for Disaster Risk Financing in Brazil 47 Coping with Losses: Options for Disaster Risk Financing in Brazil Chapter 5. Options for Disaster Risk Financing in Brazil D eveloping a national DRFI strategy could increase Brazil’s financial re- sponse capacity to adverse natural events while reducing the associated fiscal burden as response programs become more cost-effective. Brazil may work to strengthen the current risk retention approach by operationalizing a national multi-year disaster fund, by optimizing the insurance subsidies and temporary cash transfers programs, by carrying out detailed fiscal risk assess- ments, developing catastrophic risk models and by improving the current damage and loss assessment procedures and data management systems. 48 Coping with Losses: Options for Disaster Risk Financing in Brazil Chapter 5. Options for Disaster Risk Financing in Brazil Disaster risk financing and insurance is a key pillar of a those associated to small-scale, but more frequent comprehensive DRM (Disaster Risk Management) strat- events) may be appropriately prioritized. Address- egy. As prevention and mitigation efforts cannot fully pro- ing these layers of risk call for a risk retention strategy tect a country against adverse natural events, the main based on reserve funds, budgetary allocations and con- goal of a DRFI strategy is to increase financial response tingent funds. The design of the details of the risk re- capacity while reducing the fiscal burden generated by tention strategy are important to avoid financing gaps the government liabilities associated to natural hazards. and ensure a cost-effective and timely use of disaster response funds, while promoting the right incentives In order to support countries that are vulnerable to nat- for prevention and the adequate conditions for resilient ural hazards to design a DRFI strategy adequate to their reconstruction. hazard profiles, the World Bank has developed a DRFI framework (box 5.1) that takes into account different Within this context, this chapter lists a set of actions to layers of risk that a country might be exposed to and be considered as part of efforts to establish a broader considers the financial instruments that are the most ap- DRFI strategy in Brazil. Establishing a multi-year finan- propriate for financing response to disasters. Countries cial platform, extending fiscal risk assessments, further such as Indonesia, Vietnam, Mexico, Colombia and oth- understanding the governments liability associated to ers have successfully designed and implemented sover- temporary cash transfers and improving damage and eign catastrophe risk financing strategies, as well as risk loss assessment procedures and data management sys- transfer programs. tems could be next steps to be taken by government bodies such as the Ministries of Planning, Finance and Given Brazil’s hazard profile, the low layer of risk (i.e. National Integration. 49 Coping with Losses: Options for Disaster Risk Financing in Brazil Box 5.1 The World Bank Framework for Disaster Risk Financing and Insurance The World Bank has supported countries such as Mexico, Indonesia, Vietnam, and others implementing di- saster risk financing and insurance strategies based on the risk-layering approach shown in the figure below. Figure B.5.1 Three-tier financial strategy against natural disasters Disaster Risk Disaster Risks Financing Instruments Low Major High Risk Layer Disaster Risk Insurance (e.g., major earthquake, (e.g., parametric major tropical cyclone) insurance, cat bonds) Frequency of event Severity of impact Medium Risk Layer (e.g., floods, Contingent credit small earthquake) Low Risk Layer Contingency budget, (e.g., localized reserves annual High Minor floods, landslides) budget allocation Source: World Bank Disaster Risk Financing and Insurance Program 2010. The DRFI framework identifies risk layers according to risks in a particular country and then structures differ- ent financial tools accordingly: - Low risk layer (with a return period of up to 4 years): annual budget allocation and contingency budgets could finance more frequent, recurrent and lower losses. - Medium layer (with a return period of between 4 and 20 years): contingent credit finances less frequent but more severe events. - High risk layer (return period of over 20 years): low frequent and severe risk can be transferred to the private sector. Details of a strategy are tailored to each countrys specific features. In Brazil, evidence suggests that the priority should be to improve financial instruments for the low and medium risk layers, since catastrophic events are not the main component of the country’s hazard profile. 50 Coping with Losses: Options for Disaster Risk Financing in Brazil 5.1 Implementing a national disaster fund The regular occurrence of financing gaps, the excessive reliance on ex-post funding sources, and a still cumbersome process for disbursement at the local level are some of the issues previously highlighted in the current institutional framework for disaster response. The Ministry of Planning (MPOG) by completing the standard operating procedures and by formalizing the funding flow arrangements between FUNCAP and local institutions, such as the state-level calamity funds, would help address these weaknesses. Based on this new platform, the MPOG could promote awareness of fiscal impact, as well as support correct capital reserve funds by proposing annual allocations to calamity funds that reflect the AAL and by ensuring that top-ups made for major events meet reconstruction needs. Building on successful international experiences such as the Mexi- can FONDEN (box 5.2), FUNCAP could become a facility flexible enough to manage disaster related resources in a manner that takes into consideration different liquidity and planning needs, as well as the vertical flow of funds between national and local institutions. One difficulty that has been precluding FUNCAP from becoming operational concerns the sources of funding. As with the previous proposals presented since its establishment, the current proposal also reserves a share of Federal Government revenues to the fund, but in this case indexed by what is raised from lottery revenues. This particular feature still encounters resistance in the political arena, and the current proposal for FUNCAP likewise might not be approved. Therefore, actually operationalizing FUNCAP will necessarily involve a difficult discussion on how to ad- dress the controversial proposal for linking tax revenues to the fund. Monitoring and accountability mechanisms could be strengthened as a means to reduce reconstruction gaps as well. Improved institutional capacity might increase and speed up access to recovery and reconstruction resources, while reducing corruption and waste. These are factors that might be more important determinants of financing gaps for recovery and reconstruction than the availability of resources per se. The use of the Civil Defense card is a successful example. It has been used to release and track resources during the emergency stage. If recovery and reconstruction activities were also financed through mechanisms with features simi- lar in nature to those of the Civil Defense Payment card, the related savings may be significant. 51 Coping with Losses: Options for Disaster Risk Financing in Brazil Box 5.2 FONDEN: Mexico’s National Disaster Fund In order to handle the continuous need for ex post budget reallocations for disaster response, the Government of Mexico established the Fund for Natural Disasters (FONDEN) in 1996 with the aim of providing adequate financial resources for reconstruction without compromising committed government spending. Since then, the fund has evolved from an ex ante budgetary tool to a major component of Mexico’s disaster risk management strategy. Its operation relies on a framework for damage and loss assessments, resources allocation, funding channels and implementation timelines between federal and state institutions. To manage high variability of disaster losses, since 2004 FONDEN has been allowed to use risk transfer instruments. This is part of the Government of Mexico’s DRFI strategy covering different risk of layers to complement the FONDEN’s risk retention coverage. FONDEN finances 100% of the reconstruction of federal assets and 50% of local assets the first time it is required. Since 2006, FONDEN has had a specific program to finance disaster risk reduction activities such as producing risk maps and undertaking small structural interventions. The “build back better” principle that allows for reconstruction at higher standards and the relocation rebuilding of assets was also introduced to FONDEN, as well as an Emergency Relief Fund to cover activities immediately before and after a disaster. A brief comparison between the current FUNCAP structure and the FONDEN is presented in table B.5.2 below. Table B.5.2. FUNCAP versus FONDEN FUNCAP FONDEN Objective Rehabilitation and reconstruction of ar- Relief in aftermath a disaster; eas hit by natural catastrophes. Under Rehabilitation and reconstruction of: special circumstances, relief in the after- • federal and state infrastructure math of a disaster. • low-income houses Funding prevention activities such as: • Risk assessment • Risk reduction • Local community capacity building around disaster prevention Sources • Voluntary transfers from the Federal At least 0.40% of the federal budget of funds Government • Voluntary transfers from states and municipalities • Involuntary transfers from the Fed- eral Government (it must provide three times the funds provided by the local governments) Coverage 100% of reconstruction costs (note that 100% of reconstruction costs for the local governments are responsible federal assets and 50% of those for for 25% of FUNCAP’s funding. local assets (this percentage declines if insurance is not purchased for the reconstructed assets) Funds Municipalities and state governments Federal and State Governments can be (states traditionally restores munici- used by pal assets) Source: World Bank, 2013b and author’s compilation 52 Coping with Losses: Options for Disaster Risk Financing in Brazil 5.2 Optimizing subsidies and temporary cash transfer programs Agricultural insurance is relatively well established in Brazil, and since 2003 a subsidies program (PSR) sup- ports the observed increase in penetration rates. In ad- dition, public transfers solutions for small farmers are also commonly used. Apart from in the agriculture sec- tor, temporary cash transfers are used as a response to damage in the housing sector. Funded by the Federal Government (with some exceptions in the housing sec- tor), linking these programs to a broader DRFI strategy could increase their cost-effectiveness. As mentioned above, financial management of the subsidies program (PSR) has been challenging, which may curtail expansion of agricultural insurance pen- etration in the country. Assessing the government subsidy programs for natural hazard insurance, par- ticularly in the agriculture sector (PSR), and updating these schemes after defining the optimal relationship between insurers, banks, beneficiaries, and the Gov- ernment, could provide the condition for expanding coverage at reduced cost. frequent, events that cannot be analyzed based on his- torical data. Cost-effectiveness in the temporary cash transfers, both in the agriculture and housing sectors, may be The financial risk assessment presented in Chapter 2 enhanced by a better understanding of the contingent is a preliminary analysis based on incomplete historical liabilities (AAL and PML) associated to the cash trans- series of disaster material losses. The data does not ac- fer programs such as the PROAGRO, Garantia Safra, count for many of the less severe but frequent events Bolsa-Estiagem, and the cash transfers for tempo- that characterize Brazil’s hazard profile. Therefore, in rary housing programs. The possibility of consolidat- order to successfully design and implement a broader ing these payments into a stand-alone risk pool that DRFI strategy, it is recommended that the MPOG and can be managed through risk retention or risk transfer MoF carry out an improved fiscal risk assessment, po- schemes may also be considered. tentially focused at the State level. The resulting risk metrics could then be used to inform how transfers to 5.3 Improving Financial local governments should be managed and in the bud- getary management of the FUNCAP . Risk Assessment Beyond assessing damage and losses from actual analysis and developing events, the development of a national exposure da- tabase and hazard modules for major perils would be catastrophe risk models very useful for policy makers interested in assessing the fiscal impacts of specific or possible events. They Countries such as Colombia, Mexico and Honduras could be used, for example, by the Ministries of Plan- have been updating and strengthening their DRFI strat- ning and Finance to develop financial catastrophe risk egies. To do this they have carried out fiscal risk assess- models and use them to prepare annual budgets and ments and developed catastrophe risk models to assess analyze risk transfer strategies. In fact, a recent trend contingent liabilities associated to more severe, but less has been to produce hybrid LEC curves that use 53 Coping with Losses: Options for Disaster Risk Financing in Brazil both historical data and hazard modelling to assess disbursement processes and keep track of disbursed the PML associated to both high frequency and rare resources. The idea is that the use of a damage and events. This information would also be useful for pri- loss assessment methodology based on objective cost vate insurers interested in offering affordable catas- estimates can support the objective allocation of funds trophe insurance to improve product development and reduce political competition for resources in the and pricing methods. This would be a more costly aftermath of a disaster: when the criteria are not well initiative however that, if not feasible in the short- defined, line Ministries and local Governments might run, could be maintained as a medium-term goal. overestimate their funding needs knowing that the requested amount will not be fully allocated. If an ob- jective damage and loss assessment methodology and 5.4 Enhancing disaster resources tracking systems are adopted, rationalizing data management systems post-disaster funding allocation and disbursement could become an easier task. and damage and loss Once again Mexico provides a good example of how assessment procedures improving damage and loss assessment procedures has helped to strengthen the country’s response capacity. In Mexico, a damage assessment committee (DAC) is Adopting an objective damage and loss assessment established before the declaration of natural disaster is methodology would provide both (i) a clear guideline formally issued (24 hours following the technical confir- for allocating disaster response funds, and (ii) the ap- mation of a disaster). The DAC comprises sectorial sub- propriate basis for the data collection efforts that are committees composed of members of both federal and needed to keep track of the historical material losses, local institutions. Each subcommittee has a maximum which in turn are inputs required to design and update of 20 working days to complete fieldwork, to docu- a DRFI strategy. ment and photograph (using GIS devices) damage, and Allocating response resources based on damage and to itemize reconstruction needs and the related costs. loss assessment methodology can help to speed up While the full damage assessment is being completed, 54 Coping with Losses: Options for Disaster Risk Financing in Brazil federal and local agencies have 7 days to request funding might be sensitive to changes in these hypotheses, with for the emergency phase, which guarantees immediate possible implications on how a financial strategy for di- liquidity and at the same time enables the DAC to com- saster response should be calibrated to the country’s plete the full damage and loss assessment free from the hazard profile. A consolidated data set on disaster ma- pressure for immediate resources. terial losses could therefore be a valuable input in the design of a DRFI strategy. The MI has already taken important steps in this same direction: a simplified version of the damage assessment The private insurance sector could also benefit by im- form has been recently implemented to allow for timely proving agriculture and property insurance pricing, ex- provision of funds in the immediate aftermath of a disas- panding coverage and reducing prices. As a result, (i) ter. A series of training sessions on the DaLA (Damage the PSR (agricultural insurance subsidies) could become and Loss Assessment Methodology) was provided to more cost-effective and (ii) a higher coverage of prop- the Finance and Planning State Secretariats, as well as to erty insurance in the commerce and industry sectors local Civil Defense institutions. Meanwhile discussions could also reduce the Governments liability regarding on standardizing damage and loss assessment proce- these sectors. dures continue at the national level. Future discussions could focus on how to establish a standard damage and By consolidating the historical material losses this would loss assessment procedure, how to track historical ma- also inform the MI about the main issues involved in the terial losses, and how to ensure that this information current damage and loss assessment procedures: as is made available to planning and finance institutions to flaws in the historical series are detected, the current facilitate designing, updating and assessing of a national system could be updated to prevent these same flaws DRFI strategy for Brazil. from infecting future data. The Ministry of National In- tegration has already taken steps to build disaster his- As for the availability and quality of data on historical torical data, but only information on the human toll has material disaster losses, it was necessary to rely on very been made available to date. Taking the next step and restrictive assumptions in order to carry out the pre- providing data on historical material losses could gener- liminary analysis presented above. Despite every effort ate the positive abovementioned externalities. to make reasonable assumptions, the results obtained 55 Coping with Losses: Options for Disaster Risk Financing in Brazil References Adami, A., Ozaki. V. 2012. Modelagem estatística dos De Moura, F. S.; Pedroso, F. F.; Toro, J. 2012. “2011 Rio prêmios do seguro rural. Revista de Política Agrícola. de Janeiro State Flash Floods and Landslides: Damage and Loss Assessment”. World Bank Working Paper Assad, E. 2004. O Zoneamento Agrícola e os Riscos 73623. Climáticos. Embrapa. De Moura, F. S.; Pedroso, F. F.; Toro, J. 2012. “2010 Per- Axco, 2013. Brazil –Insurance Market Information: nambuco State Flash Floods: Damage and Loss Assess- non-life. ment”. World Bank Working Paper 73622. Beckwith, R., Boudreau, L., Lane., M.,Mahul, O. White, De Moura, F. S.; Pedroso, F. F.; Toro, J. 2012. “2010 Ala- E. Innovation in Disaster Risk Financing for Develop- goas State Flash Floods: Damage and Loss Assessment”. ing Countries: Public and Private Contributions. World World Bank Working Paper 73621. Bank/GFDRR, 2011. De Moura, F. S.; Pedroso, F. F.; Toro, J. 2012. “2008 San- Bloch, R., Jha, A., Lamond, J. Cities and Flooding: A ta Catarina State Flash Floods and Landslides: Damage Guide to Integrated Urban Flood Risk Management for and Loss Assessment”. World Bank Working Paper. the 21st Century. Banco Mundial/GFDRR. 2012. Ernst & Young , 2013. Latin America insurance outlook: BRASIL. Constituição (1988). Constituição da República Poised for growth. Federativa do Brasil.Brasília, DF, Senado,1998. DOS SANTOS, Iracy Monteiro. Efetividade do Anexo BRASIL. MINISTÉRIO DA FAZENDA. SECRETARIA de Riscos Fiscais como instrumento de transparência DO TESOURO NACIONAL. Manual de Despesa Na- dos passivos contingentes, no framework da União. In- cional. Brasília: STN/Coordenação-Geral de Contabili- stituto Serzedello Corrêa – ISC/TCU, Brasília, DF, 2010. dade, 2008. 120p. GAR. Global Assessment Report on Disaster Risk Re- BRASIL. MINISTÉRIO DO PLANEJAMENTO, ORÇA- duction. From shared risk to shared value: the business MENTO E GESTÃO. SECRETARIA DE ORÇAMENTO case for disaster risk reduction. 2013 FEDERAL. Orçamentos da União Exercício Financeiro 2001: Projeto de Lei Orçamentária. Brasília: MPOG/ Ghesquiere, F., Mahul, O. Financial Protection of the SOF, v. 2, 2000. State against Natural Disasters. Policy Research Work- ing Paper n. 5429, Banco Mundial, Setembro de 2010. BRASIL. SECRETARIA DO TESOURO NACIONAL. Manual de demonstrativos fiscais : aplicado à União e GOBETTI, Sérgio Wulff. Estimativa dos Investimentos aos Estados, Distrito Federal e Municípios. Ministério Públicos: Um Novo Modelo de Análise da Execução da Fazenda, Secretaria do Tesouro Nacional. – 5. ed. Orçamentária Aplicado às Contas Nacionais. Brasília: – Brasília : Secretaria do Tesouro Nacional, Coorde- ESAF, 2006. 51p. Award-winningmonograph in 1st place nação-Geral de Normas de Contabilidade Aplicadas à in the XI NationalTreasureAward – 2006, Tributação, Federação, 2012. 699 p. : il. ; 28 cm. Orçamentos e Sistemas de Informação sobre a Admin- istração Pública, Brasília (DF). Buainain, A.; Vieira, P. 2011. Seguro Agrícola no Brasil: desafios e potencialidades. Revista Brasileira de Risco e Gunawan, I., Mahul, O., et al. Indonesia: Advancing a Seguro. National Disaster Risk Financing Strategy – Options for Consideration. Banco Mundial/GFDRR. 2012. Cummins, J., Mahul, O.Catastrophe Risk Financing in Developing Countries: Principles for Public Interven- IPEA , 2011. Agricultura: do subsídio à política agrícola. tion. Banco Mundial, 2009. Ano 8, Edição 68. 56 Coping with Losses: Options for Disaster Risk Financing in Brazil ISTOÉ Dinheiro, 2012. Os gigantes querem os mi- Guia_Siafi_Gerencial_v1_2010.pdf>.Access in June 7, crosseguros. 2012. Jha, A., Mahul, O., et al. “Advancing Disaster Risk Financ- SECRETARIA DO TESOURO NACIONAL. Siafi Geren- ing and Insurance in ASEAN Countries: Framework and cial: Itens e Grupos de Informação Recomendados. Bra- Options for Implementation”, GFDRR Draft document sília: STN/Subsecretaria de Contabilidade Pública, 2012. for discussion with ASEAN Member States. 2011. SENADO FEDERAL. Orçamento da União/Processo Linnerooth-Bayer at al., 2006. Insurance for assisting Orçamentário. Disponível em: . Access in: May 24, 2012. Lloyd´s, 2012. Global underinsurance report. Wang, X. Mahul, O., Stutley C. Weathering the Storm: Options for Disaster Risk Financing in Vietnam. GFDRR. Mahul, O. Financial Solutions for Catastrophe Risk Man- 2011. agement. World Bank. World Bank, 2010. Agricultural Insurance in Latin Amer- Microinsurance Network/Munich Re Foundation, 2012. ica: Developing the Market. Protecting the poor: a Microinsurance Compendium. World Bank, GFDRR, UNISDR. The Role of Hydro- MINISTÉRIO DA FAZENDA. Manual de Demonstrati- meteorological Services in Disaster Risk Management. vos Fiscais. Secretaria do Tesouro Nacional (STN), 4ª 2012. ed., Brasília, DF, 2011. World Bank, 2011b. GFDRR, ISDR. Disaster Risk Man- MINISTÉRIO DA INTEGRAÇÃO NACIONAL. Manual agement Programs for Priority Countries. de Planejamento em Defesa Civil. Secretaria de Defesa Civil (SEDEC), vol. 1, Brasília, DF, 1999. World Bank, 2011a. Indonesia: Advancing a National Disaster Risk Financing Strategy: Options for Consider- MINISTÉRIO DA INTEGRAÇÃO NACIONAL. Atlas ation. Jakarta and Washington, DC. Brasileiro de Desastres Naturais, Brasília, DF, 2012. World Bank, 2013a. De Moura, F. S.; Pedroso, F. F.; Toro, MINISTÉRIO DA INTEGRAÇÃO NACIONAL. Retro- J. 2013a. “The Economic Impacts of Adverse Natural spectiva Histórica da Evolução da Defesa Civil no Brasil. Events in Brazil”. Available in: . Access in April 23, 2012. World Bank, GFDRR, 2013b. “FONDEN: Mexico’s Na- tional Disaster Fund” Policy Note. MINISTÉRIO DO PLANEJAMENTO, ORÇAMENTO E GESTÃO. Manual Técnico de Orçamento - MTO. Sec- retaria de Orçamento Federal (SOF), Brasília, DF, 2012. PRESIDÊNCIA DA REPÚBLICA. Legislação. Serial Links .Availablein:. Access in April 24, 2012. SECRETARIA DO TESOURO NACIONAL. Siafi Ger- encial: Itens e Grupos de Informação Recomendados. Coord. De Contabilidade, v. 1, 2010. Available in: