80846 Indonesia Advancing a National Disaster Risk Financing Strategy – Options for Consideration October 2011 Disaster Risk Financing and Insurance Program, GFDRR and FCMNB East Asia and Pacific Disaster Risk Management Team Indonesia Country Management Unit East Asia Finance and Private Sector Unit Global Facility for Disaster Reduction and Recovery Indonesia Advancing a National Disaster Risk Financing Strategy – Options for Consideration October 2011 Disaster Risk Financing and Insurance Program, GFDRR and FCMNB East Asia and Pacific Disaster Risk Management Team Indonesia Country Management Unit East Asia Finance and Private Sector Unit Global Facility for Disaster Reduction and Recovery © 2012 The International Bank for Reconstruction and Development/The World Bank 1818 H Street NW Washington DC 20433 Telephone: 202-473-1000 Internet: www.worldbank.org All rights reserved This publication is a product of the staff of the International Bank for Reconstruction and Development/The World Bank. The findings, interpretations, and conclusions expressed in this volume do not necessarily reflect the views of the Executive Directors of The World Bank or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, 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 The material in this publication is copyrighted. Copying and/or transmitting portions or all of this work with- out permission may be a violation of applicable law. The International Bank for Reconstruction and Develop- ment/The World Bank encourages dissemination of its work and will normally grant permission to reproduce portions of the work promptly. For permission to photocopy or reprint any part of this work, please send a request with complete informa- tion to the Copyright Clearance Center Inc., 222 Rosewood Drive, Danvers, MA 01923, USA; telephone: 978-750-8400; fax: 978-750-4470; Internet: www.copyright.com. All other queries on rights and licenses, including subsidiary rights, should be addressed to the Office of the Publisher, The World Bank, 1818 H Street NW, Washington, DC 20433. Cover design: Miki Fernandez Table of Contents Acknowledgements...................................................................................................................... iv Abbreviations and Acronyms....................................................................................................... v Executive Summary...................................................................................................................... 1 Chapter 1. Introduction................................................................................................................. 5 Chapter 2. Fiscal Management of Natural Disasters........................................................................ 9 Chapter 3. Fiscal Disaster Risk Assessment..................................................................................... 17 Chapter 4. Review of Private Catastrophe Risk Insurance Market in Indonesia................................ 25 Chapter 5. Options for National Disaster Risk Financing Strategy in Indonesia................................ 29 Annexes Annex 1. Exposure to Natural Hazards in Indonesia..................................................................... 46 Annex 2. Post-Disaster Risk Financing – Indonesia Case Studies................................................... 48 Annex 3. Property Catastrophe Insurance in Indonesia................................................................ 53 Annex 4. Insurance of Public Assets in Indonesia......................................................................... 55 Annex 5. Disaster Risk Financing and Insurance Framework......................................................... 57 Annex 6. World Bank Development Policy Loan with Catastrophe Draw Down Option................ 62 Annex 7. Parametric insurance – Basic concepts.......................................................................... 64 Annex 8. Mexican Natural Disaster Fund FONDEN....................................................................... 72 Annex 9. Catastrophe Bonds in Mexico....................................................................................... 76 Annex 10. Caribbean Catastrophe Risk Insurance Facility............................................................... 78 Annex 11. Turkish Catastrophe Insurance Pool.............................................................................. 80 Annex 12. Disaster Risk Management Roles in Indonesia............................................................... 81 Annex 13. BNPB Database on Natural Disasters............................................................................. 84 Annex 14. Post-Disaster Operational Phases.................................................................................. 86 Annex 15. Disaster Risk Management Framework in Indonesia...................................................... 87 Annex 16. Borrowing capacity of Indonesia................................................................................... 89 Acknowledgements Acknowledgements T he report was produced by a team led by Olivier Mahul (FCMNB and GFDRR, World Bank) and Iwan Gunawan (EASIS, Word Bank), and comprising Djauhari Sitorus (EASFP, World Bank), Suntan Hidayat (Insurance Specialist, Consultant), Ahmad Zaki Fahmi (Public Finance Specialist, Consultant), Bambang Soetono (EASID, World Bank), Hari Purnomo (EASPR, World Bank), Charles Scawthorn (Catastrophe Risk Modeling Specialist, Consultant), Marc Forni (Financial Specialist, Consultant), and Daniel Clarke (Actuary, Consultant). The report greatly benefited from the data and information provided by the Ministry of Finance – Fiscal Policy Office (BKF), Bureau of Insurance, Capital Market Financial Institution Supervisory Agency (Bape- pam LK), Director General (DG) Budget, DG State Assets – BNPB and BAPPENAS. The report is the result of an extensive stakeholder consultation, including high-level meetings with Ministers and Vice Ministers, DG, and Parliament Members. It also builds on consultations with local govern- ments in Padang, Aceh and Yogyakarta. The team is grateful to the peer reviewers Francis Ghesquiere (SASDU, World Bank), Xiaolan Wang (ECSS6, World Bank) and Issam Abousleiman (TRE, World Bank). The team also thanks Chris Hoban (EACIF) for his comments. The report has been prepared under the overall guidance of Stefan Koeberle (Country Director, World Bank), Franz Drees-Gross (EASIS, World Bank), Loic Chiquier (FCM, World Bank), Shubham Chaudhuri (EASPR, World Bank), Subrahmanya Pulle Srinivas (EASFP, World Bank) and Abhas Jha (EASIN, World Bank). The team gratefully acknowledges funding support from the Global Facility for Disaster Reduction and Recovery (GFDRR). Abbreviations and Acronyms AAUI Indonesia General Insurance Association AEL Annual Expected Loss AIDRF Australia-Indonesia Disaster Risk Facility BAKORNAS National Coordinating Board for Disaster Management Bapepam LK Bureau of Insurance, Capital Market Financial Institution Supervisory Agency BAPPENAS National Development Planning Agency BBB Build Back Better BKF Ministry of Finance – Fiscal Policy Office BLU A non-profit public service entity (Badan Layanan Umum) BNPB National Agency for Disaster Management BPBD Regional Agency for Disaster Management CARICOM Caribbean Common Market and Community Cat DDO Catastrophe Draw Down Option CCRIF Caribbean Catastrophe Risk Insurance Facility CENAPRED Mexican National Center for Disaster Prevention DANA Damage and Needs Assessment DG Director General DPL Development Policy Loan ECLAC Economic Commission for Latin America and the Caribbean ENSO El Niño-Southern Oscillation FONDEN Mexican Fund for Natural Disasters GDP Gross Domestic Product GFDRR Global Facility for Disaster Risk Reduction and Recovery GoI Government of Indonesia GR Government Regulation IMDFF-DR Indonesia Multi Donor Funding Facility for Disaster Recovery MoF Ministry of Finance NAP-DRR National Action Plan for Disaster Risk Reduction NDRF Natural Disaster Reserve Fund PBB Land and Property Tax PDNA Post Disaster Needs Assessment PML Probable Maximum Loss SINAPROC Mexico National Civil Protection System SPV Special Purpose Vehicle TCIP Turkish Catastrophe Insurance Pool TSI Total Sum Insured <1> Executive Summary This study presents options for a national disas- allocation and to improve the timely disbursement ter risk financing strategy in Indonesia, draw- of its post-disaster financial assistance. ing heavily on international experience. The study discusses a series of complementary options This technical assistance is part of the broader for a national disaster risk financing strategy, based partnership with the GoI on disaster risk man- on a preliminary fiscal risk analysis and a review of agement and climate change adaptation. The adoption of Law 24/07 on Natural Disaster Man- the current budget management of natural disasters agement emphasizes the importance of disaster risk in Indonesia. It benefits from the international ex- management. The National Action Plan for Disaster perience of the World Bank, which has assisted sev- Risk Reduction 2010-2012 includes the design and eral countries in the design and implementation of implementation of a national disaster risk financing sovereign catastrophe risk financing strategies (for strategy within a three year time frame. instance, in Mexico, Colombia, Vietnam, Philippines, and the Caribbean island states) and property catas- The potential cost of a major disaster in Indo- trophe risk insurance programs (for instance, in Tur- nesia could exceed 3 percent of GDP. While the key, Romania and Eastern Europe). This experience annual economic impact of natural disasters is es- is tailored to the institutional, social and economic timated at 0.3 percent of Gross Domestic Product characteristics of Indonesia. (GDP) over the last decade, simulations show that a major earthquake (occurring once every 250 years) The Government of Indonesia (GoI) requested could cause losses in excess of US$30 billion, that is, the World Bank’s technical assistance to im- 3 percent of GDP of Indonesia. Damage and loss as- prove its financial response capacity in the sessment reports from recent major disasters show aftermath of natural disasters. The World Bank a consistent ranking of reconstruction needs with has assisted the Ministry of Finance in developing housing accounting for the largest expenditures a national disaster risk financing strategy for the followed by public infrastructure (primarily roads, financial protection of the state against natural di- schools and health facilities). sasters. It has also assisted the National Agency for Disaster Management (BNPB) in improving timely The Rehabilitation and Reconstruction Fund is post-disaster assistance funding mechanism. The the main budget instrument for the GoI to fi- World Bank has worked closely with Ministry of nance public post-disaster expenditures, but it Finance (Fiscal Policy Office (BKF), Bureau of Insur- is under-capitalized. Post-disaster reconstruction ance, Capital Market Financial Institution Supervi- is largely funded through the reserve of the State’s sory Agency (Bapepam LK), DG Budget, DG State General Treasury (Bendahara Umum Negara), which Assets), National Agency for Disaster Management requires parliamentary approval. An annual alloca- (BNPB) and BAPPENAS. tion of about IDR 4 trillion (US$450 million) was made through this process in 2010 and 2011. While The BNPB also requested the World Bank’s tech- this represents a 30 percent increase from 2009, it nical assistance to explore options to improve may still be insufficient to deal with a major catas- its current post-disaster assistance funding trophe or a series of moderate to severe disasters mechanism. The World Bank has assisted BNPB in in a given fiscal year. More importantly, a budget building its capacity to conduct Post Disaster Needs re-appropriation is required after almost every disas- Assessment (PDNA) as one of the basis for funding ter. The study estimates that the immediate liquidity < 2 > Indonesia: Advancing a National Disaster Risk Financing Strategy - Options for Consideration required for public post-disaster recovery spending strategy could be considered by the GoI (see Figure could exceed US$2billion in major disaster years. 1 below): ■■ Increasing the annual budget allocation up This study presents an optimal combination to US$500 million for post-disaster rehabili- of risk-retention and risk transfer instruments tation and reconstruction; that could help the GoI increase its immedi- ate financial response capacity against natural ■■ Securing a contingent credit line of US$500 disasters and better protect its fiscal balance. million; Building on the three-tier risk layering approach pro- moted by the World Bank and the preliminary fis- ■■ Purchasing (parametric) catastrophe risk cal risk assessment analysis, the following financial coverage (e.g., insurance and/or cat bonds) of US$800 million. Figure 1. Indicative disaster risk financing strategy for Indonesia Disaster Risks Disaster Risk Financing Instruments Low Major High risk layer Catastrophe risk transfer (e.g., major earthquake, major tropi- (e.g., parametric insurance, cal cyclone) cat bonds) Frequency of Event Severity of Impact Medium risk layer Contingent credit (e.g., floods, small earthquake) Low risk layer Contingent budget, reserves, an- High Minor (e.g., localized floods, landslides) nual budget allocation Source: Authors. This strategy would provide the GoI with access medium risk layer (with a return period between 4 to immediate liquidity in the aftermath of a di- and 20 years). Catastrophe risk transfer solutions saster at a competitive cost. The strategy would have proven to be cost-efficient against major disas- allow the GoI to access up to US$1.8 billion liquidity ters (high risk layer with a return period of more than in the aftermath of a disaster in order to finance im- 20 years). Should the GoI want to set up a multi-year mediate post-disaster expenditures, such as grants reserve fund, contracting contingent credit could al- for livelihood and low income housing reconstruc- low the multi-reserve fund to retain additional risk tion. Preliminary disaster fiscal risk assessment analy- and purchase less reinsurance, thereby more than sis shows that this would protect the GoI against di- doubling the reserves at the end of ten years. sasters occurring every 100 years. The combination of reserves, contingent credit and parametric insur- A National Disaster Reserve Fund (NDRF) with ance offers a cost-effective strategy. Reserves and/ a fast-disbursement mechanism could be es- or annual budget allocation are efficient to finance tablished as a vehicle for the rapid financing of recurrent low severity events (low risk layer with a public post-disaster reconstruction operations. return period of 4 years or less). Contingent credit Recent experience shows that it can take several is more cost-effective than risk transfer solutions for months to draw down funds from the Rehabilita- Executive Summary < 3 > tion and Reconstruction Fund under the State’s Gen- by the Ministry of Finance and the BNPB. It could eral Treasury. This generates delays in post-disaster rely on a combination of financial instruments (in- recovery operations, including the livelihood and cluding reserves, contingent credit and insurance) housing compensation programs. The GoI could es- to respond quickly to the post disaster needs. This tablish a NDRF, akin to a financial trust, which would Facility could build on the successful example of the disburse funds quickly after a disaster to allow for national disaster fund, Mexican Fund for Natural Di- rapid post-disaster operations. It could be managed sasters (FONDEN). Figure 2. National Disaster Reserve Fund for Indonesia Ministry of Finance BNPB Line Ministries National Disaster Reserve Fund (NDRF) Fund drawn from On-Call-Fund Sectoral Disaster Contingency Reserve Contingency Fund Parliamentary in State General Treasury Process Disaster Contingency Fund Non-profit Agency within Gov- Bridging Funding Facility ernment (BLU) for Rapid Recovery for Disaster Financing The NDRF could be established through a the purpose of providing public goods and/or servic- non-profit entity BLU under the existing legal es. An NDRF-BLU could be considered as an option framework. The existing legal structure, as outlined in the short-term. The BLU could initially finance and in Law 24/2007 on Natural Disaster Management, manage social assistance grants for livelihood and offers an enabling framework for the establishment housing recovery. Initial funding for the BLU could of the proposed NDRF, through the amendments of come as capital contribution from the regularization the Government Regulation Number 22/2008. Sev- of a fraction of the current Rehabilitation and Re- eral options could be considered for the financial construction Fund. management of the NDRF under the current legal and regulatory framework, including the expansion A funding facility could be set up to provide of the scope of the current Contingency Fund and bridge financing for post-disaster rapid recov- the integration with the On-Call Fund as well as sec- ery. In addition to the proposed BLU, GoI could set toral and sub-national contingency budgets. Under up a Bridging Funding Facility for Post Disaster Rapid Law 1/2004 of State Treasury, a non-profit entity Recovery. This Facility could be initially financed from Badan Layanan Umum (BLU) could be established a contingent credit. This line of credit could be used within a line ministry or government agency with as a revolving fund, where funds would be drawn < 4 > Indonesia: Advancing a National Disaster Risk Financing Strategy - Options for Consideration down when budget liquidity is insufficient (usually at The GoI could establish a Joint Disaster Reserve the end of fiscal year) and it would be replenished at Fund for Indonesia’s local governments. The the beginning of the following fiscal year. The Facil- Fund would build on risk diversification to offer In- ity could provide funding to the BLU or to other line donesian provinces/municipalities access to immedi- ministries and/or provincial and local governments ate non-earmarked liquidity in case of disasters at to undertake urgent recovery activities that cannot the lowest possible cost. Building on the successful be delayed until a new fiscal year starts. example of the Caribbean Catastrophe Risk Insur- ance Facility (CCRIF), the Fund would act as a joint The GoI could support the establishment of a reserve mechanism for the provinces/municipalities disaster risk insurance program for key public of Indonesia. Participating local governments would assets in partnership with the private insurance contribute to the Fund based on their own risk pro- industry. Most of the public assets, including criti- file and desired coverage level. These contributions cal assets such as hospital and schools, are not cur- would be used to maintain a reserve level sufficient rently insured against natural disasters. This program to absorb annual payouts to local governments af- would aim to offer technical assistance to the public fected by adverse natural events. The Fund could entities in the design of their catastrophe insurance also benefit from initial contributions from the cen- coverage of public assets. Standardized terms and tral government and/or the donor partners. To man- conditions for the property insurance policies would age the potential variability in financial outflow from be developed, which would assist public managers the fund, the Fund could secure additional financial in identifying their risk exposure and their insurance capacity on the international reinsurance and capital needs. The program could also structure a national markets. insurance portfolio of public assets to be then placed on the private (re)insurance market. A national prop- The implementation of a national disaster risk erty catastrophe insurance program for public assets financing strategy would require significant in- would create economies of scale and diversification stitutional capacity building. Disaster risk financ- benefits, thus lowering reinsurance premiums. It ing is one component of a comprehensive fiscal risk would also provide incentives to the local entities to management strategy, which requires specific finan- report their assets to the Central Government. cial and actuarial expertise. Major capacity building on disaster risk assessment and management of The GoI could promote property catastrophe natural disasters would be required to develop and insurance of private residential dwellings. use financial tools to guide the GoI in its national Building on the example of Turkey, the GoI could disaster risk financing strategy. strengthen its partnership with the private insurance sector to further develop property catastrophe in- surance of private residential dwellings. In particular, the insurance supervision of property catastrophe insurance could be strengthened. Chapter 1 Introduction Indonesia is highly exposed to natural disasters. Indonesia is situated in one of the world’s most active disaster hot spots, where several types of disasters such as earthquakes, tsunamis, volcanic eruptions, floods, landslides, droughts and forest fires frequently occur. <5> < 6 > Indonesia: Advancing a National Disaster Risk Financing Strategy - Options for Consideration According to a global risk analysis by the World The annual economic impact of natural disas- Bank, Indonesia is among the top 35 countries that ters is estimated at 0.3 percent of Indonesia’s have high mortality risks from multiple hazards. Ap- GDP. Over the last 10 years, the annual average cost proximately 40 percent of the population at risk, of natural disasters in Indonesia is estimated at 0.3 that is, more than 90 million lives. The increase in percent of national GDP or US$1.5 billion (EMDAT population and assets exposed to natural disasters, CRED). A summary of losses to major natural disas- combined with the rise in the number and inten- ters over the past decade can be found in Figure 1.1 sity of hydro-meteorological events resulting from below. It is estimated that a major earthquake (oc- climate change, may further increase the economic curring once every 250 years) could cause losses in and human impact of natural disasters in Indonesia. excess of 3 percent of national GDP. See Annex 1 for a description of key hazards in In- donesia. Figure 1.1. Estimated annual losses caused by major natural disasters in Indonesia, 2000-2009 6,000 5,000 4,000 US$ million 3,000 2,000 1,000 – 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 n Economic losses n 2009 GDP adjusted economic losses Source: Authors, from EMDAT CRED and World Bank. Note: Years 2003, 2005 and 2006 showing no disaster losses is the consequence of the definition of disasters by EMDAT CRED, which reports only major disasters. Hence the economic impact of an accumulation of small disasters is not reported in the EMDAT CRED database. Natural disasters have a larger economic im- earthquake in the province of Yogyakarta caused pact at local and sub-national levels. The eco- losses estimated at 30 percent of regional GDP. The nomic impact of the 2004 earthquake in the prov- regional economic impact of recent disasters can be ince of Aceh was estimated at US$4.5 billion (i.e., found in Table 1.1. one percent of national GDP), which represents 54 percent of the provincial GDP. Likewise, the 2006 Chapter 1: Introduction < 7 > Table 1.1 Impact of selected natural disasters on regional GDP Estimated losses Estimated Losses Event Province (US$ billions) (% regional GDP) Tsunami (2004) Aceh 4.5 54% Earthquake (2006) Yogyakarta 3.1 41% Earthquake (2009) West Sumatra 2.3 30% Source: Authors, from EMDAT CRED and World Bank. The high frequency of disasters has an impor- tion of a national disaster risk financing strate- tant impact on public expenditures. According gy. Indonesia is among the first developing countries to the GoI disaster data1, between 2001 and 2007 in Asia to formulate a national action plan for disas- alone, there have been more than 4,000 occur- ter risk reduction (NAP-DRR). The second NAP-DRR rences of disasters including floods (37 percent), covering the period 2010-2012 calls for the imple- droughts (24 percent), landslides (11 percent), and mentation of a national disaster risk financing strat- windstorms (9 percent). Loss to public infrastructure egy within a three year time frame. The strategy will and private dwellings, mostly uninsured, has created include budget reserve funds and disaster risk trans- a major burden on public expenditure to restore af- fer instruments such as insurance. fected facilities. At the request of the GoI, the World Bank has provided technical assistance for the develop- Indonesia’s Law on Natural Disaster Manage- ment of a national disaster risk financing strat- ment of April 2007 provides a framework for egy. The non-lending technical assistance aims to: disaster risk management in Indonesia. The i) assess the fiscal exposure of the GoI to natural adoption of Law 24/07 on Natural Disaster Man- disasters; ii) propose options for the development agement emphasizes the importance of disaster of a national strategy for the financial protection of risk management for the GoI. The Disaster Manage- the state against natural disasters; and iii) promote ment Law established a dedicated agency to deal property catastrophe risk insurance for public and with disasters, the National Disaster Management private dwellings. This technical assistance is part of Agency (BNPB), where previously only an ad-hoc the broader assistance of the World Bank to the GoI inter-ministerial council existed. BNPB is empowered on disaster risk management and climate change with a strong mandate to coordinate the line min- adaptation. istries in implementing preventative measures and leading recovery from the impact of disasters. In line Disaster risk financing and insurance is one with the law, all 33 provinces and 306 districts (out of the five pillars in the proactive and strate- of 450+) have established a disaster management gic framework for disaster risk management agency. However not all established agencies have (DRM) promoted by the World Bank. The World been given proper budget, staffing and equipment. Bank has been promoting a pro-active and strate- As of July 2011, this process is still ongoing. gic framework for DRM. This framework is based on The National Action Plan for Disaster Risk Re- five pillars: (i) risk assessment, (ii) institutional capac- duction 2010-2012 promotes the implementa- ity building, (iii) risk reduction investments; (iv) emer- gency preparedness; and (v) disaster risk financing and insurance. Despite prevention and mitigation 1 DiBi database (Data and Information on Disaster in Indone- efforts, no country can fully insulate itself against sia), National Disaster Management Agency (BNPB). http:// dibi.bnpb.go.id/. major natural disasters. Disaster risk financing and < 8 > Indonesia: Advancing a National Disaster Risk Financing Strategy - Options for Consideration insurance allows countries to increase their financial larly on the fiscal impact of natural disasters. Chap- response capacity in the aftermath of a disaster and ter 3 provides an overview of the budget process for to reduce the economic and fiscal burden of natural the financing of natural disaster losses for each of disasters by devising financial strategies combining the three post-disaster phases: emergency response, post-disaster financing (for example, post disas- recovery, and reconstruction. Chapter 4 provides a ter credit) and ex ante risk financing (for example, short description of the state of the private catas- reserves, contingent credit and risk transfer instru- trophe insurance market. Chapter 5 is devoted to a ments like insurance). review of options for the future financing of natural disaster recovery and reconstruction expenditures in This report presents the main findings and rec- Indonesia. This section includes options for sovereign ommendations of the technical assistance. It risk financing and for the promotion of commercial consists of five chapters including this introduction. catastrophe insurance for the private property sec- Chapter 2 presents a preliminary financial disaster tor. The report is complemented by fourteen techni- risk assessment for Indonesia and focuses particu- cal annexes that offer further analyses and results. Chapter 2 Fiscal Management of Natural Disasters <9> < 10 > Indonesia: Advancing a National Disaster Risk Financing Strategy - Options for Consideration A regulatory framework for post disaster fi- from recent major disasters show consistent ranking of nancing has been established by the GoI and reconstruction needs, with housing accounting for the is described under Law 24/2007. It provides the largest needs, followed by public infrastructure (pri- definition of natural disasters and identifies the re- marily roads, schools and health facilities). sponsibilities of the central and local governments as well as the functions and duties of the National The source of emergency response funds, cov- and Regional Disaster Management Agencies. The ering the first weeks after a disaster, depends regulation outlines the disaster risk financing frame- on whether the event is declared a National work, which is a shared responsibility between the Disaster or disaster of national significance3. central and local governments, stipulating the three If so, the Central Government takes responsibility phases of a disaster as: emergency, recovery, and re- through BNPB with line ministries and the BNPB dis- construction. Additional provisions not included in bursing resources through their “On call” funds for Law 24/2007 regarding the management of disaster emergency response. “On call” funds are a separate events follow Government Regulation (GR) 21/2007 line of the budget that can be engaged to support on disaster response, and GR 22/2007 on financing post-disaster early recovery activities while emer- and management of natural disasters. Finally, Law gency status is still in effect. If not declared a Na- 33/2004 stipulates how local authorities can request tional Disaster, local governments provide financing emergency funds from the Central Government in through their contingency budgets. case of a disaster. See Annex 13 for a detailed de- scription of the post-disaster operational phases. Budget appropriations for post-disaster recov- ery activities, financed through the Rehabilita- The financial responsibility of central and lo- tion and Reconstruction Fund, are made during cal governments is defined by Law 24/2007 and the budget discussions at the Parliament. Bud- further elaborated in GR 22/2008. Major disasters get preparation and approval affects the availability are financed with support from the central budget of additional financing resources in specific months through exceptional transfers to the provincial bud- of the year; budget preparation (before December) gets. Post-disaster financing of minor scale disasters and mid-year budget revision (June). Timing of assis- are generally assigned to the local and provincial gov- tance is also contingent on any delays in the prepa- ernments. Central Government financing of recovery ration process, while the regulations for budget re- efforts must be approved by parliament with funds visions and evaluation also affect the flexibility for being drawn from the State’s General Treasury (Ben- reallocations of funds. dahara Umum Negara) and disbursed through the Rehabilitation and Reconstruction Fund. However, the public financial contribution for the reconstruction of pri- definition of major disasters receiving assistance from vate dwellings greatly varies by disaster event. For example, the Central Government and minor disasters that do in the case of Central Java earthquake (2006), households not receive Central Government funding is unclear. of heavily damaged houses received a maximum of IDR20 million (US$2300). In the disaster of West Java and West Sumatra (2009), the affected households received IDR15 mil- According to the Law 24/2007, the government is lion (US$1800) for houses with heavy damage, IDR5 million responsible for the post-disaster financing of: i) (US$600) for medium damage, and IDR1 million (US$120) for low damage. emergency/relief operations; ii) recovery and recon- 3 In the cases of five recent disasters (West Java and West Su- struction of public infrastructure and buildings; iii) (par- matra earthquakes of 2009, Wassior flash flood, Mentawai tial) financial assistance for the reconstruction of pri- tsunami, and Mt. Merapi eruption of 2010), the Government vate dwellings2. Damage and loss assessment reports did not declare them as national disasters. However, the Government considered them as disasters where supports from the National Government is required in both response 2 Experiences from several recent disasters show that the and recovery phases. Chapter 2: Fiscal Management of Natural Disasters < 11 > Identification and transfer of funds for post-di- tary approval as the fund is drawn from the State’s saster reconstruction can take several months, General Treasury. This allocation was IDR 4 trillion depending on when during the fiscal year a di- (US$450 million) in 2011 (up from IDR 3.8 trillion in saster strikes. For example, the Aceh tsunami oc- 2010 and IDR 3 trillion in 2009). curred on December 26, 2004, at the end of the fiscal year, when the budget allocation for the next fiscal Reconstruction resources are made available year had already been approved. Major additional by the Central Government. These funds usually budget appropriations for disaster response had to come from the reallocation of capital expenditure be completed through the mid-year budget revision budgets of next fiscal years, and/or drawn from the in 2005, affecting the need for additional financing unexpended budget of the previous years. The dis- for emergency response. In contrast, the Yogyakarta bursement of these resources can take one year or earthquake occurred on May 29, 2006, just in time more. for the mid-year budget revision that started in June, resulting in faster execution of funds. The overall post-disaster response expendi- tures of GoI represented less than one per- During the recovery phase, covering approxi- cent of the total government budget in 2010. mately the first three to six months after a It included IDR 4 trillion from the Rehabilitation and disaster funds are available through the Re- Reconstruction Fund and IDR 2.5 trillion from the habilitation and Reconstruction funding assis- budget of the line ministries. It also included IDR 2.5 tance for major disasters. Every year, the GoI sets trillion for the reconstruction operations of the West aside a specific budget allocation for rehabilitation Sumatra earthquake. About IDR 38 trillion (US$4.5 and reconstruction through the Rehabilitation and billion) in post-disaster emergency and reconstruc- Reconstruction Fund to provide resources to recover tion expenditures were financed by the GoI follow- from recurrent events that occurred in the preced- ing the 2004 Aceh tsunami. See Figure 2.1. ing years. The use of these funds requires parliamen- Figure 2.1. Post-Disaster response budget of Government of Indonesia, 2004-2010 16 14 12 IDR trillion 10 10.5 10.9 4.0 10.4 8 2.5 6 1.9 2.5 4 1.6 1.4 1.2 1.0 1.3 2 3.3 2.9 2.7 3.0 3.0 4.0 0 2005 2006 2007 2008 2008 2010 n Disaster contingency fund n Line ministry disaster budget n Aceh emergency operations n Aceh recovery and reconstruction operations n West Sumatra reconstruction operations 2005 2006 2007 2008 2009 2010 Post disaster response budget (% of total government budget) 2.3% 2.1% 1.9% 1.5% 0.4% 0.8% Source: Ministry of Finance. < 12 > Indonesia: Advancing a National Disaster Risk Financing Strategy - Options for Consideration Donor assistance can represent a significant, ment can ask the provincial government for budget although uncertain, part of the financing of support, or directly request support from the central natural disasters and is likely to decrease in government. If additional financing is still needed, the future. For example, most of the reconstruc- the provincial government asks the central govern- tion financing after the tsunami in Aceh in 2004, ment for budget support. The post-disaster evalua- which caused losses estimated at US$4.5 billion, tion committee, which assesses the financing needs, was financed through donor assistance. In contrast, includes representatives of government entities and donor assistance represented less than 30 percent is chaired by BNPB. Once approved by the evalua- of the financial resources required in the case of tion committee, a financial request is submitted by the Yogyakarta earthquake, which caused losses of BNPB to the Ministry of Finance. The Ministry of Fi- US$3.1 billion. Given the economic growth of Indo- nance seeks approval for the financing plan from nesia (and Indonesia being now part of the G20), Parliament during the semi-annual budget revision. it is likely that the donor assistance will decrease in Budget support from the Rehabilitation and Recon- the future, as Indonesia has larger domestic financial struction Fund is then transferred to the local gov- capacity to cope with the economic losses caused by ernments. The recovery activities are executed by the natural disasters. local government and the line ministries under the coordination of BNPB and BPBD. From Municipal to Central post-disaster recov- ery financing. In the aftermath of a disaster, the The remaining part of this chapter further describes district government establishes a recovery plan. This the role and responsibility of the local and central plan is expected to be financed out of its own bud- public entities for each of the three post-disaster get (on contingency line and/or post-disaster budget phases. Figure 2.2 below summarizes the main reallocation). In case the recovery budget exceeds sources of post-disaster financing. Additional infor- 20 percent of the total budget, the district govern- mation can also be found in Annex 12. Figure 2.2. Financing of post-disaster operations in Indonesia Disaster Emergency Recovery Reconstruction Phase response/Relief On call fund/ Rehabiliation and Capital expendi- Budgetary contingency reconstruction tures of next fis- vehicle budget fun cal year budget Financing Central/local Central/local Central/local sources budget budget budget Source: Author. Note: See Annex 12 for further details. Chapter 2: Fiscal Management of Natural Disasters < 13 > Emergency Response/Relief Phase Recovery Phase During the first few weeks after a disaster, the The recovery phase (also called rehabilitation emergency response phase includes activities such phase) starts after the emergency response as rescue and evacuation, and the provision of ba- phase and lasts 3 to 6 months. During this post- sic supplies (e.g., food, water), health services, and disaster phase, lifeline infrastructure (water, electrici- emergency shelters. ty, sanitation, etc) and key public buildings (hospitals, bridges, etc) are repaired. Housing rehabilitation as- Funds for emergency response activities are sistance is also provided to the affected households. immediately available from a variety of sourc- es, depending on the size of the disaster. The The type of financial assistance provided dur- local governments use their own financial resources ing the recovery phase depends on the severity for emergency response through their contingency of the event. According to the government regula- budget line. If these funds are not sufficient, funds tion GR 21/2007, if the disaster damage is less than may be provided by the provincial government. In 20 percent of the municipal budget, recovery is fi- the event of a disaster, BNPB disburses funds from nanced at the local level. If the disaster damage ex- its “on call” fund. ceeds 20 percent of municipal budget, assistance is provided by the provincial government. If additional Post-disaster emergency response operations funds are required, they are provided by the Central are mostly executed by local governments Government through the Rehabilitation and Recon- and financed through their own contingency struction Fund. It should be noted that in the rule is budget at a relatively fast speed. Local govern- not always followed in practice. ments do not have specific “on call” funds to be triggered after a disaster, but instead draw monies The Rehabilitation and Reconstruction Fund from their contingency budget line. These funds may offer budget resources for the financing are quick-disbursing budget items that can be im- of natural disasters. The current Rehabilitation mediately executed after a disaster. Provincial gov- and Reconstruction Fund (Dana Rehabilitasi dan ernments also allocate a portion of their budget Rekonstruksi4), which is the continuation of funding for unforeseen events such as natural disasters and assistance practice for post-disaster previously man- other non-planned activities through a contingency aged by the Coordinating Ministry for Social Wel- budget line. Due to the contingent nature of this fare, is voted by the Parliament in the annual bud- budget item, funds can only be used for emergency get. It is not a multi-year reserve fund, but instead, response and cannot be allocated for activities that a line item that is renewed every year on the basis can be planned in advance, such as reconstruction. of proposal from the executive branch (i.e., today from BNPB). The GR 22/2008 stipulates a Disaster The BNPB ensures coordination of emergency Contingency fund. However, Article 6 (2) of the GR response activities through its “on call” fund. 22/2008 limits the use of the disaster contingency In the event of a National Disaster, “on call” funds fund only for preparedness activities at the pre- are transferred to the local governments through 4 It should be noted that the term ‘Dana Rehabilitasi dan the local disaster agencies BPBD. For larger disasters Rekonstruksi’ is not recognized in the GR 22/2008. Instead, Ministry of Social Affairs is involved in emergency article 6 (5) stipulates that: “Social assistance fund through response operations using their own “on call” fund grant is provided in the State Budget (APBN) for activities at within their budget. In addition, some portion of the the post-disaster stage”. The current practice of requesting to parliament to draw from contingent reserve at the State’s Rehabilitation and Reconstruction Fund can be used General Treasury leads to the approved fund being called as “on call” funds. ‘Rehabilitation and Reconstruction (RR) Fund’. < 14 > Indonesia: Advancing a National Disaster Risk Financing Strategy - Options for Consideration disaster stage, possibly mixing the term contingency disbursement must be approved by Parliament. This with that of contingency plan, which indeed is a generates delays in post-disaster recovery and re- pre-disaster measure. This restriction contradicts the construction operations. See Annex 2. concept of ‘contingent’ being conditional of disaster events that may occur. The size of the Rehabilitation and Reconstruc- tion Fund can be revised based on past events In practice disaster contingency funds are not and has remained relatively constant in recent specifically allocated, but instead are drawn years, except in 2011 where it increased signifi- from a contingency fund in the State’s General cantly. Figure 2.3 shows the annual allocation to Treasury. The current practice in providing funds the Rehabilitation and Reconstruction Fund within for post-disaster rehabilitation and reconstruction the initial budget and the revised budget, as well has been to propose special budget allocation from as the actual disbursements, as a percentage of the the line item “others” (budget code 999) based on total government expenditures. The major increase proposal submitted to the Parliaments to draw from in 2005 in the revised budget is the consequence of the broad contingency reserve in the State’s General the Tsunami in Aceh in December 2004. Likewise, Treasury (BUN). Proceeds are mainly used for post- the 2006 increase in the revised budget was voted disaster recovery expenditures and are predomi- by the Parliament to finance the emergency and re- nantly targeted to the affected households (e.g., as covery expenditures in the aftermath of the earth- social compensations). In 2011, IDR 4 trillion were quake in Yogyakarta of May 2006. In 2007-2009, allocated to this fund, or 0.4 percent of government the initial budget allocation was enough to cover expenditures. This allocation, referred to as Rehabili- the post-disaster emergency and recovery activities, tation and Reconstruction Fund, represents a 5 per- and no budget revision was needed. The Rehabili- cent increase from IDR 3.8 trillion in 2010 and a 33 tation and Reconstruction Fund received an annual percent increase from IDR 3 trillion in 2010. budget allocation of 0.3 and 0.4 percent of govern- ment expenditures in 2010 and 2011, respectively. To engage funds from the Rehabilitation and Reconstruction Fund, BNBP sends a proposal to The Ministry of Social Affairs disburses liveli- the Ministry of Finance, which submits it to the hood grants and housing compensation to the Parliament for approval. The parliamentary bud- households affected by a disaster, out of its get process occurs annually and is revised halfway own budget. The amount varies by disaster. For through each year. Recovery expenditures can also example, the grant distributed to disaster affected be financed during the fiscal year of the disaster households in Wasior Papua in October 2010 was through a budget reallocation of capital expendi- IDR 5,000 per person, per day, during the recovery tures of the line ministries, but the funds available phase (normally three to six months). The Ministry are limited and are not yet well coordinated. of Social Affairs receives an annual allocation for this purpose and if the funds are insufficient, the Minis- The current Budget Law does not allow for the try can request an additional allocation in the mid- fast disbursement of funds in case of disaster year budget revision. and can generate a liquidity crunch. Contrary to emergency spending from the “on call” funds, The distribution of these grants can take some the execution of funds from the Rehabilitation and time. In the case of Aceh Tsunami in late December Reconstruction Fund to finance post-disaster recov- 2004, for example, the grants were received in May ery operations has to follow the state budget cycle. and June 2005, that is, five months after the oc- Recent experience shows that it can take several currence of the disaster. In the case of Yogyakarta months to draw down funds from the Contingency earthquake in May 2006, the grants were distrib- Fund in BUN because any specific allocation and uted between June and August 2006. See Box 2.1. Chapter 2: Fiscal Management of Natural Disasters < 15 > Figure 2.3. Rehabilitation and Reconstruction Fund, as percentage of government expenditures, 2005-2011 0.70% 0.60% n Initial allocation n Revised allocation n Actual disbursements 0.50% US$ million 0.40% 0.30% 0.20% 0.00% 2005 2006 2007 2008 2009 2010* 2011* (*) Revised allocation and actual disbursements not available. Source: Ministry of Finance, 2010. Box 2.1. Post-Disaster allocation of resources The 2006 Yogyakarta earthquake is an example of funds meeting recovery needs in a timely fashion. As the di- saster occurred in May, the government was able to revise its budget in the mid-year budget revision of June in order to allow for the financing of post-disaster emergency and recovery operations. By October 2006, US$270 million of the assessed needs for housing reconstruction were available for disbursement and the funds were distributed between October and December. The remaining US$270 million was budgeted in the next fiscal year and was disbursed in 2007. The 2009 West Java earthquake highlights the possible inefficiency resulting from the current budget approval process. Approximately one-third of the post-disaster recovery needs (about IDR 500 billion) was available in December 2009 from the Rehabilitation and Reconstruction Fund. However, the affected households waited until October 2010 (one year after the disaster) to receive the full amount because the second tranche was budgeted under the Rehabilitation and Reconstruction Fund of the next fiscal year. Reconstruction Phase of the total government expenditures. The tsunami in Aceh in December 2004 had a major impact on The reconstruction of public assets is mainly the post-disaster expenditures in the following fiscal financed from budget allocations of capital years. Removing Aceh, and considering that recov- expenditures of future fiscal years. Figure 2.4 ery funds are equivalent to 0.3 percent of govern- shows the total annual post-disaster expenditures in ment expenditures in 2008 and 2009, reconstruc- the government budget in 2005-2009, as a percent tion financing has been limited in recent years. < 16 > Indonesia: Advancing a National Disaster Risk Financing Strategy - Options for Consideration Figure 2.4. Post-disaster public recovery and reconstruction expenditures, as percentage of Central Government expenditures 2.50% n Total with Aceh event n Total without Aceh event 2.00% 1.50% 1.00% 0.50% 0.00% 2005 2006 2007 2008 2009 Source: Ministry of Finance, 2010. Line ministries are responsible for the recon- implementation of regional autonomy and fiscal de- struction of their assets. The Ministry of Finance centralization that began in 2001, regional govern- allocates budget to the capital expenditures budget ments receive transfers of funds from the Central line of each line ministry to allow them for the re- Government to finance specific projects and activi- construction of their assets damaged or destroyed ties. These funds constitute an essential component by the natural disaster. The line ministries then of the local budget revenues. For example, in 2009 transfer those funds to their regional offices. This total funds transferred from the Central Govern- reconstruction phase can take several years. Since ment to the regions represented 42 percent of the the establishment of BNPB, line ministries have as- total state budget expenditures.6 Therefore, in case sumed in several occasions that BNPB is responsible of a disaster, the local governments are expected to to secure funding for all reconstruction needs5. This finance the reconstruction of their assets out their had further delayed the budgeting for the line min- annual budget. In case of a major disaster, the Cen- istries’ assets as they tend to wait for possible top tral government can exceptionally allocate addition- up funding. al funds for rehabilitation and reconstruction7. The local governments are responsible for the reconstruction of their own assets. With the 6 These funds can reach 60 percent of state expenses if other national programs for decentralization are included. See In- donesia Finance and Budget Memorandum, 2010. 5 The case of 2009 West Sumatra earthquake was a concrete 7 There have been mixed cases where the line ministries allo- examples of situation where some line ministries initially cated sub-national grants to repair sub-national assets (e.g., expected that BNPB will secure funding for them to repair DAK for Education in 2010 allowed the fund to be used to assets under their responsibilities. It was finally recognized repair schools damaged by past disasters), and where BNPB that Law 24/2007 does not remove the responsibilities of line provided subsidy grants to local governments to repair dam- ministries to repair damaged assets under their jurisdiction, ages local assets (e.g., for repair and/or upgrade of local whether the assets were damaged by disaster or otherwise. roads damaged by floods or earthquakes). Chapter 3 Fiscal Disaster Risk Assessment < 17 > < 18 > Indonesia: Advancing a National Disaster Risk Financing Strategy - Options for Consideration The assessment of the financial and fiscal risk related disaster. In the aftermath of a disaster, immediate to natural disasters is the first step in devising di- resources must be mobilized to fund post-disaster saster risk financing strategies. Such an assessment emergency and recovery activities. Once the recov- ideally requires both historic loss data and modeled ery phase is completed, the GoI must mobilize lon- losses from catastrophe risk models. This presents a ger term resources to meet reconstruction needs. In preliminary fiscal risk profile of Government of In- general, there are three broad categories of post- donesia. disaster spending needs for which the Government assumed contingent liability, namely: 1) repair of na- Indonesia has conducted post-disaster damage tionally owned public assets such as national roads, assessments to measure the physical and finan- major water infrastructure, and national govern- cial losses caused by adverse natural events for ment buildings (typically in medium-term)); 2) repair several years. The damage assessment procedure of sub-nationally owned public assets such as pro- was upgraded in 2006 with the introduction of the vincial and district roads, schools, or local markets PDNA system, based on the Economic Commission (typically in short-to-medium term); and 3) stimulus for Latin America and the Caribbean (ECLAC) meth- grants for livelihood recovery and housing recon- odology. This damage assessment system is intend- struction (typically in short term). ed to record direct physical damage to public and private property in order to facilitate post-disaster A major challenge for governments in the af- recovery and reconstruction financing decisions by termath of a disaster is to access immediate government. liquidity to finance its short-term spending needs. While there are various financial instruments that can be mobilized for the post-disaster recon- Contingent liability and post-disaster struction phase, including additional credit and tax spending needs increases, financial instruments that ensure access to The contingent liability of the government due immediate liquidity after a disaster are more chal- to natural disasters creates major fiscal risk. lenging to access. See Annex 5 which describes the However, the government’s contingent liability is not potential financial instruments available. clearly defined in the laws and thus makes the fiscal risk difficult to assess. Beyond its explicit contingent Assessing the short-term post-disaster spend- liability and its associated spending needs, such as ing needs is essential. To devise a cost-effective the reconstruction of public assets and infrastruc- disaster risk financing strategy, especially for the ture, the Government may have a moral and social funding of short-term post-disaster public spend- responsibility (implicit contingent liability) to assist ing needs, it is critical to assess those possible public the population in case of a disaster. For example, spending needs that create additional fiscal risk for the Government provides not only emergency assis- the government. tance (such as food, shelters, and medication) but also can finance recovery/reconstruction activities such as stimulus grants for rebuilding of low-income Fiscal disaster risk modeling housing. The fiscal disaster risk profile of Indonesia, which reflects the government’s contingent lia- The contingent liability of the GoI related to bility of natural disasters, should build on both natural disasters can be categorized in short historic disaster losses and simulated disaster term and medium term post-disaster spend- losses. Historic disaster loss data, as reported by ing needs. All financial resources do not need to the provincial and the central governments, informs be mobilized immediately after the occurrence of a about the recurrent losses caused by small but fre- Chapter 3: Fiscal Disaster Risk Assessment < 19 > quent disasters, such as localized floods, small earth- The reinsurance company PT Maipark has de- quakes, etc. Simulated catastrophe losses are com- veloped an earthquake risk model for the in- puted from catastrophe risk models for a specific surance industry in Indonesia, which still needs peril, such as earthquake, and inform about possible some further technical improvements. While catastrophe losses caused by a major disaster occur- the hazard module relies on the state-of-the-art ring once every 20 years or less frequently. seismic technology and a unique catalogue of his- torical events, some further development would im- Probabilistic catastrophe risk models offer the prove the model, such as the development of a set government innovative tools to assess the fi- of country-specific vulnerability classes. This model nancial exposure to natural disasters. Govern- is currently being used to develop scenario-based ments in both developed and developing countries earthquake analysis to guide the Ministry of Finance are increasingly using catastrophe risk modeling in its fiscal risk management. See Box 3.2. techniques to guide their disaster risk management and financing decisions. Such tools allow for the probabilistic assessment of low-frequency, high se- verity disasters, such as a major earthquake or tropi- cal cyclone occurring once every 20 years or less fre- quently. See Box 3.1. Box 3.1. Probabilistic catastrophe risk modeling Financial disaster risk assessment for governments can be developed using inputs from probabilistic catastrophe risk models. This technique was originally developed by the insurance industry to assess the risk on a portfolio of assets and is increasingly used by governments to assess their exposure to adverse natural events. A typical risk model is made of the following modules: Hazard module: This module defines the frequency and severity of potential perils (e.g. earthquake, tropical cyclone) at specific locations within the region of interest. This is done by analyzing historical frequencies and reviewing scientific studies performed on the severity and frequencies in the region of interest. This module then generates thousands of stochastic events based on historical data and experts’ opinions. Exposure module: This is a geo-referenced database of assets at risk, assigning a list of attributes (e.g., exact location, construction type, number of stories) for each asset. This information is used to determine the area’s vulnerability, captured though vulnerability functions. At a larger scale, for example when analyzing an entire country, proxies are used to define the vulnerability of entire neighborhoods or even cities. Loss module: This module combines the hazard module and the exposure module to calculate different risk metrics, such as the annual expected loss (AEL), which is an expression of the long-term (for example, 1,000 years) average annual loss, and the probable maximum loss for a given return period, which represents the expected loss severity based on likely occurrence, such as the 1-in-50-year loss or the 1-in-100-year loss. Risk matrices generated by probabilistic risk models can be used to complement historical analysis and are particularly useful to policy makers in assessing the probability of losses and the maximum loss that could be generated by major events (e.g. an earthquake affecting a major city or a cyclone affecting a major port). < 20 > Indonesia: Advancing a National Disaster Risk Financing Strategy - Options for Consideration Box 3.2. Scenario-based earthquake risk analysis for Ministry of Finance The fiscal department of Ministry of Finance (BKF) set up a working group to develop an earthquake risk model to be used for the fiscal strategy against natural disasters. This working group includes the Bandung Institute of Technology, the private reinsurance company PT Maipark, the Ministry of Public Works, BKMG, LIPI, and the Geological Agency. The model builds on the earthquake hazard model developed by PT Maipark. An exposure database (includ- ing both public and private assets) will be developed with assistance from the World Bank and the Australia- Indonesia Facility for Disaster Reduction. A scenario-based analysis will be conducted to assess the economic and fiscal impact of major earthquakes, including probable maximum losses of selected earthquake events. Three areas have been selected for in depth scenario analysis including: i) area of Sunda Straits; ii) coastal area of Central and East Java; and iii) Nusa Tenggara Barat near to Bali island. These areas were selected based on their high vulnerability and economic importance. A disaster risk modeling and management tool, Preliminary fiscal disaster risk profile called Risk-in-a-Box is under development. This of Government of Indonesia initiative, undertaken by BNPB with support from The public spending needs related to post- the World Bank, the Global Facility for Disaster Re- disaster recovery and reconstruction operations duction and Recovery (GDFRR) and the Australia-In- are difficult to trace. As a first attempt to assess donesia Disaster Risk Facility (AIDRF), aims to guide this fiscal disaster risk in Indonesia, historic budget disaster risk management decisions by providing de- expenditures related to natural disasters are used to tailed risk assessment. The purpose of Risk-in-a-Box estimate the impact of natural disaster on the govern- is to develop a tool, to be run on laptops, that will ment budget. Unfortunately, such data is not easily model impacts of different hazard events on popula- available because most of the post-disaster expendi- tion or infrastructure according to given guidelines. tures are not identified as such in the budget. This is The objective of the tool is to support the overall particularly true for the recovery expenditures to be process of sub-national risk assessments in Indone- financed during the year of the disaster. Reconstruc- sia and the software developed is likely to be useful tion activities, which usually start several months after more broadly as a general impact modeling tool. a disaster, are typically planned in the budget of fu- ture fiscal years and thus may be easier to trace. An analysis of historical losses should comple- ment the earthquake risk modeling approach. The fiscal disaster risk related to the pub- The Indonesian earthquake model, once completed, lic spending needs for post-disaster recovery can assist the Ministry of Finance in identifying the operations is estimated using the number of fiscal impact of major disasters. However, such an buildings affected by disasters, as reported by approach: (i) is limited to earthquake risks; and, (ii) BNPB. Post-disaster recovery expenditures financed does not efficiently capture the more recurrent loss- by the government in the first months after a disas- es usually caused by localized floods or landslide. Re- ter are estimated using an indirect approach based current localized losses must be taken into account on the number of buildings damaged or destroyed since the accumulation of such events can create as reported by BNPB. Figure 3.1 shows the num- significant fiscal losses for the government. Chapter 3: Fiscal Disaster Risk Assessment < 21 > Figure 3.1. Buildings reported as destroyed or damaged by natural disasters in Indonesia, 2004-2009 n Other perils n Earthquake/Tsunami n Flood/Landslide 600,000 Number of buildings destroyed or 500,000 damaged, by peril 400,000 300,000 200,000 100,000 – 2004 2005 2006 2007 2007 2009 Source: BNPB. ber of buildings reported by BNPB as destroyed or is for the reconstruction of critical public assets.9 See damaged by natural disasters over the period 2004- Figure 3.2.10 2009.8 It should be noted that the number of build- ings destroyed was significantly higher in 2009 com- Actuarial techniques have been used to provide pared to the previous years (including the 2004 year preliminary estimates of future possible public where the Tsunami in Aceh occurred). spending needs for post-disaster recovery op- erations. Public spending data of past events, as The fiscal cost of a building reported as de- estimated from the number of buildings destroyed stroyed or damaged is estimated at US$1500. and damaged, have been used to fit a parametric Based on recent disasters in Indonesia, it is estimat- distribution and simulate possible future spending ed that the GoI allocates on average US$1,500 for needs (or fiscal losses) related to natural disasters. every house destroyed or damaged by a disaster. A In particular, the risk metrics such as the annual ex- portion of this cost is the direct financial compensa- pected loss (AEL) and the probable maximum loss tion for the affected households and the remaining (PML) have been estimated. The AEL is an estimate of the long-term annual average loss, after account- ing for historic trends in the historic data. The PML is defined as an estimate of the maximum loss that Experiences from several recent disasters show that the public fi- 9 nancial contribution for the recovery of private dwellings greatly varies by disaster event. For example, in the case of Central Java Longer time series is available but the analysis of the data 8 earthquake, households of heavily damaged houses received a seems to show a structural break in BNPB data on build- maximum of IDR 20 million. In the recent disaster of West Java ings destroyed or damaged by natural disasters following and West Sumatra, the affected households received IDR 15 mil- the 2004 Tsunami. This structural break, possibly caused by lion for houses with heavy damage, 5 million for medium dam- a change in reporting practices, means that BNPB data on age, and IDR 1 million for low damage. buildings damaged or destroyed between 2004 and 2009 is 10 The actuarial model developed to assess the fiscal risk related likely to be more relevant for estimation of future experience to natural disasters allows for any value of fiscal cost of a than data before 2004. destroyed or damaged building. < 22 > Indonesia: Advancing a National Disaster Risk Financing Strategy - Options for Consideration Figure 3.2. Indicative historic fiscal risk related to natural disasters, 2004-2009 n Other perils n Earthquake n Flood 1,000 900 800 700 US$ million 600 500 400 300 200 100 – 2004 2005 2006 2007 2008 2009 Source: Authors, from BNPB. Note: Fiscal losses estimated at US$1,500.00 per building damaged/destroyed is likely to arise on the occurrence of an event or se- indicative fiscal loss excedance curve, the indica- ries of events considered to be within the realms of tive AEL and selected PML. In an average year, the probability, ignoring remote coincidences and pos- fiscal losses are estimated in the range of US$420- sible but unlikely catastrophes. For example, a PML 550. Every 10 years they could exceed US$800-950 with a 100-year return period is the estimated loss million; and every 100 years, losses could be in caused by an event occurring once every 100 years excess of US$1.5-1.6 billion. It should be noted on average (or with a 1 percent chance per year on that the AEL is mainly driven by high frequency, low average). Two actuarial methods have been tested. severity events such as floods and landslides, while First, historic fiscal loss data over the period 2004- the PMLs with return periods of 50 years and more 2009 are used to fit a parametric distribution (Ac- are mainly driven by low frequency, high severity tuarial Method 1). Second, historic fiscal loss data, events such as earthquakes and tsunamis. adjusted over the period 2000-2009 using a linear trend, are used to fit a parametric distribution (Actu- Although this fiscal disaster risk analysis arial Method 2).11 12 should be seen as preliminary, it provides the GoI with order of magnitude of their pos- The preliminary fiscal disaster risk analysis sug- sible public spending needs for post-disaster gests that the annual fiscal disaster losses are recovery operations. As discussed before, the in the range of US$420-500 million and that above actuarial analysis, which uses estimated his- once every 100 years these losses are close to toric fiscal loss data, should be complemented by US$1.5-1.6 billion. Figure 3.3 below shows the catastrophe risk modeling techniques, particularly A simple linear detrending method (statistically significant) 11 for the assessment of future possible losses caused is applied to the historic fiscal loss data. This upward linear by major disasters, like earthquake. However, the trend captures increase in asset exposure, better damage re- analysis above does provide a preliminary fiscal risk porting system, etc. profile of natural disasters for the GoI, to be used Although this report presents figures assuming that fiscal 12 losses follow a Log-Normal distribution, the results are broad- to guide the GoI in the development of a nation- ly robust to alternative distributional assumptions. al disaster risk financing strategy (see Chapter 5). Chapter 3: Fiscal Disaster Risk Assessment < 23 > Figure 3.3. Fiscal Disaster Risk Profile for Government of Indonesia – Indicative Exceedance Probability Curve Estimation method 1 Estimation method 2 2,000 1,800 1,600 1,400 US$ million 1,200 1,000 800 600 400 200 0 0 50 100 150 200 250 Source: Authors, from BNPB. Note: Fiscal losses estimated at US$1,500.00 per building damaged/destroyed Actuarial Method 1 Actuarial Method 2 Indicative risk metrics US$ million US$ million Annual Expected Loss 423 554 Probable Maximum Loss: 10 year return period 796 945 50 year return period 1,320 1,299 100 year return period 1,570 1,448 150 year return period 1,725 1,550 250 year return period 1,947 1,647 Chapter 4 Review of Private Catastrophe Risk Insurance Market in Indonesia < 25 > < 26 > Indonesia: Advancing a National Disaster Risk Financing Strategy - Options for Consideration Indonesia’s non-life insurance market is un- insurance pricing; iii) develop a hazard and exposure der-developed. The non-life insurance market in database for earthquakes; and, iv) build local capac- Indonesia is less developed than in its neighboring ity for earthquake resilience. A full description of countries. The non-life insurance penetration rate, Maipark can be found in Annex 3. measured as a percentage of GDP, is estimated at 0.6 percent in Indonesia compared to 1.6 percent in Earthquake insurance penetration is very low Malaysia and 1.1 percent in Thailand. with less than 5 percent of assets insured (main- ly commercial and industrial assets). The volume The current insurance regulatory framework does of Maipark’s coverage for non-life catastrophic in- not monitor or control catastrophe risk. There is no surance was estimated at US$2.6 billion in 2008. specific regulatory treatment of catastrophe risks, However, market penetration is increasingly quickly. such as specific capital or reinsurance protection re- Premiums written in 2009 totaled US$8.8 million quirements or catastrophe reserves. compared to US$6.5 million in 2008, a growth rate of 27 percent. The maximum commission paid on earthquake cessions is 8.5 percent. Private property catastrophe insurance Recent earthquake events show that insured PT Maipark is the only specialized earthquake losses represent less than 10 percent of the insurance company in Indonesia. The General damage.13 Out of the estimated US$4.5 billion Insurance Association of Indonesia, supported by damage caused by the Tsunami in 2004, less than the GoI and the Bureau of Insurance, established 1.5 percent were insured. About 6 percent of the the specialized catastrophe insurance company PT damage caused by the earthquake in Bengkulu in Maipark in 2004. The mandate of this insurer, owned 2007 was insured. See Figure 4.1. by the domestic non-life insurance companies, is to: Estimates of economic losses vary according to sources and 13 i) promote discipline and proper handling of earth- methodologies, estimates as high as 12% have been released quake insurance; ii) set a benchmark for earthquake by entities in the private sector. Figure 4.1. Insured losses as a percentage to estimated total damage 7% 6% 5% 4% 3% 2% 1% 0% Nabire Aceh Padang Yogyakarta Pangandaran West Sumatera/ Bengkulu Earthquake Earthquake – Earthquake Earthquake Earthquake Padang Earthquake Tsunami Earthquake Source: Ministry of Finance. Chapter 4: Review of Private Catastrophe Risk Insurance Market in Indonesia < 27 > Earthquake insurance premium rates vary by The West Sumatra Province government has in- zone and type of buildings. Earthquake premium sured its local assets against earthquake risks rates, developed by Maipark and approved by the since 2008. This insurance coverage protects 42 lo- Insurance Association, vary by zone (5 zones), use cal government buildings, four hospitals, 73 local (commercial and industrial, residential) and type of government official and guest houses. The insurance construction. They vary from 0.85 per mill to 4.7 per premium is paid from a specific budget item “ex- mill of the total sum insured. See Annex 3 for addi- penditure for insurance of local government assets”. tional coverage details. Total insurance premiums were around IDR 200 mil- lion in 2010. The first claim experience was after the Other forms of catastrophe risk insurance have earthquake of September 2009 and was settled in been piloted, such as a microinsurance against May 2010 (i.e., seven months after the earthquake) floods. Developed by Whana Tata in 2009, this for an amount of IDR 20 billion. product was established in partnership with the German development agency GTZ and the reinsurer The municipality of Yogyakarta has insured its Munich Re. This product is not a property insurance public assets since 2003, including government product, as the payout is not based on actual prop- buildings, schools, hospitals, traditional mar- erty losses. Rather it serves as livelihood coverage ket places, and motor vehicles. After the 2006 against floods, providing immediate cash to the in- earthquake, the municipality received a payout of sured households located in flooded areas. This bi- IDR3.4 billion, which represents 14 times the annual nary insurance product is sold for IDR 50,000 and premium that year In addition to protecting public pays IDR 250,000 when flooding in Jakarta reaches assets, the municipality plans to set up compulsory a certain predefined level. Coverage cards (or sev- property catastrophe insurance for private dwellings eral) can be purchased by any person that can prove with insurance premiums bundled with the land and residence in the covered area, which was available property tax (PBB) payment. in 23 sub-districts of Jakarta, and about 500 cards were sold through local insurance agents. The pro- gram was not renewed in 2010. Catastrophe risk insurance of public assets Public assets are usually not insured for cata- strophic events, although some provinces/mu- nicipalities have begun insuring their critical assets. Assets owned by the Central Government are not insured, and similar to many developed countries, the Central Government is, de facto, its own insurer. However there have been recent initia- tives in some provinces to insure selected public as- sets against natural disasters. Chapter 5 Options for National Disaster Risk Financing Strategy in Indonesia < 29 > < 30 > Indonesia: Advancing a National Disaster Risk Financing Strategy - Options for Consideration A national disaster risk financing strategy A national geo-referenced exposure database should be designed to improve the capacity of should be built. This database would include attri- the GoI to access immediate financial resources butes of public and private buildings and infrastruc- in case of natural disaster while maintaining its ture exposed to natural disasters, such as schools, fiscal balance. Building on the country disaster risk hospitals, public buildings, roads, bridges. It could financing framework promoted by the World Bank, also include information on private assets such as six options for a comprehensive disaster risk financ- houses. This database, combined with the catastro- ing in Indonesia are discussed below. Table 5.1 pres- phe risk model, would allow, among other applica- ents a summary of the proposed options. tions, for the assessment of the financial and fiscal impact of natural disasters. This information would Table 5.1. Options for a national disaster risk also be critical for the insurance industry to offer financing strategy in Indonesia sustainable and affordable property catastrophe in- surance products. Timeframe Options for disaster risk financing Short Term Develop financial disaster risk assessment A financial catastrophe risk model should be tools developed for Ministry of Finance. The earth- Short Term Develop a national disaster risk financing quake risk model could be the basis of a financial strategy relying and risk retention and risk transfer catastrophe risk model to be used by the Ministry Short Term Establish a National Disaster Reserve of Finance. This model would include an actuarial/ Fund as fast-disbursement mechanism financial model that would build on the modeled for the financing of post-disaster losses of the catastrophe risk model and the historic operations losses. This tool would assist the Ministry of Finance Medium Term Establish a disaster risk insurance in the design of the national disaster risk financing program for public assets strategy, including the size of the annual budget al- Medium Term Promote property catastrophe risk insurance of private dwellings location to the Rehabilitation and Reconstruction Fund and any disaster risk transfer strategy (such as Longer Term Establish a Joint Disaster Reserve Fund for Indonesia’s Local Governments insurance). Such a financial model is currently being used by the Ministry of Finance in Mexico. See Box 5.1. Develop financial disaster risk assessment tools Develop a national disaster risk The design of a national disaster risk financing financing strategy strategy starts with a detailed disaster risk as- The national disaster risk financing strategy sessment. Catastrophe risk modeling techniques should rely on a risk layering approach. This can complement the actuarial analysis of historic approach offers an optimal mix of risk retention loss data to assess the financial and fiscal exposure (through reserves/contingency budget and contin- to natural disasters. gent credit) and risk transfer such as insurance. See Annex 5 for further details and a comparative analy- Hazard modules for major perils should be de- sis of risk financing and risk transfer products. veloped. PT Maipark has developed an earthquake hazard module, based on a unique catalogue of Disaster risk layers could be financed through historic earthquakes. A flood hazard module could an optimal combination of financial instru- also be developed for major urban areas like Greater ments. Figure 5.1 depicts the three tiered financial Jakarta. strategy described below. Chapter 5: Options for National Disaster Risk Financing Strategy in Indonesia < 31 > Box 5.1. R-FONDEN – The financial catastrophe risk model of Ministry of Finance in Mexico The Government of Mexico developed, for its national disaster fund FONDEN, a catastrophe risk model called R- FONDEN. This probabilistic risk model offers catastrophe risk analysis for four major perils (earthquake, floods, tropical cyclones, and storm surge), for infrastructure in key sectors (education, health, roads, and low-income housing) at the national level, state level and sub-state level. The analysis can be performed on a scenario-basis or on a probabilistic basis. R-FONDEN takes as input a detailed exposure database (including details of buildings, roads and other public assets, and produces) as outputs risk metrics including AEL and PML. This model is currently used by the Ministry of Finance, in combination with the actuarial analysis of historic loss data, to monitor the disaster risk exposure of the portfolio of FONDEN and to design disaster risk transfer strategies, such as the placement of indemnity-based reinsurance and the issuance of catastrophe bonds. Figure 5.1. Bottom up approach to three-Tier financial strategy against natural disasters Disaster Risks Disaster Risk Financing Instruments Low Major High risk layer Catastrophe risk transfer (e.g., major earthquake, major (e.g., parametric insurance, tropical cyclone) cat bonds) Frequency of Event Severity of Impact Medium risk layer Contingent credit (e.g., floods, small earthquake) Low risk layer Contingent budget, reserves, High Minor (e.g., localized floods, landslides) annual budget allocation Source: Authors. ■■ Low Risk Layer (with return period up to 4 Development Policy Loan with Catastrophe De- years): The annual budget allocation/contingen- ferred Drawdown Option, (DPL with Cat DDO). cy budget could finance recurrent disaster losses. See Box 5.2. An annual budget allocation would finance re- current disasters such as localized floods, land- ■■ High Risk Layer (with return period higher slides, or minor earthquakes. than 20 years): Low frequency, high sever- ity risks can be transferred to the international ■■ Medium Risk Layer (with return period be- capital/reinsurance markets through catastrophe tween 4 years and 20 years): Contingent credit reinsurance, cat bonds and/or cat derivatives. Di- would finance more severe, but less frequent, di- saster risk transfer instruments, such as disaster sasters. This budget instrument would allow The insurance, would finance major disasters. The GoI to draw down funds quickly after a natural GoI could purchase parametric insurance against disaster. The GoI may consider the World Bank major disasters like earthquakes or tropical cy- < 32 > Indonesia: Advancing a National Disaster Risk Financing Strategy - Options for Consideration Box 5.2. World Bank Catastrophe Deferred Drawdown Option The Development Policy Loan (DPL) with catastrophe draw down options (Cat DDO) offers a source of im- mediate liquidity that can serve as bridge financing while other sources (e.g. concessional funding, bilateral aid or reconstruction loans) are being mobilized after a natural disaster. Borrowers have access to financing in amounts up to US$500 million or 0.25 percent of GDP (whichever is less). The Cat DDO has a “soft” trigger, as opposed to “parametric” trigger; funds can be drawn down upon the occurrence of a natural disaster resulting in the declaration of a state of emergency. See Annex 6 for additional details. clones. Payouts would be disbursed based on Additional financial capacity could be secured parametric triggers, such as the magnitude of an through parametric insurance. The GoI could earthquake or the intensity of a tropical cyclone. complement its reserves and/or contingent credit This type of insurance is transparent and allows with parametric insurance. Parametric insurance for fast claims settlement (usually within 2 to 4 products are insurance contracts that make pay- weeks). ments based on the intensity of an event (for exam- ple, wind speed, earthquake intensity) rather than A “bottom-up” disaster risk financing approach the actual loss. Unlike traditional insurance settle- should be considered. The GoI should first secure ments, which require an assessment of individual financing for recurrent events (bottom risk layer) losses on the ground, parametric insurance relies on through risk retention (reserves and/or contingent an assessment of losses using a predefined formula credit) and then move up by increasing its level of based on variables that are exogenous to both the financial resilience through disaster risk transfer in- individual policyholder and the insurer, but which struments. have a strong correlation to individual losses. Para- metric instruments allow for fast claims settlement Although the national budget does not explicitly (usually within 2 to 4 weeks) and are less exposed to prohibit the purchase of insurance, there is cur- moral hazard and adverse selection. However, para- rently no specific budget item allowing for the metric products are exposed to basis risk, i.e., the payment of insurance premiums. Under the cur- possibility that claims payments may not perfectly rent budget law, BNPB cannot use its annual budget match individual losses. Careful design of index in- allocation to purchase insurance. It should be allowed surance parameters is important to help reduce ba- under the budget law to use part of its resources to sis risk. Key features for parametric insurance cover- purchase insurance. In Mexico, for example, the bud- age are defined in Box 5.3 and Table 5.2. See Annex get law authorizes the national disaster fund FON- 7 for a complete description of parametric insurance DEN, through its Trust Fund, to use part of its annual coverage. budget allocation for the purchase of financial risk transfer instruments such as insurance and catastro- phe bonds. Such transactions are made through the public reinsurance company Agroasemex. Chapter 5: Options for National Disaster Risk Financing Strategy in Indonesia < 33 > Box 5.3. Parametric insurance coverage– potential key features Insured party: The National Disaster Reserve Fund. Coverage: BNPB and Ministry of Finance would identify the post-disaster activities to be financed, including emergency needs, affected low-income households, reconstruction of critical infrastructure and buildings. Perils covered: Parametric insurance is designed for specific perils such as earthquakes (possibly including tsu- nami) and tropical cyclones (possibly including storm surge). Localized risks, such as floods and landslides are more difficult to model and thus to cover under parametric triggers. Triggers and payouts: Payments would be made upon the trigger event conditions, such as magnitude/depth of an earthquake and wind-speed/central pressure of a tropical cyclone. Modeled losses (as estimated by a catastrophe risk model) could be used to better correlate with the actual losses. Risk zones: Risk zones would be identified in Indonesia for major perils, such as earthquakes. Specific para- metric insurance triggers would be designed for each risk zone. They can be covered under a single aggregate coverage (portfolio of risk zones and perils) or for each risk zone separately. Table 5.2. Summary of various trigger types Modeled Loss 1st Generation 2nd Generation Parametric Hybrid Methodology Model estimated Model probability Model probability of Model probability of losses directly. of certain types of exceeding a calculated index exceeding a calculated index events occurring in a score. Index is comprised score. Index is comprised predefined geographic of weighted measurement of weighted simulated area. stations. measurements. Required Inputs Event intensity (e.g., Event basic parameters Numerous measurements Event basic parameters to trigger epicenter location (e.g., epicenter and from a network of recording (e.g., epicenter location and magnitude, or magnitude, or landfall stations (e.g., ground motion and magnitude, or landfall landfall and central location and central intensity or wind speed). location and central pressure). pressure). pressure). Advantages Easy to understand, Easy to understand, Perceived as more closely Fast trigger with no with no basis risk can be triggered correlating event intensity reporting network required. between the original quickly after an event. parameters to losses than 1st Can be adapted to simulate loss estimate and the Generation. other damage metrics such trigger methodology. as affected population (e.g., PAGER). Disadvantages Requires a post- Higher basis risk than Basis risk (but less than More basis risk than a event remodeling of other types of triggers. 1st Generation). Requires straightforw2deled and the losses by a risk reliable, independent actual parameter values. modeling firm. reporting network. Potential measurement error/damaged measurement stations. < 34 > Indonesia: Advancing a National Disaster Risk Financing Strategy - Options for Consideration Box 5.4. Mexican Catastrophe Bond MultiCat In 2009, the Government of Mexico issued a four-tranche cat bond (totaling US$290 million) with a three-year maturity under the MultiCat Program. The issuer is a Special Purpose Vehicle (SPV) that indirectly provides para- metric insurance to FONDEN against earthquake risk in three regions around Mexico City and hurricanes on the Atlantic and Pacific coasts. The cat bond will repay the principal to investors unless an earthquake or hurricane triggers a transfer of the funds to the Mexican government. See Annex 9 for additional details. The GoI could complement its financial disas- World Bank DPL with Cat DDO, would be triggered ter risk transfer strategy by issuing catastrophe on average every 4 years, when the annual budget al- bonds against extreme losses caused by spe- location is exhausted.14 Whilst the GoI could increase cific perils. Catastrophe bonds are index-linked se- the annual budget allocation to US$1 billion instead curities that secure financial resources on the capital of securing a contingent credit line, it may be politi- markets, to be disbursed in case of the occurrence cally more sustainable to pre-fund losses through the of a pre-defined natural disaster. Cat bonds gener- reserve fund that are expected to occur once every ally cover the highest level of risk and are mainly 4 years and post-fund larger losses (by repaying any issued to specific perils with an annual probability of drawn down debt) expected to occur on average ev- occurrence of 2 percent or less (that is, a return pe- ery 4 to 20 years. A contingent credit line enables riod of 50 years or more). Mexico issued cat bonds the government to save reinsurance costs without the in 2006 and in 2009. The 2009 MultiCat program is need to increase the annual budget allocation beyond described in Box 5.4 and Annex 9. what would be politically sustainable. An indicative national disaster risk financing The contingent credit could be used as a bridg- strategy is proposed below. The fiscal disaster ing facility. The contingent credit can act as bridg- risk profile described in Chapter 2 and the above- ing funding facility for post-disaster rapid recovery mentioned risk layering approached are the basis for to suit Indonesia’s specific budgetary needs. The fa- the design of a comprehensive national disaster risk cility can be set with flexible triggers, to allow GoI to financing strategy. It relies on a three-tier risk financ- access short-term liquidity (especially when a disas- ing approach: reserves, contingent credit and risk ter occurs at the end of budget revision cycle), and transfer (e.g., parametric insurance). Figure 5.2 illus- to allow GoI to repay the loan once the new fiscal trates the indicative disaster risk financing strategy. year starts. The annual budget allocation could increase to A (parametric) disaster insurance coverage of 0.5 percent of the annual government budget US$800 million could be purchased to leverage expenditures, or US$500 million. This annual allo- the financial capacity of the GoI. This insurance cation is estimated to allow the government to cover coverage would protect the government against the recovery costs of recurrent natural disasters, for events with a return period up to 4 years. The GoI 14 The Cat DDO is shown to be at least 40 percent less expen- has already increased the annual budget allocation sive (in expected net present value) than the average disaster to IDR 4 trillion (approx. US$450 million) in 2011. insurance premium for medium risk layers. This analysis is based on the comparison, for a given risk layer, between the average disaster insurance premium and the expected net A contingent credit line of US$500 million could present value of the cost of the CAT DDO (which depends on be secured to increase the retention capacity the free end fee, the interest rate and the discount rate). See of the GoI. This contingent credit line, such as the Clarke and Mahul (2011). Chapter 5: Options for National Disaster Risk Financing Strategy in Indonesia < 35 > major disasters, occurring once every 20 years or less nesia can quickly access up to US$1.8 billion in frequently. It would ensure quick access to liquidity case of major natural disasters occurring once to cover the recovery costs faced by the government every 100 years. If the GoI wants to strengthen in the first three months after a disaster. (resp. reduce) its level of financial resilience, it could increase (resp. decrease) the insurance coverage ac- The national disaster risk financing strategy cordingly. would ensure that the Government of Indo- Figure 5.2. Indicative disaster risk financing strategy Return Exhaustion point Coverage AEL Loss on Period Prob. of first loss (US$ million) (US$ million) (US$ million) Line (years) 1,800 Catastrophe risk insurance 16 2% 19 5% 800 1,000 Contingent Credit 66 13% 4 28% 500 500 National Disaster Fund 338 68% - 100% 500 Note: AAL: Annual Average Loss; Loss on Line: AAL/Coverage; Probability of first loss: probability that the risk layer is triggered. Source: Authors. Contingent credit could also be used by the tion, simulations using a dynamic financial analysis GoI to build up multi-year reserves quickly at a (DFA) model show that a US$500 million Cat DDO lower cost. The GoI could establish a multi-year re- would allow the GoI to more than double the aver- serve mechanism, where the unused annual budget age net reserves of a dedicated disaster reserve fund allocation can accumulate over time. As an illustra- at end of 10 years. See Figure 5.3 and Table 5.3. Figure 5.3. Net reserves of the multi-year reserve fund, with US$500 million Cat DDO 3,000 Reserves minus any outstanding 2,500 2,000 debt (US$ millions) 1,500 US$ million 1,000 500 – -500 -1,000 -1,500 -2,000 1 2 3 4 5 6 7 8 9 10 11 Year Source: Authors. < 36 > Indonesia: Advancing a National Disaster Risk Financing Strategy - Options for Consideration Table 5.3. Net reserves at the end of 10 years, with and without a US$500 million Cat DDO Without Cat DDO With Cat DDO Percentage change 391 1,121 187% 19 61 228% 193 1,121 480% 875 2,293 162% Assumptions: WB interest rate = 4.4%; Interest rate on reserves = 3%; discount rate = 5%; post-disaster interest rate = 8%. Source: Authors. Establish a National Disaster disburse funds quickly after a disaster to allow for Reserve Fund as fast-disbursement rapid implementation of recovery operations. This mechanism for the financing Facility could build on the successful example of of post-disaster Mexico, which established the national disaster fund FONDEN. See Box 5.5. The post-disaster budget allocation process is currently slow and can generate a liquid- The NDRF could rely on existing disaster fund- ity crunch. The Rehabilitation and Reconstruction ing mechanisms. The NDRF as a disaster risk fi- Fund, with an annual budget allocation of IDR 4 nancing vehicle would establish sufficient transpar- trillion in 2011, is one of the main funding sources ency and internal controls to ensure coordinated for post-disaster recovery and early reconstruction. allocation and efficient use of post-disaster funds. Recent experience shows however that it can take Monies allocated under sectoral and disaster con- up to several months to draw down funds from the tingency budget would be immediately available for Rehabilitation and Reconstruction Fund because the disbursement to BNPB, or implementing line agen- disbursements must be approved by Parliament. This cies during the recovery phase of a disaster, thereby can generate delays in post-disaster recovery and re- removing the bottleneck encountered while await- construction operations. ing parliamentary approval. The existing On-Call budget should be part of NDRF, whereas the cur- Without improving this budget allocation pro- rent restriction (in the GR 22/2008) on disaster con- cess, the national disaster risk financing strate- tingency funds being only for pre-disaster activities gy would be ineffective. The national disaster risk should be removed such that it fully functions like a financing strategy suggested in the previous chapter financial trust. aims at allowing the GoI to access immediate fund- ing in case of a disaster. However, if this funding The NDRF could be established as a public ser- takes time to be allocated and executed, it does not vice agency. Several options can be considered for serve the ultimate objective to allow for fast imple- the financial management of the NDRF under the mentation of post-disaster recovery and reconstruc- current legal and regulatory framework. Under Law tion operations. 1/2004 of State’s General Treasury, a Badan Layanan Umum (BLU) is a non-profit institution that can be A National Disaster Reserve Fund (NDRF) could established within a line ministry or government be established as a mechanism for the rapid fi- agency with the purpose of providing public goods nancing of post-disaster operations. A basket of and/or services. Although the BLU reports to a line mechanisms and instruments could be developed as ministry or government agency, it is managed inde- part of a NDRF, akin to a financial trust, which would pendently, similar to a firm. The GOI has established Chapter 5: Options for National Disaster Risk Financing Strategy in Indonesia < 37 > Box 5.5. The Mexican Fund for Natural Disasters (FONDEN) The Federal Government created FONDEN in 1996 as a financial mechanism to provide the Federal agencies and the Mexican states with post-disaster financial resources. The mandate of the FONDEN is to: (i) finance post-disaster emergency assistance (through a revolving fund); and, (ii) provide the 32 Mexican States and the line ministries (e.g., Ministry of Infrastructure, Ministry of Health, Ministry of Education, Ministry of Human Development) with financial resources in case of natural disasters that would exceed their budget capacity. The FONDEN finances the post-disaster recovery and reconstruction of public assets (100 percent of Federal assets and 50 percent of state and municipal assets) and low-income houses. A Trust Fund was especially established to finance the FONDEN program. The Government of Mexico created FONDEN as a response to the delays faced in the post-disaster financing of emergency and recovery activities. See Annex 8 for further details. public entities like hospitals, universities, the public public buildings (e.g., schools, hospitals). A proper R&D agency, and the road management agency as fiduciary oversight would be developed. It could BLUs. An important feature of the BLU is that it is draw lessons from the newly established Indonesia given flexibility to manage its revenue and expendi- Multi Donor Funding Facility for Disaster Recovery tures as a corporation, without going through the (IMDFF-DR). regular government budget appropriation process that has to be approved by the Parliament every For long-term reconstruction and recovery time requiring modification. Hence, a BLU could be needs sectoral and sub-national budget alloca- the financial vehicle that ensures fast disbursement tions should remain the main source of funding for the financing of post-disaster emergency and as part of NDRF. As disaster risk management is a recovery operations (such as for housing and liveli- shared responsibility between national, sub-national hood subsidy). However, a requirement to establish and sectoral authorities, major long-term and multi- a BLU is that the entity must have a regular revenue year reconstruction and recovery needs should re- stream. As an alternative to a BLU structure, other main within the respective authorities having over- options - such as the creation of a public insurance sight over the assets and activities impacted by company - could also be considered. disasters. As part of NDRF, the Government may de- velop an inter-sectoral reconstruction budget coor- The NDRF would have a bridging facility and dination process for rehabilitation of national assets, would be responsible for the financing of emer- whilst using regional transfer mechanisms such as gency assistance and post-disaster recovery ac- DAK, deconcentration and co-administration (Tugas tivities. Prioritization and implementation planning Pembantuan) to assist sub-national governments. of activities to be financed would be completed by BNPB, local disaster management agencies, and line The NDRF would report to both BNPB and the ministries. The NDRF would be mandated to pro- Ministry of Finance. The Ministry of Finance would vide funds for: i) post-disaster emergency expenses provide the NDRF with an annual budget allocation managed by BNPB; ii) a compensation scheme for approved by the Parliament, to be deposited in a households affected by natural disasters; and, iii) financial entity akin to a Trust and managed by a disbursement for post-disaster and recovery to the public entity (BLU). The NDRF would operate ac- national and local agencies in case of critical damag- cording to a detailed Operations Manual (including es caused to infrastructure (e.g., roads, bridges) and detailed disbursement and execution procedures) < 38 > Indonesia: Advancing a National Disaster Risk Financing Strategy - Options for Consideration specifically designed for post-disaster response and agencies in charge of the emergency and/or recov- recovery operations. In the event of a disaster, BNPB ery activities. Parliamentary approval would no lon- would submit a financial request to the NDRF which, ger be needed to access the funds already allocated if in compliance with the Operations Manual, would in the NDRF. Figure 5.4 below depicts a possible in- trigger direct payments from the NDRF to the local stitutional framework for the NDRF. Figure 5.4. National Disaster Reserve Fund Ministry of Finance BNPB Line Ministries National Disaster Reserve Fund (NDRF) Fund drawn from On-Call-Fund Sectoral Disaster Contingency Reserve Contingency Fund Parliamentary in State General Treasury Process Disaster Contingency Fund Non-profit Agency within Gov- Bridging Funding Facility ernment (BLU) for Rapid Recovery for Disaster Financing The NDRF would be allowed to build up multi- designing and implementing a comprehensive risk year reserves. The NDRF would build up reserves financing strategy that may include contingent debt from the unspent amount of its annual budget al- agreements, the purchase of indemnity and para- locations over time in order to increase its retention metric insurance, and the issuance of catastrophe capacity. A combination of applicable existing public bonds or alternative risk transfer mechanisms. financial management mechanisms such as through establishing BLU, escrow account, and others should The NDRF could be established under the ex- be used, while creating a dedicated multi-year fund- isting legal framework. The existing legal struc- ing framework for disaster may be considered. ture, as outlined in Law 24/2007 on Natural Disaster Management, provides a framework for establishing The NDRF would be allowed to purchase disas- the NDRF. Government Regulation Number 22/2008 ter risk transfer instruments in order to lever- could be amended to allow for the formalization age its financial capacity in case of a disaster. of this funding mechanism. Annex 15 provides an Government regulation would be required to allow analysis of the existing Government regulation and the NDRF to pay disaster insurance premiums out options available to establish the NDRF. of its annual budget allocation. With this approval, the NDRF through its BLU would be responsible for Chapter 5: Options for National Disaster Risk Financing Strategy in Indonesia < 39 > Establish a disaster risk insurance lic assets would allow for economies of scale and program for public assets diversification benefits, and thus, lower reinsurance premiums. Public assets, such as schools and hospitals, and public infrastructure, such as roads and bridges, can be severely affected by natural di- Promote property catastrophe sasters and are currently not insured in Indo- insurance of private dwellings nesia. Given easy access to capital markets, most The current penetration of catastrophe prop- developed countries self-insure their public assets. erty insurance is low in Indonesia. Despite the In other words, because these countries have access efforts of the specialized reinsurer PT Maipark, less to bond market funding, they bear the full cost of recovery/reconstruction when a disaster strikes. In than 5 percent of the properties are currently in- Indonesia, most public assets and infrastructure are sured against natural disasters - and most of them not insured, although some provinces/municipalities are commercial and industrial properties. This low have recently insured selected public assets. See An- penetration is a direct consequence of the low de- nex 3 for recent developments. velopment of the non-life insurance market in Indo- nesia. In some middle-income countries, where fis- cal resources and access to capital are limited, The GoI may want to promote property catas- some governments require by law that pub- trophe insurance for private residential dwell- lic assets have a property insurance coverage ings. A developed domestic property catastrophe against natural disasters. This is the case in Latin insurance market would reduce the GoI’s implicit American countries such as Costa Rica, Mexico, and contingent exposure to major disasters. To promote Colombia. However, in practice, most public assets market development, the GoI could finance and remain uninsured or under-insured, partly because make available exposure and loss models to pri- the public managers are reluctant to spend part of vate insurers. It could also support information and their limited budget to pay an insurance premium awareness campaigns. and they often lack basic information to select a cost-effective insurance coverage. Turkey provides an interesting example of a homeowner’s catastrophe insurance program. A Disaster Risk Insurance Program for Pub- The Turkish Catastrophe Insurance Pool (TCIP) was lic Assets could be established in Indonesia to established in 2000 to overcome problems of market promote disaster insurance of public assets in failure in Turkey, namely a lack of local market earth- collaboration with the private insurance indus- quake capacity. The World Bank provided technical try. This program would aim at offering technical and financial assistance in the design stage of the assistance to public entities in the design of their TCIP to model and rate the earthquake exposure, catastrophe insurance coverage of public assets. as well as a contingent loan in the start-up imple- Standardized terms and conditions for the property mentation phase to cover claims as part of the risk insurance policies would be developed in collabora- financing program. A key feature of the coverage is tion with the private insurance industry that would that it is a simple property, earthquake only, policy assist public managers in identifying their risk expo- that is provided at affordable rates. Given the very sure and their insurance needs. The program could low voluntary demand by Turkish home-owners for also structure a national insurance portfolio of pub- insurance, earthquake insurance was made compul- lic assets to be then placed on the private (re)insur- sory for registered houses in urban centers. See Box ance market. A national approach to insuring pub- 5.6 for a short description. < 40 > Indonesia: Advancing a National Disaster Risk Financing Strategy - Options for Consideration Box 5.6. Turkish Catastrophe Insurance Pool The Turkish Catastrophe Insurance Pool (TICP) is a public sector insurance company that is managed on sound technical and commercial insurance principles. The TCIP purchases commercial reinsurance and the Govern- ment of Turkey acts as a catastrophe reinsurer of last resort for claims arising out of an earthquake with a return period of greater than 300 years. The TCIP policy is a stand-alone property earthquake policy with a maximum sum insured per policy of US$65,000, an average premium rate of US$46, and a 2 percent of sum insured deductible. Premium rates are based on construction type (2 types) and property location (differentiating between 5 earthquake risk zones) and vary from less that 0.05 percent for a concrete reinforced house in a low risk zone to 0.60 percent for a house located in the highest risk zone. Since inception, TCIP has averaged a penetration rate of about 20 per- cent, or 3 million domestic dwellings. See Annex 11 for further details. Should the GoI want to establish a private resi- ban areas. In Indonesia much of the rural hous- dential catastrophe insurance program, a num- ing stock is unlikely to meet minimum building ber of key decisions would need to be made, standards required by local insurers and their including whether: reinsurers; and, ■■ to form a public-sector catastrophe insurance ■■ to involve government in the program through fund, as in the case of Turkey, or to promote a public-private partnership. This could include some form of “coinsurance pool” through the the provision of start-up funding (research and involvement of the existing non-life private development costs). The GoI could also act as commercial insurers. The specialized reinsurer PT a reinsurer of last resort for extreme insured Maipark could play a central role; losses, when the financial capacity of the private sector is insufficient. ■■ to make homeowners property insurance com- Alternative disaster microinsurance products pulsory or to market the coverage on a volun- designed to protect the livelihoods of affected tary basis. The Turkish example showed that the households could be developed, as part of a demand by homeowners for property insurance comprehensive coverage against natural disas- was low due to the lack of an insurance culture ters. In light of the flood microinsurance program by Turkish homeowners and it was necessary to piloted by Wahana Tata, similar microinsurance make coverage compulsory; products could be designed to protect households ■■ to bundle property catastrophe insurance with impacted by recurrent natural disasters. Such micro- mortgages for homeowners or to keep as insurance products could be linked to some savings stand-alone coverage. Mortgage-linked ca- and/or credit mechanisms in order to offer a com- tastrophe insurance coverage could be made prehensive coverage against natural disasters. For compulsory. An alternative to make coverage example, microinsurance could build on the com- compulsory could be to bundle coverage with munity empowerment programs such as the PNPM property taxes; (Program Nasional Pemberdayaan Masyarakat), which already have a large network. ■■ to target the product at urban property own- ers alone or to target all households. In Turkey, Improved quality of insurance supervision would be earthquake insurance is only compulsory in ur- required to effectively promote catastrophe risk cov- Chapter 5: Options for National Disaster Risk Financing Strategy in Indonesia < 41 > erage among private insurers. The quality of insur- cient, thus significantly reducing the cost of securing ance supervision in Indonesia should be further im- additional capacity. proved through the use of a risk-based assessment of insurers’ retention capacity and reinsurance strate- The proposed Fund would act as a joint reserve gies based on catastrophe risk modeling and actuarial fund for the local governments. Participating tools. This would include the development of an ac- local governments would contribute to the tuarial model to further refine the commercial earth- Fund based on their own risk profile and de- quake premium rates and to assess the impact of sired coverage level. These contributions would natural disasters on the insurers’ portfolio. A scoring be used to maintain a reserve level sufficient to ab- tool to assess the quality and adequacy of the insur- sorb annual payouts to local governments affected ers’ reinsurance strategies could also be developed. by adverse natural events. To manage the potential variability in financial outflows, the Fund would se- cure additional financial capacity on the internation- Establish a Joint Disaster Reserve al reinsurance and capital markets. See Figure 5.5 Fund for Indonesia’s Local for an illustration of the proposed Fund. Governments Experience from the last ten years shows that The proposed Fund would help Indonesia’s local Indonesian local governments (e.g., municipali- governments access non-earmarked resources ties or provinces) often lack the financial re- quickly in the aftermath of a natural disaster. sources required to respond effectively in the To ensure transparency and avoid delays, payouts aftermath of natural disasters. Due to their limit- would be processed on a parametric basis. In con- ed size and economic base, many local governments trast to traditional indemnity insurance that makes (municipalities or provincial governments) do not claims payments based on formal confirmation of a have the capacity to set aside the required reserves loss, parametric instruments disburses funds based needed to finance the disaster losses not covered by on the occurrence of a pre-defined easily verifiable the Central Government. event - without having to wait for an on-site loss assessment. With immediate access to liquidity pro- The proposed Fund would help Indonesian lo- vincial governments would not be dependent on cal governments improve their response ca- Central Government transfers to finance emergency pacity at very low cost. Local governments are and recovery efforts after significant disasters. responsible for a fair share of the cost of natural disasters and they often struggle to mobilize the The Fund would provide the participating lo- required financial resources to support emergency cal governments with access to catastrophe and recovery activities. The proposed Fund would risk insurance at the lowest possible cost. Pre- act as a reserve facility holding financial reserves liminary analysis shows that a joint reserve Fund to become available in case of an adverse natural would allow for the participating local governments event. The joint nature of the Fund means that lo- to pool their natural disaster risks into one, better- cal governments would have immediate access to diversified portfolio, thus significantly reducing the significant resources without having to build these cost of reserves. The cost of financial protection is reserves over time. At the same time, local govern- highly dependent on the variability of the risk that ments would have access to significant resources is being insured. Since disaster risk among the local without having to bear the financial (and political) governments is not perfectly correlated, the cost of cost of holding these resources. Finally, by pooling coverage for a pooled portfolio would be less than their risks, local governments would be able to ac- the sum of coverage on an individual province-wise cess the reinsurance market where it is more effi- basis (See Box 5.7). < 42 > Indonesia: Advancing a National Disaster Risk Financing Strategy - Options for Consideration Figure 5.5. Joint Disaster Reserve Fund for the Indonesian Local Governments Reinsurance/ART (Purchased on international financial markets) Local gvt. 1 Local gvt. 2 Premium Payout Insurance Local gvt. 3 Premium Initial central GROWTH government Local gvt. 4 contribution Insurance Local gvt. 5 Payout Reserves Initial donor Local gvt. 6 contribution Local gvt. 7 Pooling disaster risks would reduce the cost The NDRF could initially serve as the manager of join reserves by 50 percent or more. To un- of the proposed joint reserve fund. Since provin- derstand the principles of the proposed Fund one cial and local governments are autonomous, local could consider a system through which several local government s would join the proposed fund on a governments would agree to combine their reserve voluntary basis. The NDRF would serve as the fund funds into a common pool. If each local government manager. were to build up its own reserves to sustain a po- tential catastrophic event, the sum of these local re- The Caribbean Catastrophe Risk Insurance Fa- serves would be much larger than the actual needs cility offers a successful example of such a pool. of the pooled local governments in a given year. Fig- The CCRIF is the result of two years of collaborative ure 5.6 below shows how the participation of each work between the Caribbean Common Market and province affects the level of risk capital needed by Community (CARICOM) governments, key donor the proposed Fund. The relative capital requirements partners, and the World Bank Group. The Facility of the fund to sustain a 1-in-150 year catastroph- became operational on June 1, 2007. Since then, ic event is reduced by 50% when seven provinces the Facility has disbursed more than US$30 million participate. It can be further reduced by 55% if the to the participating Caribbean countries affected by fund includes 15 provinces or more. natural disaster to help them finance their immedi- ate post-disaster expenditures. See Box 5.7 for a de- scription of the CCRIF. Chapter 5: Options for National Disaster Risk Financing Strategy in Indonesia < 43 > Figure 5.6. Pooling benefits among Indonesian provinces 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Number of participating provinces Note: Ratio of the 1-in-150 year PML of the aggregate risk portfolio of the participating provinces compared and the sum of the 1-in-150 PML of each participating province. Box 5.7. Caribbean Catastrophe Risk Insurance Facility The CCRIF functions as a mutual insurance company controlled by participating governments. The Facility was initially capitalized by participating countries, with support from donor partners. CCRIF helps Caribbean countries lower the cost of insurance by pooling risks. Insured countries pay an annual premium commensurate with their own specific risk exposure and receive compensation based on the level of coverage agreed upon in the insurance contract upon the occurrence of a major disaster. A portion of the pooled risk is retained through reserves, which helps to reduce the cost of insurance premiums. The CCRIF transfers the risks it cannot retain by purchasing reinsurance and catastrophe swaps. Coverage provided by the Facility is “parametric” in nature. Unlike traditional insurance settlements that re- quire an assessment of individual losses on the ground, parametric insurance relies on a payout disbursement contingent on the intensity of an event (e.g., wind speed, ground acceleration). In the case of CCRIF, payouts are proportional to the estimated impact of an event on each country’s budget, which is derived from a proba- bilistic catastrophe risk model developed specifically for the Facility. See Annex 10 for further details. < 45 > ANNEXES < 46 > Indonesia: Advancing a National Disaster Risk Financing Strategy - Options for Consideration Annex 1. Exposure to Natural Hazards in Indonesia Indonesia ranks 12th among the most vulnerable recorded - predominantly in Java. Most recently, in countries to high mortality risk from multiple haz- 2010, Mount Merapi erupted, killing 324 and dis- ards. The country is situated in one of the most active locating over 320,000 people. In 1815 the Tambora disaster hot spots, where several types of disasters volcano on the northern coast of Sumbawa, West such as earthquakes, tsunamis, volcanic eruptions, Nusa Tenggara Province erupted claiming more floods, landslides, droughts and forest fires fre- 92,000 lives and in 1883 the Krakatoa eruption quently occur. According to a global risk analysis by claimed more than 36,000 lives and created tsuna- the World Bank15, about 40 percent of the popula- mis as far away as South Africa. The islands of Java tion lives at high mortality risk. For a country that and Sumatra are also prone to landslides because of has more than 230 million people, this implies that their topographic and unstable soil conditions. more than 90 million people live at risk. The high frequency of disasters has an important Hydrometeorological  impact on public expenditures. According to the Within the past century, floods have been the most Government’s disaster data16, between 2001 and frequent disaster for Indonesia. A high rainfall regime 2007 alone, there have been more than 4,000 oc- in the west and a dry zone in some eastern provinces currences of disasters including floods (37 percent), are subject to recurring floods and droughts. Floods droughts (24 percent), landslides (11 percent), and often impact major population centers such as Ja- windstorms (9 percent). As the disasters damage karta (with a population of more than 13 million), public infrastructure and private homes, mostly un- Medan (more than 2 million), and Bandung (more insured, they create an enormous burden on public than 4 million). The government estimated that the expenditure to restore affected facilities. 2007 flood in Jakarta created total damage and losses of more than US$900 million17. According to the Ministry of Public Works, the annual flood in the Hazard Profile Bengawan Solo Watershed that occurred in 2007 cost the government more than US$200 million or Geologic equal to the total emergency allocation for all disas- Situated in the earthquake belt and Pacific ring-of- ters for the entire year of 200818. fire, Indonesia is highly vulnerable to earthquakes and volcanic eruptions. The areas most vulnerable Climate variability and change to earthquakes are Sumatera, Java, Bali, Nusa Teng- gara, Maluku, Sulawesi and Papua. Sumatera alone Deforestation and prolonged drought intensify the has suffered from over 15 large earthquakes in the occurrence of forest fires. The wildland fire and past 100 years. Indonesia also has 129 active volca- smoke-haze episodes in Indonesia during the 1980s noes, 70 of which are classified as dangerous. Be- and 1990s were the first documented influence of tween 2001 and 2007, 26 volcanic eruptions were drought impact triggered by the El Niño-Southern See World Bank, Natural Disaster Hotspots, A Global Risk 15 Analysis (Washington, DC: Disaster Risk Management Series, Laporan Perkiraan Kerusakan dan Kerugian Pasca Bencana 17 2005), Table 1.2 Banjir Awal Februari 2007 di WIlayah Jabodetabek, National DiBi database (Data and Information on Disaster in Indone- 16 Development Planning Agency (BAPPENAS) 2007. sia), National Disaster Management Agency (BNPB). http:// Source: Center for Strategic Assessment of the Ministry of 18 dibi.bnpb.go.id/. Public Works, April 2009. Annex 1. Exposure to Natural Hazards in Indonesia < 47 > Oscillation (ENSO). In East Kalimantan, nearly 3.5 mal settlements. The combination of the poor qual- million hectares of forests were affected by drought ity settlements and inadequate infrastructure has and fire. Nearly 0.8 million hectares of primary rain increased Indonesia’s vulnerability, especially when forest were burned, though impacts were more larger scale disaster events occur. widespread in logged-over and secondary forests (mainly in the vicinity of settlement areas). The cli- mate anomaly brought by El Nino also induced a de- Overall Risk Profile crease in rainfall impacting food production by an average of 3.06 percent. More frequent events, increased exposure, and lower coping capacity leading to great- er impacts Factors of Vulnerability The overall risk profile implies increased exposure and lower coping capacity in coming years. A com- Population increase and urbanization bination of Indonesia’s unique geological setting As in many other developing countries, economic and the complexity of its population settlements growth in Indonesia has shown a strong correlation have generally led to increased disaster occurrence with urbanization, both in the sense of people mov- with a tendency for significant human impacts (e.g., ing from rural areas to the cities and in terms of the loss of life and economic disruption). High and in- urbanizing of rural settlements. By 2008, at least 50 creasing population density, coupled with growing percent of the population was living in cities, and unplanned development in high risk zones, contin- urban areas were increasing in population at 4.4 ues to increase Indonesia’s vulnerability to recurrent percent per year, well beyond national population and large scale disasters. growth. Currently, more than 110 million people live in or around 60 cities that are predominantly locat- ed in the coastal areas, exposing them to common hazards such as earthquakes, flooding and commu- nicable diseases. The high population density and unplanned development in many of the larger cities has also increased the vulnerability of the population in the case of large-scale disasters. Increased exposure due to poorly enforced zoning and poorly maintained infrastruc- ture The high rate of urbanization in Indonesia, and lim- ited capacity of urban centers to provide adequate shelters and infrastructure, has led to the emer- gence of many unplanned settlements. Poor quality and enforcement of land use zoning in turn has led to many hazard prone locations being occupied by settlements, thereby increasing the exposure of the population to disasters. The Ministry of Public Works estimated that a quarter of urban population (or around 25 million people) lives in slums and infor- < 48 > Indonesia: Advancing a National Disaster Risk Financing Strategy - Options for Consideration Annex 2. Post-Disaster Risk Financing – Indonesia Case Studies Aceh (2004) the previously planned target of 1 percent of GDP (IMF 2005). A major tsunami impacted Aceh and other coun- In response to the event, the GoI spent US$206 tries located on the coast of the Indian Ocean on million for relief operations and US$204 million for December 26, 2004. In Aceh, the waves reached emergency response operations. The emergency up to six kilometers inland and destroyed 800 ki- response budget was executed by several public lometers of coastline. By January 2005, about agencies including Ministry of Social Affairs (food 125,000 people had died and 93,000 had been and clothing assistance), Ministry of Health, Min- reported missing. istry of Education, Ministry of Public Works (emer- The economic impact of the 2004 earthquake in gency road and bridges repair, shelter), and other the region of Aceh was estimated at two percent ministries such as Home Affairs, Defense, Police of national GDP. Losses were concentrated in one Forces, etc. region, Aceh, and amounted to US$4.5 billion, A total of US$7.8 billion was pledged by the inter- representing half of the provincial GDP. national community to support reconstruction. Of The direct fiscal impact of the earthquake on the that total, about US$6.5 billion (83 percent) was national budget was limited due to the substantial allocated to specific projects. This amount is high- donor pledges for reconstruction. Financial assis- er than the estimated cost of damage and losses, tance from external donors limited the increase of which was around US$4.5 billion, allowing for ad- the government budget deficit to only 1.25 per- ditional investments using the build back better cent of GDP, which was only slightly higher than (BBB) principle. See Figure A2.2. Figure A2.2. Aceh reconstruction - Financial Needs vs. Fund Available US$ Billion 10 9 8 Upgrading facilities in tsunami- and earthquake-affected areas and post-conflict reintegration 7 Donors–soft loans; 0.4 Donor grants; 0.1 and development; 0.8 6 Inflation; 1.3 Donors; 2.3 5 Nias; 0.4 4 3 NGOs; 2.0 Aceh damage and loss 2 assessment; 4.5 1 GOI; 2.2 0 Source: World Bank (2010) Annex 2. Post-Disaster Risk Financing – Indonesia Case Studies < 49 > The GoI disbursement rates for reconstruction cated funds (US$684 million) were executed by were slower compared to the fast disbursement December 2005 (one year after the disaster). Dis- for emergency response. The majority of the bud- bursement picked up in 2006, particularly in the get for tsunami reconstruction was approved in final months of the year and, by June 2008, almost June 2005 and less than 10 percent of the allo- US$5 billion were executed. See Figure A2.3. Figure A2.3. Fund Disbursed for Aceh Reconstruction (2005-2008) US$ million 6,000 n Donor n NGO n GOI 4,858 5,000 4,178 4,000 3,371 3,000 2,814 2,197 2,000 1,493 1,000 902 684 0 Nov-09 Dec-09 Jan-10 Feb-10 Mar-10 Apr-10 May-10 Jun-10 Jul-10 Aug-10 Sep-10 Oct-10 Nov-10 Dec-10 Jan-11 Feb-11 Mar-11 Apr-11 May-11 Jun-11 Jul-11 Aug-11 Sep-11 Oct-11 Nov-11 Dec-11 Jan-12 Feb-12 Mar-12 Apr-12 May-12 Jun-12 Source: World Bank (2010) Yogyakarta and Central Java (2006) tary, local governments, and various regional Bakor- nas.19 Regional Bakornas received a budget of around An earthquake with magnitude 5.9 RS occurred US$7.5 million for emergency response efforts. Yogyakarta and Central Java on May 27, 2006, District authorities were responsible for the distribu- resulting in major damage, particularly in the dis- tion of in-kind support from the Central Government. tricts of Bantul in Yogyakarta Province and Klaten Such support included 10 kilograms of rice per per- in Central Java Province. Estimated total dam- son, US$0.3 per person per day, a one-time grant of ages and losses caused by the earthquake were US$10 per person for clothing, and another US$10 US$3.1 billion. About 5,800 people were killed, for kitchen equipment per household. In addition, the over 38,000 were injured, and more than 127,000 Central Government provided full living expenses for houses were completely destroyed. three months to over 820,000 people whose homes were severely damaged. For houses that suffered only The National Coordinating Board for Disaster Man- minor damages, a one month allowance was provided. agement (BAKORNAS) coordinated the emergency response efforts with cooperation from Ministry of 19 The national disaster management agency BNPB was not es- People’s Welfare, Ministry of Social Affairs, the mili- tablished yet. < 50 > Indonesia: Advancing a National Disaster Risk Financing Strategy - Options for Consideration The cost of emergency operations for Yogyakar- of temporary shelters and their infrastructure (com- ta earthquake through Bakornas’ budget totaled mon kitchen, water and sanitation facilities, etc); US$21.5 million, out of which about UD$18 mil- iii) emergency repair of the damaged facilities; iv) lion was financed through the central budget and in-kind assistance to victims (food, clothing); and, US$3.5 million from international assistance. See iv) cash grants of US$215 for families had a family Table A2.1 for a full breakdown. member die during the flood. Table A2.1. Distribution of Funds for Emergency The central and provincial governments provided Operations after Yogyakarta’s earthquake funding and equipment and the total amount spent during the emergency phase was US$3.1 million. Bakornas’ Emergency Relief Budget USD million The Central Government provided the biggest con- Living Allowances 14.85 tribution, with US$2.1 million, which was distrib- Health 1.50 uted to five districts and two provinces. The five Public Works 1.00 local governments contributed up to US$0.49 mil- Personnel 0.51 lion, and US$0.27 million were received from other Total 17.86 sources, such as private donations. Source: Bakornas (2007) The Ministry of Public Works, as the sectoral agency Recovery and reconstruction operations from the responsible for the management of Bengawan Solo Central Government totaled US$610 million for two river basin, led the reconstruction and rehabilitation fiscal years (2006-2007), out of which US$540 mil- operations financed by its sectoral budget. This ef- lion was used for housing reconstruction. In addition fort was complemented by the Ministry of Forestry to these Central Government funds, the sub-nation- efforts in critical forest areas. al governments (provincial and local) contributed The reconstruction of public infrastructure and pri- US$140 million. Contributions from international vate dwellings was under the responsibility of the and national donors were estimated at US$107 mil- local governments. At least two local governments, lion, that is, only 11 percent of total public financing Karang Anyar and Wonogiri, allocated funds from for reconstruction. Insured losses were anecdotal. In their local budget to provide cash grants to repair total, less than 30 percent of the post-disaster dam- the damaged houses. For example, the amount al- age was covered by public financing. located in Wonogiri was US$400 per house with an additional US$200 provided in the case of heavily Bengawan Solo Flood (2007) damaged houses. After heavy rains in December 2007, floods affected The local governments of Wonogiri and Karang several districts in Central and East Java. Accord- Anyar spent US$0.0584 million for and US$0.318 ing to official figures from BAKORNAS, 109 people million for the post-disaster operations, respectively. were reported dead and 1,793 houses were heavily These two most affected local governments received damaged. The damage and losses were estimated at a reconstruction grant from the disaster response about US$173.1 million. budget of the Central Government to rebuild their public infrastructure in fiscal year 2008, including Local governments in the affected districts respond- US$3.075 million for Karang Anyar and US$2.621 ed immediately to the disaster. The following activi- million for Wonogiri. Unfortunately, there is no in- ties were carried out: i) evacuation of survivors and formation on how much was spent for reconstruc- victims of the flood; ii) organization and preparation tion from the local budget. Annex 2. Post-Disaster Risk Financing – Indonesia Case Studies < 51 > Table A2.2. Timeline of Fund for Housing Reconstruction, The case of three recent large disasters Post Disaster Housing Reconstruction by GoI Time lag until Disaster Date of Budget % of Housing Recon meeting recon fund Event Date Source of Fund Appropriation Needs* needs are met Yogyakarta May 27,2006 IBRD Loan Realloca- 13 July 2006 2.5 Earthquake tion Disaster Fund (999 Phase 1: October 42.5 11 months Budget Code) 2006 Phase 2: April 2007 42.5 MDF/JRF June 2007 10 West Java Sept. 2,2009 Disaster Fund (999 Phase 1: December 30 Earthquake Budget Code) 2009 60 11-12 months Phase 2: Aug-Sept 2010 West Suma- Sept. 30, 2009 Disaster Fund (999 Phase 1: December 4 tra Earth- Budget Code) 2009 > 11 months quake Phase 2: Aug-Sep 67 2010 Source: BNPB, Technical team for Yogyakarta and Central Java Reconstruction. ** Percentage is based on reconstruction plan Timeliness of post-disaster response allocation for March 2005 was only received by ben- eficiaries between May and June 200520. After the Funds for emergency relief operations are generally Yogyakarta earthquake in May 2006, funds were available in a relatively timely manner. Relief activi- distributed between June and August 200621. ties are mostly executed by local governments and financed through contingency budgets. The BNPB Timeliness of fund availability for rehabilitation and provides coordination and financial support from its reconstruction activities is more problematic. In the “on call” budget. For larger disaster events, many case of three recent disaster events, the time re- line ministries, such as Health, Education, Public quired to distribute the full allocations for housing Works, and Social Affairs, are involved in relief op- reconstruction took approximately 11 months. The erations. These activities are financed through the indicated timeline is summarized in the table below. emergency/”on call” budget lines allocated to the ministry budgets. In the case of Yogyakarta earthquake, which oc- curred May 26, 2006, the government was able to Subsequent to the relief phase, the Ministry of So- include the required funds in the mid-year budget cial Affairs provides a living cost grant (Jadup-Jatah revision. Therefore, by October 2006, US$270 mil- Hidup) to the displaced persons and receives an an- lion of the government allocation was appropriated nual budget allocation for the grants. If the funds and funds were distributed to beneficiaries between are inadequate, the Ministry can request replenish- October and December. The remaining 50 percent ment through the bi-annual budget revision process. of the government allocation was appropriated in April 2007, 11 months after disaster. The distribution of the living cost grant to the ben- eficiaries is often delayed. For example, after the 20 Source: BPK Audit Report Aceh Tsunami in December 2004, the living grant 21 Source: Various Media Reports < 52 > Indonesia: Advancing a National Disaster Risk Financing Strategy - Options for Consideration In the case of West Java earthquake of September fully disburse as the West Java earthquake. The first 2, 2009, 30 percent of the required funds, approxi- tranche was appropriated in December 2009; how- mately $50 million, were appropriated in December ever only US$31 million was appropriated, which 2009. The second tranche of funds was appropri- was enough to cover about 4 percent of housing re- ated in August-September 2010, roughly one full construction needs. The second tranche was appro- year after disaster. priated in August 2010, approximately 11 months after disaster. The West Sumatra earthquake occurred on Sep- tember 30, 2009 and experienced similar delays to Annex 3. Property Catastrophe Insurance in Indonesia < 53 > Annex 3. Property Catastrophe Insurance in Indonesia Insurance for catastrophic events is usually included astrophic reinsurance (see Box A3.1 for additional as an add-on product in the insurance policy. For ex- details). The premium rates were initially developed ample, it can be included under individual or indus- by Maipark and approved by the Indonesia General trial property insurance for all risks or for fire coverage Insurance Association (Asosiasi Asuranci, AAUI). The with extended risks (where extended risks can include insurance coverage provided by Maipark includes: earthquake, volcano eruptions, floods, landslides, etc). earthquakes, volcano eruptions, fire and explosions following an earthquake and/or volcano eruptions, In the case of earthquake coverage, a system of tsunamis, and business disturbances. compulsory cession of earthquake risks to special- ist earthquake insurer PT Maipark is in place. The As shown in Table A3.1, MAIPARK’s gross premium Ministry of Finance (MoF) established a compulsory growth between 2004 and 2008 was on average 21 earthquake pool in 2003 that was replaced by PT percent, while net premium growth between 2007 Maipark, a company that specializes in writing cat- and 2008 was around 4 percent. Table A3.1. MAIPARK Gross and Net Premiums and Claims (2004-2008) Gross Premium (US Gross Claim Claim Ratio Net Premium Net Claim Claim Ratio Year $ million) (US $ million) (%) (US $ million) (US $ million) (%) 2004 3.34 – – 0.01 0.02 221 2005 4.83 0.86 18 – – – 2006 6.04 2.07 34 – – – 2007 7.15 2.53 35 3.81 0.28 7 2008 8.23 0.77 9 4.51 0.42 9 Source: MoF Annual Report, 2008. Under current regulation, insurance companies are quired to cede five percent of the Total Sum Insured required to issue separate earthquake policies and are (TSI) up to a maximum amount of US$2.5 million per not permitted to coinsure them. They are also required insurance risk in West Java, Jakarta and Banten prov- to cede a specific portion of the earthquake risk to ince (Zone 5), or 25 percent of TSI to a maximum of Maipark at agreed rates. The deductible is set at 2.5 US$2.5 million per insurance risk in other areas. Pre- percent of the sum insured. Domestic insures are re- mium rates can be found in Table A3.2 below. Table A3.2. MAIPARK’s Insurance Premium Tariff (Zones I-V) Commercial and Industrial (Non Dwelling House) Construction Class Zone I Zone II Zone III Zone IV Zone V Steel, Wood and RC Frame ≤ 9 Stories 0.90 0.95 1.25 1.50 1.90 > 9 Stories 1.35 1.45 1.55 1.60 2.00 Others 1.00 1.10 1.55 3.00 4.70 Dwelling House – occupation code 2976 Construction Class Zone I Zone II Zone III Zone IV Zone V Steel, Wood and RC Frame 0.85 0.95 1.15 1.35 1.60 Others 0.90 1.00 1.55 2.75 4.50 Note: Rates are per mill (tenth of a percent) of TSI, with a 2.5% deductible. < 54 > Indonesia: Advancing a National Disaster Risk Financing Strategy - Options for Consideration Box A3.1. Maipark Background Maipark was founded in 2003 to take over the functions of Pool Reasuransi Gempa Bumi Indonesia (Indonesia Earthquake Reinsurance Pool). Shareholding participants in the company are non-life insurance and reinsurance companies operating in Indonesia. Maipark’s primary mission, as established by the Ministry of Finance, is to develop a national database on ca- tastrophe insurance in order to provide affordable insurance premiums in line with the government regulation 73/1992, which stipulates business conduct and adequacy of premium practices. In order to establish premium prices, it has divided the country into five earthquake zones, with the highest premiums being charged in the most dangerous zone (Zone 5). Furthermore, it is currently developing a comprehensive earthquake hazard map through a research and development project to identify all geological hazards in the country. In addition, Maipark acts as a national capacity coordinator for earthquake insurance, with the objective of managing and coordinating earthquake risks. Coverage provided by the insurance company includes earth- quakes, volcano eruptions, fire and explosions following an earthquake and/or volcano eruptions, tsunamis, and business disturbances. Property coverage currently provided includes: dwelling houses, home offices, office buildings, malls, factories, communication towers, schools, etc. Earthquake Risk Zone Map of Indonesia Annex 4. Insurance of Public Assets in Indonesia < 55 > Annex 4. Insurance of Public Assets in Indonesia The Central Government does not currently insure age was extended to earthquake risks. This insur- public assets against natural disaster risks and the ance coverage covers 42 local government build- majority of provincial and local governments also ings, four hospitals, 73 local government official self insure their critical assets. In recent years, there and guest houses. Insurance premiums were 1.25 has been some progress among local governments per mill (based on Maipark’s tariff). to buy insurance coverage for their public assets against natural disasters The insurance premium allocation is paid from the provincial government budget. The first insurance premium payment was for 42 local government PT Asuransi Bangun Askrida buildings, and the second payment was for local gov- ernment official’s houses and guest houses. The total PT Asuransi Bangun Askrida was established by the insurance premium of about US$20,000 is budgeted government as a state-owned company in 1989 to in the local provincial budget as “expenditure for in- provide insurance coverage for government build- surance of local government assets”. The first claim ings and property. Askrida was initially owned by experience was after the earthquake of September the Regional Development Bank (Bank Pembangu- 2009 and was settled in May 2010 (that is, seven nan Daerah), although in 1996 the shareholding months after earthquake) for an amount of US$2 mil- was extended to all local provincial governments. lion. The final claim settlement is still pending. Askrida now offers indemnity insurance products for fire, motor vehicle, engineering, personal accident, The Padang municipal government has also insured fraud, marine, surety bond and counter identity. some of their local government assets. However, the policy was not yet effective when the earthquake Given the lengthy insurance premium budget pro- took place in 2009. In fiscal year 2010, additional as- cesses, Askrida has proposed to some local govern- sets were insured among Padang municipal govern- ments that they include a catastrophe property insur- ments, including Puskesmas, a local government- ance protection plan within their budgets. Given the owned small clinic, and some schools. complexity and difficulty in monitoring a nationwide program, Askrida has only approached local gov- ernments for regional/provincial schemes. However, Yogyakarta Askrida would support a national plan to finance in- surance premium payments, at the local level, through Yogyakarta is one of many regions in Indonesia which is highly exposed to earthquakes, volcanic the national budget. Through this approach, payouts eruptions, tsunamis, tornados and landslides. Since would be allocated to local governments that would 2003, the municipality of Yogyakarta has insured distribute them to the community. Two examples of its public assets. These assets include government local governments insuring assets include West Su- buildings, schools, hospitals, traditional market matra Province Pandang, and Yogyakarta. places and motor vehicle. The aggregate insurance premium is estimated at US$90,000 in 2010 and West Sumatra Province, Padang US$20,000 million in 2011. The premium in 2011 decreased due to a smaller budget capacity of the In 2007, the West Sumatra Province government in- municipality of Yogyakarta. The average premium sured local assets—fire risk only—with PT Asuransi rate is 5 percent of the sum insured. The insurance Bangum Askrida insurance. Since 2008, the cover- scheme is opened to all insurance companies and < 56 > Indonesia: Advancing a National Disaster Risk Financing Strategy - Options for Consideration contract is given based on bidding process. The in- province. Losses and damages caused by this earth- surance companies Bumid, Sinar Mas and Ramayana quake were around US$3.12 billion, affecting most- are the main insurers of the city of Yogyakarta. ly private houses (Table A4.1.). After the earthquake, the City received a payout of US$340,000, that is, On May 27, 2006, an earthquake of magnitude of 14 times higher than the annual premium that year. 6.2 hit Yogyakarta and some areas in Central Java Table A4.1. Yogyakarta Earthquake Disaster Damage and Losses, 2006 Sector Damage Losses Total (Us$ million) (Us$ million) (Us$ million) Housing 1,390 140 1,530 Social 390 10 4,00 Economic productive 430 670 1,100 Infrastructure 40 20 60 Cross-sectoral 20 10 30 Total 2,270 850 3,120 Source: mof ( 2008) However, the vast majority of the buildings affected was received about 5 months after the disaster, de- had no insurance protection. As a result, most of laying recovery and reconstruction efforts. the damage and losses were borne by the local gov- Going forward, the City plans to set up compulsory ernment and the households, with some help from property catastrophe insurance for dwelling houses. the Central Government and the donor community. Insurance premium would be bundled with the PBB Financial assistance from the Central Government payment. Annex 5. Disaster Risk Financing and Insurance Framework < 57 > Annex 5. Disaster Risk Financing and Insurance Framework To help countries reduce their (over-)reliance on post rapid response once a disaster occurs. The World disaster external assistance, the World Bank has pro- Bank country catastrophe risk financing framework moted a disaster risk financing and insurance frame- is based on three pillars: work, which is partly based on corporate risk man- agement principles but also considers economic and ■■ Assessment of the government’s contingent lia- social factors such as the government’s fiscal profile bility. The first step in understanding the govern- and the living conditions of the poor (Gurenko and ment’s contingent liability is to develop precise Lester 2003, Cummins and Mahul 2009). risk models that accurately reflect the country’s risk exposure to natural hazards and the losses This risk management approach relies on the identi- associated with various events. Second, a dia- fication and assessment of the (implicit and explicit) logue must take place regarding the roles and re- contingent liability of the government in the event sponsibilities of the government and individuals of natural disasters and on the financing of this con- in the aftermath of a catastrophic event. The con- tingent liability, possibly using market-based finan- tingent liability of the government due to natural cial instruments. By ensuring that sufficient liquid- disasters is often implicit, as the law usually does ity exists immediately following a disaster, modern not clearly define the financial responsibility of funding approaches can help speed recovery, ensure the government when a disaster hits the country. that scarce government funds are well used, and re- The government thus acts as a (re)insurer of last duce the risk-enhancing effects of moral hazard. resort, without knowing precisely its catastrophe risk exposure. By understanding the full exposure With sufficient liquidity following a disaster, the gov- and the extent of public intervention in recovery ernment can immediately focus on early recovery efforts, it is possible to ascertain the contingent and not be distracted by having to close short-term liability carried by the government. funding gaps. At the same time, authorities can ■■ Promotion of commercial property catastrophe jumpstart reconstruction, particularly of key pub- insurance. The government can reduce its con- lic infrastructure (including bridges, hospitals, and tingent liability by encouraging private competi- schools). Finally, catastrophe risk management can tive insurance solutions for the transfer of pri- assist countries in the optimal allocation of risk in vately-owned risks, including property insurance the economy, which may result in higher economic and agricultural insurance. This can be done by growth, better risk reduction, and more effective creating an enabling environment that allows pri- poverty alleviation. vate insurers and reinsurers to offer competitive products and, possibly, through the establish- The sovereign catastrophe risk financing framework ment of catastrophe insurance programs based is part of a broader disaster risk management frame- on public-private partnerships, including catas- work promoted by the World Bank, which also in- trophe insurance pools. This allows the govern- cludes: i) risk assessment; ii) emergency prepared- ment to reduce its contingent liability in the case ness; iii) risk reduction; and, iv) institutional capacity of a natural disaster. The government can thus building. Catastrophe risk financing complements concentrate its financial support on the poor and other disaster risk management activities and pro- disadvantaged. tects against extreme events that cannot be effi- ciently mitigated. It can also provide incentives for ■■ Sovereign financial protection against natural prevention and preparedness activities and allow disasters. The government can manage its re- < 58 > Indonesia: Advancing a National Disaster Risk Financing Strategy - Options for Consideration maining contingent liability arising from natural ning and include reserves or calamity funds, bud- disasters by promoting the insurance of public get contingencies, contingent debt facility and risk assets and by protecting its budget against exter- transfer mechanisms. Risk transfer instruments are nal shocks through sovereign risk financing solu- instruments through which risk is ceded to a third tions, including reserves, contingent credit and party, such as traditional insurance and reinsurance, insurance. parametric insurance (where insurance payouts are triggered by pre-defined parameters such as wind Source of Financing Post-Disaster speed of a hurricane) and Alternative Risk Transfer (ART) instruments such as catastrophe (CAT) bonds. Governments have access to various sources of fi- nancing following a disaster. These sources can be The analysis of the fiscal management of natural categorized as ex-post and ex-ante financing in- disasters in Indonesia has identified possible post- struments. Ex-post instruments are sources that do disaster resource gaps. This time-sensitive analysis not require advance planning. This includes budget supports the design of a cost-effective disaster risk reallocation, domestic credit, external credit, tax financing strategy, as different financial instruments increase, and donor assistance. Ex-ante risk financ- are available at different periods after a disaster (Fig- ing instruments require pro-active advance plan- ure A5.1). Figure A5.1. Availability of Financial Instruments Over Time Short term Medium term Long term (1-3 months) (3 to 9 months) (over 9 months) Ex-post financing Contingency Budget Donor assistance (relief) Budget reallocation Domestic credit External credit Donor assistance (reconstr.) Tax increase Ex-ante financing Reserve fund Contingent debt Parametric insurance Traditional insurance Source: Ghesquiere and Mahul (2007) Among the ex post (post-disaster) financing tools, reconstruction loans from international financial in- contingency budget is the first to be immediately stitutions, such as the World Bank. available after a disaster. Other ex-post financing tools usually take more time to mobilize and are Ex ante financing instruments can provide immediate mainly available for the reconstruction phase. These liquidity after a natural disaster. These instruments include emergency recovery loans and post-disaster are designed and implemented before a disaster oc- Annex 5. Disaster Risk Financing and Insurance Framework < 59 > curs. These instruments include national disaster re- be used to cover the recurrent losses. Other sources serve funds, contingent credit and insurance. Small of financing such as contingent credit, emergency but recurrent losses can be retained through reserves loans and possibly insurance should enter into play and/or contingent credit. More severe but less fre- only once reserves and budget contingencies are ex- quent events, occurring for example once every 7 hausted or cannot be accessed fast enough. A “bot- years or more, can be transferred to the insurance tom-up” approach is recommended: the govern- or capital markets. Finally, international post-disaster ment first secures funds for recurrent disaster events donor assistance plays a role after the occurrence of and then increases its post-disaster financial capacity an extreme natural disaster. to finance less frequent but more severe events. The level of fiscal resilience to natural disasters, which Catastrophe risk layering can be used to design a drives the optimal financial strategies against natural risk financing strategy (see Figure A5.2). Budget disasters, is a decision to be taken by the govern- contingencies together with reserves are the cheap- ment based on economic and social considerations. est source of ex-ante risk financing and will generally Figure A5.2. Catastrophe Risk Layering High severity International Donor Assistance Insurance Linked Securities Risk Transfer Insurance/Reinsurance Contingent Credit Risk Retention Low severity Reserves Low frequency High frequency Source: Authors. A comparative analysis of the ex ante risk financing and risk transfer instruments is provided in Table A5.1. < 60 > Indonesia: Advancing a National Disaster Risk Financing Strategy - Options for Consideration Table A5.1. Contingent financing instruments for natural disaster. Product Benefits Costs/Risks/Constraints Risk Transfer Indemnity CAT No basis risk Works better in mature markets with solid local (Re)Insurance delivery systems and insurance regulatory framework Less technical work/investments involved in product design (follow the fortune approach) Market focused on asset based approach (concepts of interest for sovereigns like emergency relief, low Technology transfer expertise from international income housing, safety nets are considered usually markets being replicated worldwide for decades non insurable) Less restriction of geography/peril for a specific Difficult to create investor confidence on potential contract moral hazard when sovereign risk is involved Liability is transferred from gov’t balance sheet Up front premium to financial markets One year protection is the norm Counterparty credit risk Settlement of claims can take a long time Parametric (Re) No moral hazard, and more transparent for Basis risk Insurance risk-assuming counterparty Extensive and high-quality data sets are required Rapid disbursement of funds to model the hazard and quantify probability of a loss to the contract Multi-annual protection may be feasible22 High up-front costs (including cost of product Less insurance market infrastructure required (e.g. development and premium) claims verification) Counterparty credit risk CAT Bonds Limited credit risk. Vehicle is fully collateralized, but Basis risk for parametric and modeled loss CAT bond collateral is invested introducing some credit risk.23 triggers Access to a broader source of funding (Capital High up-front costs Markets + Insurance) Investors’ appetite for only very low probability events No moral hazard (depending on trigger type – (rarely below 1 in 75 year triggering events) indemnity trigger cat bonds still present moral hazard) Limited geography/perils by transaction Multi-annual protection (lock pricing for a period of 3 years usually) Historically has traded above CAT Reinsurance for similar risk layer Variety in options for triggers (indemnity, modeled loss, parametric and industry-loss linked products are It is regulated as an investment security (not possible) insurance) and therefore the legal framework can be complicated for sovereigns Parametric and modeled loss triggers can disburse rapidly Liability is transferred from gov’t balance sheet to financial markets 22 Parametric insurance is a relatively new concept, demonstrated for example by the Caribbean Catastrophe Risk Insurance Facility (CCRIF) established in 2007. These covers are more bespoke, and counterparties may be open to multi-year contracts such as that seen between Swiss Re and the Dominican Republic. The CCRIF paid out within 2 weeks of the devastating earthquake that hit Haiti in 2010. 23 The Total Return Swap structure, and permitted asset rules for collateral investment, in widespread use prior to the financial crisis exposed a number of bonds to credit issues during the crisis (largely due to the collapse of Lehman brothers). Since then, rules on permitted investments have tightened considerably and the current trend is to invest all proceeds in US Treasury Money Market funds. Annex 5. Disaster Risk Financing and Insurance Framework < 61 > Product Benefits Costs/Risks/Constraints Risk Transfer (continuation) CAT Derivatives Limited basis risk for large diversified portfolios Works only when there is a mature, credible (ex. Industry Loss of assets (settled on third party industry loss indices methodology to generate an aggregate industry loss Warranties) or tailor made indices) estimation which is not currently available outside of developed insurance markets24 Attractive to risk-assuming counterparty as there is no moral hazard, and product is easy to understand Typically only annual protection is offered Liability is transferred from gov’t balance sheet Counterparty credit risk (depending on where trade to financial markets occurs – many contracts are negotiated directly between counterparties) Weather Flexibility with regards to incorporate tailor made Sufficient historic data and ground measurement Derivatives indices tends to be limited in LIC Multi-annual protection available Basis risk Flexibility with regards to perils/geography of High up-front costs protection Counterparty credit risk Rapid payout Risk Financing Contingent Lower costs Financial impact is retained in gov’t balance sheet Credit Multilaterals No basis risk (Use of softer triggers that can be Institutions like the World Bank have an absolute size linked to gov’t actions like Declaration of Disaster) limit of 0.25% of GDP, which is very limiting in LIC (Ex. Cat DDO) because the potential impact of natural disasters can Flexibility on financial terms (including a longer term usually be substantially higher than any of the other risk financing alternatives) Funds are ring-fenced and are not at risk of depletion as a result of political pressure for purposes other than disaster response No counterparty credit risk (where the counterparty is the World Bank as per the Cat DDO) Structured Limited credit risk (fully funded vehicles) Basis risk (triggers/risks are usually limited on a similar Financing fashion as done in the CAT Bond space) Vehicles Possibility to generate positive cost of carry (service of debt repaid through the vehicle) Financial impact is retained in gov’t balance sheet Multi-annual availability Structured Risk Financing Finite Risk Can be used to combine risk retention (through These are ‘next generation instruments’ intended Contracts reserving), risk financing and risk transfer elements to complement existing risk retention and transfer into the program strategies. Therefore instruments are only suitable for institutions that already have a sophisticated risk Provides flexibility to include a wider spectrum of risks financing strategy in place, and that have technical (from lower to higher probability events) and flexibility capacity to accurately assess their risk in detail in how much of the risk is transferred versus retained Few countries have legislation in place to regulate Can combine both soft and tighter parametric these instruments triggers Lack of supervision has led some financial Multi-annual contracts (5 year terms are not intermediaries in developed countries to use these uncommon) tools to hide liabilities Contract includes cancellable clauses Legal language is sophisticated 24 ILWs trade for US perils, European windstorm and to a lesser extent Japanese earthquake. Third party industry loss providers recognized and accepted by the market include US Property Claims Services (PCS) and European companies (PERILS AG, Swiss Re Sigma, Munich NatCat services)funds. < 62 > Indonesia: Advancing a National Disaster Risk Financing Strategy - Options for Consideration Annex 6. World Bank Development Policy Loan with Catastrophe Draw Down Option The Development Policy Loan with Cat DDO is a opposed to “parametric” trigger, which means that contingent credit line that provides immediate li- funds become available for disbursement upon the quidity to IBRD member countries in the aftermath occurrence of a natural disaster resulting in the dec- of a natural disaster. It is part of a broad spectrum laration of a state of emergency. of World Bank Group disaster risk financing instru- ments available to assist borrowers in planning ef- The Cat DDO has a revolving feature; amounts re- ficient responses to catastrophic events. paid during the drawdown period are available for subsequent withdrawal. The three-year drawdown The Cat DDO helps develop a country’s capacity to period may be renewed up to four times, for a total manage the risk of natural disasters and should be maximum period of 15 years. part of a broader preventive disaster risk manage- ment strategy. The Cat DDO complements existing Pricing Considerations market-based disaster risk financing instruments such as insurance, catastrophe bonds, reserve funds, etc. The Cat DDO carries a LIBOR-based interest rate that is charged on disbursed and outstanding amounts. In order to gain access to financing, the borrower The interest rate will be the prevailing rate for IBRD must implement a disaster risk management pro- loans at time of drawdown. A front-end fee of 0.50 gram, which the Bank will monitor on a periodic percent on the approved loan amount and a renew- basis. al fee of 0.25 percent also applies. Key Features The Cat DDO provides an affordable source of con- tingent credit for governments to finance recurrent The Cat DDO offers a source of immediate liquidity losses caused by natural disasters. The expected net that can serve as bridge financing while other sourc- present value of the cost of the Cat DDO is estimat- es (e.g. concessional funding, bilateral aid or recon- ed to be at least 30 percent lower than the cost of struction loans) are being mobilized after a natural insurance for medium risk layers (that is, a disaster disaster. The Cat DDO ensures that the government occurring once every three years). This cost saving will have immediate access to bridge financing fol- can be even higher when the country’s opportunity lowing a disaster, which is when a government’s cost of capital is greater. post-disaster liquidity constraints are highest. Borrowers have access to financing in amounts up to US$500 million or 0.25 percent of GDP (which- ever is less). The Cat DDO has a “soft” trigger, as Annex 6. World Bank Development Policy Loan with Catastrophe Draw Down Option < 63 > Major Terms and Conditions of the Catastrophe Risk Deferred Drawdown Options Purpose To enhance/develop the capacity of the borrowers to manage catastrophe risk. To provide immediate liquidity to fill the budget gap after a natural disaster. To safeguard on-going development programs. Eligibility All IBRD-eligible borrowers (upon meeting pre-approval criteria). Pre-approval Appropriate macroeconomic policy framework. Criteria The preparation of existence of a disaster risk management program. Loan Currency EUR, JPY and USD Drawdown Up to the full amount is available for disbursement at any time within three years from loan signing. Drawdown period may be renewed up to a maximum of four extensions. Repayment Terms Must be determined upon commitment and may be modified upon drawdown within prevailing maturity policy limits. Lending Rate Like regular IBRD loans, the lending rate consists of a variable base rate plus a spread. The lending rate is reset semi-annually, on each interest payment date, and applies to interest periods beginning on those dates. The base rate is the value of the 6-Month LIBOR at the start of an interest period for most curren- cies, or a recognized commercial bank floating rate reference for others. Lending Rate The prevailing spread, either fixed or variable, for regular IBRD loans at the time of each drawdown. Spread 1. Fixed for the life of the loan: Consists of IBRD’s projected fund cost margin relative to LIBOR, plus IBRD’s contractural spread of 0.50%, a risk premium, a maturity premium for loans with average maturities greater than 12 years, and a basis swap adjustment for non-USD loans. 2. Variables set semi-annually: Consists of IBRD’s average cost margin on related funding relative to LIBOR plus IBRD’s contractural spread of 0.50% and a maturity premium for loans with average maturity greater than 12 years. The variable is recalculated on January 1 and July 1 of each year. The calculation of the average maturity of DDOs begins at loan effectiveness for the determination of the applicable maturity premium, but a withdrawal for the remaining components of the spread. Front-End Fee 0.50% of the loan amount is due within 60 days of the effectiveness date; may be financed out of loan proceeds. Renewal Fee 0.25% of the undisbursed balance. Currency Same as regular IBRD loans. Conversions, Interest rate Conversions, Caps, Collars, Payment Dates, Conversion Fees, Prepayments Other Features Country Limit: Maximum size of 0.25% of GDP or the equivalent of US$500 million, whichever is smaller. Limits for small states are considered on a case-by-case basis. Revolving Features: Amounts repaid by the borrower are available for drawdown, provided that the clos- ing date has not expired. < 64 > Indonesia: Advancing a National Disaster Risk Financing Strategy - Options for Consideration Annex 7. Parametric Insurance – Basic Concepts Parametric insurance has been developed over the Hazard Index past several years as a solution to the long delays incurred for claims payments in the aftermath of Hazard Modeling a major natural disaster and costly loss adjustment processes. Instead of indemnifying a specific loss A thorough assessment of the underlying hazard that is measured ex-post by adjustors on the ground, forms the basis of all parametric insurance contracts. a payout is made based on an ex-ante agreement of Hazard assessment is intended to accurately reflect the estimated loss caused by a specific size and type the hazard history and expected impacts. The char- of event. This ax-ante approach allows the insured acter of the hazard (its frequency and intensity) in a to receive a claim payment within a matter of days, specific territory is the primary factor in determining enabling recovery from the event to proceed more the cost of coverage. quickly. Using robust, stochastic models of tropical storms The design of parametric insurance product is based (hurricane) and earthquake hazards, the frequency on the following features: (probability) with which the hazard occurs at dif- ferent levels of intensity is determined. For the hur- ■■ Hazard Index. The hazard covered – generally ricane hazard, intensity can be expressed in wind earthquakes and hurricanes, but increasingly speed, and for earthquakes, ground acceleration floods and droughts – is modeled for probability can be used. As a result of this analysis, hazard ex- of occurrence and likely amount of loss at differ- ceedence curves are produced for each location of ent event intensities. interest. These curves (see Figure A7.1) depict the relationship between the intensity of the hazard and ■■ Payout Curve. Specific attachment and exhaus- the probability of that hazard intensity occurring. tion points are established to define the mini- mum size of event covered and the maximum coverage available for an extreme event. Based Figure A7.1. Wind Exceedence Curve on the amount of coverage desired, a contract payout curve is created, which defines the size of 1 in 500 the payout for particular event intensities. of occurrence Probability ■■ Trigger type. The mechanism by which the para- metric instrument will pay out is referred to as the “trigger.” The most appropriate type of trig- ger mechanism for a given transaction depends primarily on the issuer’s and investor’s needs and 1 in 10 risk preferences. 70 Wind Speed (mph) 145 Hazard intensity ■■ Payout Determination. When a major event occurs, the impact and location are verified by third parties. If it is determined that the hazard Loss Estimation is within the terms and conditions of the para- metric insurance policy, the payout is determined Damage and losses due to hazard impacts increase and paid to the insured party. exponentially as the intensity of the hazard increas- Annex 7. Parametric Insurance – Basic Concepts < 65 > es; that is, the rate of damage increases more rapidly Using a set of damage functions appropriate to than does the increase in the intensity of the hazard. the territory, in combination with information on A single hazard event can have varying impacts at the quantity and location of development and in- different locations within a country and hazard in- frastructure, a composite damage function (Figure tensities should be measured at multiple significant A7.2) can be derived for use with the results of the locations in the country for determining the para- hazard analysis. These damage functions are based metric trigger. on data derived from insurance claims and by engi- neering-based damage modeling. To understand the precise impact of an event, haz- ard intensities and frequencies are taken for several Payout Curve measurement points. These measurement points are pre-selected to correspond to key economic activity Attachment and Exhaustion Points areas in the covered area. The hazard values mea- sured at each of these points during a hazard event Parametric contracts include attachment and ex- are combined with a predetermined weighting that haustion points. The attachment point is the haz- reflects the economic loss potential of economic ard index value at which the contract is triggered, activities surrounding each measurement point. and functions like a deductible in a standard insur- This analysis helps ensure that the parametric haz- ance policy. Payouts are made on the policy when ard trigger reflects the actual impacts experienced the hazard index for an event in a covered territory across a territory. equals or exceeds the attachment point specified in the contract. The policyholder covers all losses for Composite Damage Function events that generate a hazard index below the at- tachment point. The relationship between the intensity of an event and the losses incurred as a result of the event is As the hazard index increases above the attachment known as a damage function. Damage functions point, the corresponding payout increases up to the take into account the resilience of the capital stock exhaustion point selected by the participating terri- in a given area and are specific to various catego- tory. The exhaustion point is the upper limit of disas- ries of infrastructure and building types in residential ter intensity to be covered. and commercial sectors. The policy limit is the difference between the attach- ment and exhaustion points (exhaustion – attach- ment) and is the maximum amount to be paid out Figure A7.2. Composite Damage Function under the contract. Payouts for events that have in- country hazard indexes that exceed the exhaustion 95% point will be paid at the policy limit. This limit applies Composite Damage to the full term (one year) of the contract; the total Percent damage Function amount paid out under the contract during the one- year period will not exceed the policy limit, whether that limit is reached due to payout from one large event or multiple smaller events that each trigger 10% payments under the contract. 70 Wind Speed (mph) 145 Hazard intensity < 66 > Indonesia: Advancing a National Disaster Risk Financing Strategy - Options for Consideration Selection of Attachment and Exhaustion Points Figure A7.4. Final Contract Payout Curve loss to government accts. When developing a parametric contract, a policy- Desired coverage for Exhaustion holder will identify a level of financial impact on the 35 government budget, beyond which it would want to receive an immediate cash injection. This value is an Attachment 15 appropriate starting point for identifying an attach- 550 980 ment point for the contract. Once an attachment point has been selected, the exhaustion point can 550 1000 Hazard index be set based on the cost of the contract and the maximum amount that the country is interested in paying for the catastrophe coverage. The frequency ■■ equation for calculating the hazard index (in- with which the hazard index exceeds the attach- cluding country-specific measurement points ment point and the range between attachment and and importance factors [weights]); exhaustion point are primary determinants of the premium cost. ■■ equation for the payout curve; and ■■ the attachment and exhaustion points. Figure A7.3. Fitting the Index Curve The cost of the final parametric contract is based on Policy Limit a pure premium charge derived directly from the fi- 20 nal payout curve, plus an administrative load to cov- Payout ($Mil) Attachment Exhaustion er costs such as reserve development, reinsurance, and administration. 0 550 980 Hazard index Trigger Mechanism Based on an index curve derived, it is possible to The mechanism by which a risk financing product identify the hazard index value that corresponds to (e.g., insurance, catastrophe (cat) bond) will pay out a specific payout amount. In the example shown in is referred to as the “trigger.” The most appropri- Figure A7.3, a contract attachment point of US$15 ate type of trigger mechanism for a given transac- million is selected. Based on the territory’s index tion depends primarily on the issuer’s and investors’ curve, this corresponds to a hazard index value of needs and risk preferences. Each trigger described 550, and an exhaustion point of US$35 million cor- below, including: i) modeled loss trigger; ii) first gen- responds to a hazard index of 980 on this same eration parametric trigger; iii) second generation curve. The selection of these two points results in a parametric trigger; and, iv) hybrid parametric trigger, policy limit of US$20 million (US$35 million – US$15 has a different level of basis risk. Basis risk is the risk million). that the payments received from the policyholder do not correlate well with the actual losses. Final Parametric Payout Curve An example of a final payout curve for one hazard Modeled Loss Trigger for a participating territory can be found in Figure A modeled loss trigger is based upon predicted A7.4. This curve is defined by the following, which losses to the underlying exposure database. Struc- will be explicitly included in the parametric contract: turing such a transaction is relatively straightforward Annex 7. Parametric Insurance – Basic Concepts < 67 > because the trigger amount and layer size are set ly and easily available, with no need to perform any directly to the levels of protection desired, such as post-event modeling or having recording stations on US$120 million policy limit coverage with an attach- the ground. The rate of occurrence of trigger events ment point equal to a 1-in-20 year event. When an in each specified zone and resulting loss probabilities event occurs, the loss to the insurance product is can then be computed and aggregated to estimate determined by simulating the actual event in its ca- the risk of the overall structure. tastrophe model and estimating its financial impact. Remodeling the event requires the collection of cer- This trigger was successfully used in the analysis tain event parameters such as epicenter location, performed for the CAT-Mex bond in 2009 for the depth and magnitude for earthquake, and landfall Government of Mexico. Note that although the location, maximum wind speed, central pressure CAT-Mex transaction triggered a full payout upon and precipitations for hurricanes. Because this type meeting its trigger, this type of binary payout is by of information is quickly available after an event, no means mandatory for this trigger type. modeled loss triggers can be structured to pay out relatively quickly. Second Generation Parametric Trigger Second generation or “index” parametric triggers First Generation Parametric Trigger are most common in areas of the world like Europe Known as the “cat-in-a-box” parametric trigger, pay- and Japan where a dense network of seismic or wind ment is based on the occurrence of predefined event stations is available, as this type of trigger requires parameters, such as an earthquake with a specified the measurement of event parameters at hundreds magnitude and hypocenter depth occurring in a or thousands of locations near concentrations of specified area, or a tropical cyclone whose eye with exposure. The event parameters recorded at each a specified maximum central pressure crosses with- station are combined with station-specific weights in a specified “box” drawn around the territory or meant to account for the relative accumulation of populated areas to be protected. The advantage of exposure by means of a formula such as described such a trigger is that the event parameters are quick- in Box A7.1 below. Box A7.1. Second Generation Parametric Trigger Loss Formula # Prov Where: Loss j = a x b x S i=1 Wj x (P i,j) z ■■ Lossj is the estimated loss from event j, expressed in millions of US$; ■■ a is the payout factor, such as 16% for earthquakes and 23% for tropical cyclones; ■■ β is a constant determined through an optimization process, whose purpose is to minimize any discrepancy between the modeled loss estimate and the parametric formula (i.e., basis risk); ■■ Wi is the weight associated with recording station i, based on the relative amount of exposure near that station; ■■ (Pj,i)z is a polynomial to the power z. P is the intensity of the peril from event j as measured at re- cording station i. It typically represents spectral acceleration for earthquakes and wind speed for tropical cyclones, though it could also be adapted to measure precipitation and storm surge levels should the proper measuring stations be available. < 68 > Indonesia: Advancing a National Disaster Risk Financing Strategy - Options for Consideration The insurance product is structured to trigger based hybrid triggers do not require the use of full-blown on exceeding a given index value which corresponds modeling software to estimate the losses and sub- to the loss level for which the issuer wants pro- sequent payment from an event. Instead, the event tection. Beyond the need for a dense network of parameter values are input into a spreadsheet that measuring stations, this structure is not appropriate would automatically apply simplified formulas to de- to provide protection for events that would be ex- termine local intensity. Such a spreadsheet could be pected to destroy or otherwise render ineffective the made widely available to risk managers, who could physical stations. Storm surge and tsunami height, use it to generate a quick preliminary estimate of both elements modeled for the South Pacific, would potential payouts from an event by entering in the likely fall in that category. spreadsheet the widely-available event parameter values. Hybrid Parametric Trigger Hybrid triggers present features of both second Determination of Contract Payout generation parametric structures and modeled loss after a Hazard Event triggers. This trigger can be identical to a traditional second generation parametric trigger; with the im- Calculation of Hazard Index portant difference being the intensity of the event To determine contract payout after a hazard event, collected near each concentration of exposure is a hazard index is calculated for the event. Since predicted based on the event parameter values, in- equipment to measure wind speed does not exist at stead of being based on field measurements from each calculation location, standard, predetermined physical recording stations. models are used to calculate these intensities, using storm information from the official reporting agency. This trigger is most useful where the reporting net- work is not dense enough (or cannot report its ob- Using the calculated hazard values for the measure- servations fast enough) to provide sufficient param- ment locations and importance factors that were eters for a traditional second generation structure. defined in the development of the hazard index To simulate local event intensity such as ground function, the index value is calculated according to motion or wind speed, event parameters such as an the hazard index formula specified in each country’s earthquake’s epicenter location, hypocenter depth, contract (see Figure A7.5). and magnitude; or, a tropic storm’s landfall location, central pressure, and radius of maximum wind, are used as inputs to recreate a simulated ground mo- Figure A7.5. Wind Speeds Calculated tion or wind speed at each point of interest. The at all Measurement Points weights associated with these points are combined with the simulated event parameter in an index for- mula similar to that described in Box A7.1. Hybrids * 85 mph can also be calibrated to accommodate other met- rics of damage such as the number of people af- * 110 fected by an earthquake, a measure similar to the USGS PAGER system. * 102 105 * * 100 The concept of collecting relatively simple predicted Hurricane event intensity values to simulate the local impact Hurricane of wind and ground shaking is very close to that of modeled loss triggers. The difference being that Annex 7. Parametric Insurance – Basic Concepts < 69 > Calculation of Payout Amount Treatment of Multiple Events Once the hazard index has been calculated for a The policy limit is an annual loss limit—it is the maxi- particular hazard event that affected a participating mum total payment from the Facility over the course territory, the index value is compared to the attach- of the contract year, whether from one or multiple ment and exhaustion points for the covered territo- events. ry. If the hazard index calculated for the event in the territory is below the attachment point, no payment In the example shown in Figure A.7.7, a payment is made to for the event. If the hazard index for the of US$9.3 million is made for the first storm to trig- event exceeds the attachment, the payout amount ger the contract. A second storm occurs, with a haz- can be determined directly based on the attachment ard index of 851, which corresponds to a payout of and exhaustion points and the policy limit, as shown US$14 million. Since the combined total of the two in the theoretic Equation 1 below. payments (US$23.3 million) exceeds the policy limit, the second payment is capped at US$10.7 million, so that the total payment for the year is equal to the Figure A7.6. Payout Calculation policy limit of US$20 million. for First Hazard Event Storm 1 payout: $9.3 mil Figure A7.7. Payout Calculation Policy Limit 20 for Multiple Events Payout ($Mil) Attachment Exhaustion 9.3 Storm 1 payout: $9.3 mil 750 Storm 2 payout: Calculated–$14 mil 0 Capped–$10.7 mil 550 980 Total payout for year: $20 mil Hazard index Policy Limit Payout ($Mil) 20 Attachment 14 Exhaustion Equation 1. Payout Calculation Formula 9.3 750 851 (event index – attachment) 0 550 980 Payout = * policy limit Hazard index (exhaustion – attachment) The resulting payout amount cannot be less than zero or greater than the policy limit. Parametric instruments have significant advantages over indemnity coverage, particularly in regards to In the example shown in Figure A7.6, the hazard index speed of payment and low costs. These benefits for a specific event was 750. Since this index value is come at a cost, most prevalent of which is the risk above the attachment value of 550, this event triggers of not receiving a payment for a hazard event, or a payment on the parametric contract. The payout for receiving a payment that does not sufficient cover this territory for this event would be approximately the losses incurred. A summary of the pros and cons US$9.3 million, as demonstrated in Equation 2. of parametric instruments can be found in Box A7.2. Equation 2. Payout Calculation (750 – 550) Payout = * US$20 million (980 – 550) 200 = * US$20 million 430 = US$9.3 million < 70 > Indonesia: Advancing a National Disaster Risk Financing Strategy - Options for Consideration Box A7.2. Benefits and Challenges of Parametric Insurance Benefits to Parametric Insurance Challenges to Parametric Insurance No moral hazard. Moral hazard arises when insured par- Basis risk. Basis risk emerges when the insurance payout ties can alter their behavior to increase the potential likeli- does not exactly match the actual loss. By definition, the in- hood or magnitude of a loss. Parametric insurance policies dex used in a parametric contract is a proxy for the real loss, are exempt from moral hazard because the indexes used and thus one cannot exclude that the parametric insurance in the calculation of the indemnity payouts (for example, indemnity may slightly underestimate (or overestimate) the wind speed, ground motion) are independent on the indi- actual loss. Careful design of the terms and conditions of vidual actions of the governments. the parametric insurance policy is critical to minimize this basis risk. Recent catastrophe risk modeling techniques al- No adverse selection. Adverse selection occurs when the low for the design of composite indexes that better mimic potential insured has better information than the insurer potential losses. At the same time, it is important to re- about the potential likelihood or magnitude of a loss, thus member that the objective of a parametric instrument is using that information to self-select whether or not to pur- not to cover the full losses in a covered territory, but to chase insurance. This informational asymmetry problem is guarantee a minimum amount of liquidity in case of a ma- eliminated, as sophisticated country-specific catastrophic jor adverse natural event. risk models are developed to assess the frequency and se- verity of hurricanes and earthquakes. Model bias. Model bias is the possibility that the catas- trophe models consistently underestimate or overestimate Lower operating costs. Unlike traditional insurance, para- the type and probability of losses resulting from certain metric insurance does not require costly monitoring pro- catastrophic events. In other words, it is the basis risk be- cesses (since there is no moral hazard or adverse selection) tween reality and its (necessarily incomplete) representa- or loss adjustment processes. Parametric insurance prod- tion through a mathematical model. Model bias can be ucts depend exclusively on the realized value of the under- reduced through a thorough understanding of the catas- lying index as measured by independent agencies (such as trophe environment of a region including the type, distri- the U.S. National Hurricane Center or the U.S. Geological bution, quantity and vulnerability of its building stock to Survey, National Earthquake Information Center). disaster. The incorporation of this information into a model calibrated based on expertise and historical loss experience should limit model bias. The more information that is avail- Transparency. Parametric insurance contracts are based on able about local catastrophe activity, local building stock independently reported indexes and transparent indemnity and local loss experience, the lower the model bias is likely formulas. As such, they give little opportunity for litigation to be. between the parties. With payouts calculated based on a predefined formula included in the contract, and input data provided by an independent agency, the parties to a con- Technical limitations of insurable hazards. Because tract can calculate the potential impact of a disaster event parametric instruments rely on a calculated index, their use immediately after it occurs and start processing a claim. is limited to hazards that can be modeled with a sufficiently high level of confidence. Hurricane and earthquake models have been developed and tested for more than a decade No cross-subsidization. The detailed risk model permits and are under constant improvement (particularly follow- the individual assessment of the risk exposure of each par- ing Hurricane Katrina in the United States in 2005). How- ticipating territory in the pool. The insurance premium will ever, catastrophe risk assessment models for hazards like thus be calculated individually based on the estimated risk volcanic eruptions or tsunamis are still under development. faced by each territory. This process will ensure that oppor- tunities for cross-subsidization are kept to a minimum and remain negligible when compared to the benefits provided Market limitations of insurable hazards. The existence by the pooled portfolio. of a catastrophe risk model developed by an independent agency is a necessary but not sufficient condition to make Annex 7. Parametric Insurance – Basic Concepts < 71 > Box A7.2. Benefits and Challenges of Parametric Insurance (continuation) Benefits to Parametric Insurance Challenges to Parametric Insurance Immediate disbursement. Because no loss assessment is this risk insurable. Financial investors generally charge an required, parametric contracts allow for the settlement of uncertainty load in the premium to accept risks that are claims shortly after an event. It is expected that claims are new in the market. This uncertainty load can make the pre- settled within four weeks following a disaster, as weather/ mium so high, compared to the expected loss, that the risk earthquake information is available on a daily basis. This becomes uninsurable. This is currently the case for tsunamis rapid claim settlement is essential if the affected states are and volcanic eruptions. to get access to liquidity to cover emergency and early re- covery expenditures. Education. Parametric insurance is a combination of in- surance concepts and financial concepts. Education of Reinsurance and securitization. Parametric insurance is policymakers and government agencies will be essential to a new type of financial product where the underlying as- ensure that the instrument is understood and used appro- set is a physical index (for example, wind speed, ground priately by local authorities. motion). Financial markets are interested in these types of products, which are uncorrelated with their asset portfolio and thus allow for further diversification. While they are sometimes reluctant to invest in insurance and reinsurance companies, because they do not fully understand the risks faced by these companies, parametric instruments are gen- erally event specific, making them more transparent and thus more attractive to investors. This facilitates the access of the capital markets through securitization (for example, index-linked securities, including catastrophe bonds). < 72 > Indonesia: Advancing a National Disaster Risk Financing Strategy - Options for Consideration Annex 8. Mexican Natural Disaster Fund FONDEN Mexico has a long history of, and broad exposure to, in 1994, legislation was passed to require federal, natural disasters. Located on the along the world’s state and municipal assets to be privately insured. In “fire belt”, where 80 percent of the world’s seismic 1996, the government created the Fund for Natural and volcanic activity takes place, Mexico is a seismi- Disasters in the Ministry of Finance (FONDEN). cally active country. The country is also highly ex- FONDEN is an instrument for the coordination of posed to tropical storms and is located in one of the intergovernmental and inter-institutional entities to few regions of the world that can be affected simul- quickly provide funds in response to natural disasters. taneously by two independent cyclone regions, the FONDEN’s main purpose is to provide immediate fi- North Atlantic and the North Pacific. nancial support to federal agencies and local govern- ments recovering from a disaster, and in particular for To address its vulnerability to adverse natural events, the: i) provision of relief supplies; and, ii) financing Mexico has developed a comprehensive institutional for reconstruction of public infrastructure and low in- approach to natural disasters. The catalyst to com- come homes. FONDEN is also responsible for carrying prehensive disaster risk management was the Mex- out studies on risk management and contributing to ico City earthquake of 1985. The earthquake killed the design of risk transfer instruments 6,000 people, injured 30,000 others and left a to- tal of 150,000 victims. Total direct losses exceeded US$4 billion. Main Features of FONDEN Mexico established the National Civil Protection Sys- FONDEN was originally established as a budgetary tem (SINAPROC) in 1986 as the main mechanism tool to allocate funds on an annual basis to pay for for interagency coordination of disaster efforts. expected expenditures for disaster losses. In 1999, SINAPROC is responsible for mitigating societal loss FONDEN was modified through the establishment and essential functions caused by disasters. Respon- the FONDEN Trust Fund, a catastrophe reserve fund sibility for SINAPROC lies with the Interior Ministry. that accumulates the unspent disaster budget of Also within the Ministry of the Interior, the National each year. Center for Disaster Prevention (CENAPRED) was es- tablished. CENAPRED is an institution that bridges the Financial support is directed towards public infra- gap between academic researchers and government structure and low-income households who, due to by channeling research applications developed by their poverty status, require government assistance. university researchers to the Ministry of the Interior. The adverse natural events covered by the FONDEN consist of geological perils including earthquake, volcanic eruption, tsunami, landslide and hydrologi- The Fund for Natural Disasters cal perils including drought, hurricane, excess rain- (FONDEN) fall, hail storm, flood, tornado, wildfire. Despite developing an institutional approach to disas- The FONDEN is based on three complementary in- ters, all levels of government in Mexico were still reg- struments, the Revolving Fund, the FONDEN Pro- ularly required to reallocate planed capital expendi- gram and the FONDEN Trust Fund. The first provides tures towards financing post-disaster reconstruction monies for disaster relief efforts, the second supports efforts. Budget reallocations created delays and scal- reconstruction of infrastructure and the third man- ing back of investment programs, while also slowing ages Mexico’s catastrophe risk financing strategy. deployment of funds for recovery efforts. In response, Annex 8. Mexican Natural Disaster Fund FONDEN < 73 > ■■ Revolving Fund: This fund finances emergency FONDEN Institutional Structure supplies to be provided in the aftermath of a natural disaster, such as shelters, food, primary ■■ Located within the Civil Protection unit of the health care, etc. In the case of high probability Ministry of the Interior, FONDEN is a trust man- of a disaster, or imminent danger, the local gov- aged by one of Mexico’s main development ernments can declare a situation of emergency banks (Banobras). The structure of FONDEN in- and obtain resources from FONDEN immediately. cludes a counterparty in each of the 32 Mexican Doing so allows local governments to take mea- states, including Mexico City, in order to facili- sures to prepare for immediate relief needs. tate the assignment and management of federal transfers. The main advantage of this structure ■■ FONDEN Program: This program finances reha- is the ability to provide resources to state gov- bilitation and reconstruction projects for public ernments immediately, on average five days after infrastructure (owned by municipalities, state the disaster. governments and federal governments), and the restoration of natural areas and private dwellings ■■ The FONDEN Trust receives an annual allocation of low-income households following a natural from the Ministry of Finance to develop and man- disaster. age its risk financing strategy. The risk is layered, with some tranches retained and others trans- ■■ FONDEN Trust: This Trust Fund manages the ferred through various instruments. To transfer assets of the FONDEN, including its risk trans- risk to the reinsurance markets for parametric fer strategy (reinsurance and/or alternative risk coverage or the capital markets for Cat bonds, transfer instruments). The Federal FONDEN Trust the FONDEN Trust places excess risk first with the manages the financial resources provided by the public insurer AGROASEMEX. This entity passes Federal Government, including the annual bud- on the risk to the markets. get allocation. The State FONDEN Trusts, set up for each of the 32 states, manage the financial resources received from the Federal FONDEN Trust after a natural disaster. Figure A8.1. Organizational Structure of FONDEN Ministry of Finance Ministry of Interior Civil Protection Department of Finance FONDEN CEPREDEN Opereations Unit Source: FONDEN (2010) < 74 > Indonesia: Advancing a National Disaster Risk Financing Strategy - Options for Consideration FONDEN Program FONDEN Trust once the State Government has transferred its contribution. The purpose of this program is to provide financ- ing to state and local governments that are over- FONDEN Trust whelmed by the occurrence of a disaster. The assess- ment of losses to be co-financed by the FONDEN is The Federal Government aims to promote the pri- based on a specific procedure involving the local and vate insurance of specific public assets owned by federal authorities. This procedure includes six main Federal agencies and State Governments, thus re- steps and should not exceed 23 days after occur- ducing its financing dependence on the FONDEN in rence of the disaster: case of a natural disaster. The Federal Government has empowered the FONDEN to develop a catastro- 1. In the aftermath of a disaster, a specialized fed- phe risk financing strategy, relying on private risk eral or state agency (e.g., meteorological depart- transfer instruments such as reinsurance and catas- ment, geosciences department) certifies the oc- trophe bonds. This helps the FONDEN to increase its currence of a natural disaster and informs the financial independence and overcome some political State Government; economy issues. 2. Within 4 days after the occurrence of a disaster, the State Government sets up a technical com- The financial structure of the FONDEN is depicted in mittee to identify and assess the damage caused Figure A8.2. The public bank Banogras acts as the by the natural disaster; account manager of the FONDEN Trust. The public 3. Within 10 days, the technical committee pro- reinsurer Agroasemex intermediates any financial vides the State Government with a technical and transactions with the international reinsurance and financial evaluation of the natural disaster; capital markets. 4. Within 15 days, the State Government informs the Federal Government. The Ministry of Interior The FONDEN Disaster Risk Financing issues a declaration of state of natural disaster. Strategy for 2011 Meanwhile, the Ministry of Finance authorizes the FONDEN to release early partial contribution The disaster risk financing strategy of the FONDEN to the State; relies on a combination of risk retention and risk transfer. To execute this strategy, the FONDEN re- 5. Within the following 2 days, the Ministry of In- ceives an annual budget allocation from the Federal terior should: i) ensure that the requested assis- budget, which is sometimes complemented by an tance is related to the natural disaster; ii) verify exceptional budget allocation in the case of a major that the damaged infrastructure has not ben- disaster. In order to purchase insurance coverage the efited from the FONDEN in the past; if this is Federal law was modified to allow the FONDEN to the case, the proof of insurance of the damage transfer risk to the reinsurance and capital markets, infrastructure is requested; and iii) formally ap- with the insurance premium being defined as a ser- prove the co-financing of the reconstruction of vice in the government budget law. The transferring the damaged assets. of risk to the reinsurance and capital markets are 6. The claims are authorized to be financed by the intermediated by the public reinsurance company FONDEN. In case of federal assets, the Federal Agroasemex. Below, Figure A8.3 describes the FON- FONDEN Trust pays directly the contractor. In DEN’s disaster risk financing strategy for 2011. case of state of municipal assets, the Federal FONDEN Trust transfers the funds to the State Annex 8. Mexican Natural Disaster Fund FONDEN < 75 > Figure A8.2. Financial Structure of FONDEN Placement of insurance Management and risk transfer prodcuts of the trust account (e.g., cat bonds) Agroasemex FONDEN Trust Banobras Reinsurance/Capital Markets Source: FONDEN (2010) the FONDEN that are borne by the Federal govern- Figure A8.3. FONDEN Disaster Risk Financing ment (that is 100 percent of the damage to Federal Strategy of the Federal Government in 2011 assets and 50 percent of the damage to state/mu- nicipal assets and low-income housing). The losses Mexico MultiCat Bond reported to FONDEN include replacement costs (on MXN3.5 billion average 75 percent of the total losses) and improve- Indemnity-based Reinsurance ment costs (on average 25 percent of the total loss- MXN6 billion es). Only replacement losses are covered under the reinsurance treaty. As of March 2011, the Federal Exceptional Budget Allocation MXN2.5 billion Government is expecting to place a XL reinsurance treaty of MXN 6 billion in excess of MXN 12.5 billion. Annual Budget Allocation MXN 10 billion The FONDEN has also secured the protection of a ca- tastrophe bond. In 2006, FONDEN issued a US$160 Note: The Mexico MultiCat bond covers only earthquakes in three million catastrophe bond (CatMex) to transfer Mexi- zones and hurricanes in three zones. co’s earthquake risk to the international capital mar- kets. It was the first parametric cat bond issued by a sovereign entity. After the CatMex matured in 2009, To implement the risk financing strategy, the Federal Mexico decided to further diversify its coverage by budget includes a budget line of 0.4 percent of the pooling multiple risks in multiple regions. In Octo- government expenditures for the financing of public ber 2009, it issued a multi-peril cat bond using the assets and the FONDEN, which corresponds to MXN1 World Bank’s newly established MultiCat Program. 0 billion in 2011. In case this annual budget allocation The Federal government issued a four-tranche cat is insufficient, the FONDEN has the ability to receive bond (totaling US$290 million) with a three-year an exceptional budget allocation from the Federal gov- maturity, called MultiCat Mexico. It provides (bina- ernment reserve funds (such as the oil fund). ry) parametric insurance to FONDEN against earth- quake risk in three regions around Mexico City and For the first time, in 2011, the FONDEN is placing hurricanes on the Atlantic and Pacific coasts. The cat an indemnity-based excess-of-loss (XL) reinsurance bond will repay the principal to investors unless an treaty on the international reinsurance market. Rein- earthquake or hurricane triggers a transfer of the surance payouts are based on the losses reported by funds to the Mexican government. < 76 > Indonesia: Advancing a National Disaster Risk Financing Strategy - Options for Consideration Annex 9. Catastrophe Bonds in Mexico The FONDEN uses various instruments to support natural disaster risks to capital markets, while local states and entities in responding to natural obviating the need to build up excessive budget disasters, including reserve funds and risk transfer reserves. solutions. In 2006, FONDEN issued a US$160 mil- lion catastrophe bond (CatMex) to transfer Mexico’s Operating structure earthquake risk to the international capital markets. It was the first parametric cat bond issued by a sov- Mexico issued a four-tranche cat bond (totaling ereign. US$290 million) with a three-year maturity under the MultiCat Program. The issuer is a Special Pur- After the CatMex matured in 2009, Mexico decided pose Vehicle (SPV) that indirectly provides paramet- to further diversify its coverage by pooling multiple ric insurance to the FONDEN against earthquake risk risks in multiple regions. In October 2009, it issued in three regions around Mexico City and hurricanes a multi-peril cat bond using the World Bank’s newly on the Atlantic and Pacific coasts. The cat bond will established MultiCat Program, which helps sover- repay the principal to investors unless an earthquake eign and sub-sovereign entities pool multiple perils or hurricane triggers a transfer of the funds to the in multiple regions and reduce insurance costs. Mexican government. The SPV structure is displayed in Figure A9.1 and the Objective institutional arrangements are described below: The purpose of a MultiCat Program is to transfer di- 1. The FONDEN enters into an insurance contract saster-related risks, covering multiple hazards, to the with local insurance company Agroasemex. capital markets in order to reduce pressure on public 2. Agroasemex enters into a reinsurance contract budgets. Doing so ensures that adequate funds are with Swiss Re to transfer all of the catastrophe in place for relief activities. risk. 3. Swiss Re enters into a derivative counterparty Outcome contract with a Cayman Islands-based special purpose vehicle (MultiCat Mexico 2009 Ltd.) to ■■ The bond was oversubscribed, with broad distri- transfer the catastrophe risk. bution among investors. With this bond, Mexico transferred a pool of disaster risk to the market 4. The SPV issues floating rate notes (Cat Bonds) to for the first time; secured multi-year protection capital markets investors to hedge its obligations for the covered risks at a fixed price; and reduced to Swiss Re under the counterparty contract. The potential pressure on public budgets. Mexico ef- proceeds received from investors are invested in fectively locked in funding for disaster relief prior US Treasury money market funds and deposited to the event happening, rather than relying only in a collateral account. on public budgets after the event. 5. A separate event payment account is established ■■ The demonstration effect of this transaction for with a third party bank to allow the FONDEN to other emerging market countries is significant. receive parametric loss payments directly from It has paved the way for other highly exposed the SPV, subject to the insurance contract. countries to manage fiscal volatility and stabilize government budgets by transferring extreme Annex 9. Catastrophe Bonds in Mexico < 77 > Figure A9.1. Financial Structure of MultCat Mexico 5 Loss Payment Loss Payment Amounts Event Amounts Collateral Payment Solution Account Investments Investment Earnings Reference Rate + Swiss Interest Spread 1 Agroase 2 Reinsurance 3 SPV FONDEN Collateral 4 Investors mex Company Insurance Reinsurance Ltd. (Zurich) Counterparty Account Note Proceeds Contract Contract Contract Lessons Learned 3. The availability of data and statistics about the probability and severity of a catastrophic event 1. Countries need to have a strong legal and in- is critical. New countries and regions attempting stitutional framework in place for disaster risk to tap the catastrophe bond market will need a financing to facilitate the implementation of risk supporting cat risk model. Donor countries with transfer mechanisms, which should be part of a a specific interest in working on the develop- disaster risk management framework. ment of disaster risk management capacity in developing countries can play an important part 2. There is potential to replicate this type of trans- by financing risk modeling and transaction costs. action for other middle-income countries. The Mexico bond was significantly oversubscribed, 4. The World Bank’s role as arranger significantly proving that investors continue to exhibit strong increased investor comfort. Future transactions appetite for non-peak risks. will benefit from the standardized fees and de- sign structure offered by the MultiCat Program. Table A9.1. Summary of Terms: Mexico MultiCat 2009 Class A Class B Class C Class D Peril Earthquake Pacific Hurricane Pacific Hurricane Atlantic Hurricane Notional (US$ million) 140 50 50 50 S&P rating B B B BB- Maturity October 2012 October 2012 October 2012 October 2012 Interest Spread 11.50% 10.25% 10.25% 10.25% (over US Treasury Money Market Fund) Expected loss 4.65% 4.07% 4.22% 2.39% Multiple 2.47 2.52 2.43 4.29 Source: FONDEN. < 78 > Indonesia: Advancing a National Disaster Risk Financing Strategy - Options for Consideration Annex 10. Caribbean Catastrophe Risk Insurance Facility On average, one to three Caribbean countries are Structure and Description affected by a hurricane or an earthquake each year, although during severe hurricane seasons this num- The CCRIF functions as a mutual insurance company ber can climb much higher. In 2004, the region suf- controlled by participating governments. It was ini- fered a disastrous hurricane season, with 15 named tially capitalized by the participating countries, with storms. Hurricane Ivan, the strongest storm of the support from donor partners. CCRIF helps Carib- season, wrought devastation on the Cayman Islands, bean countries lower the cost of insurance by pool- Grenada, and Jamaica. In Grenada, 89 percent of ing risks. A portion of the pooled risks is retained the country’s housing stock and more than 80 per- through reserves, which reduces the cost of insur- cent of its public and commercial building structures ance premiums. The CCRIF transfers the risks it can- sustained damage. The damage was estimated at not retain by purchasing reinsurance and catastro- over US$800 million, or approximately 200 percent phe swaps. of Grenada’s GDP. The Heads of Government of the CARICOM were compelled by their experiences dur- The coverage provided by the Facility is parametric ing this catastrophic season to ask for World Bank in nature. Unlike traditional insurance settlements assistance in improving access to catastrophe risk that require an assessment of individual losses on insurance. the ground, parametric insurance relies on a pay- out disbursement contingent on the intensity of an Objectives event (e.g., wind speed, ground acceleration). In the case of CCRIF, payouts are proportional to the esti- The main objective of the CCRIF is to provide its mated impact of an event on each country’s budget. members with access to affordable and effective The estimated impact is derived from a probabilistic coverage against natural disasters. For a number of catastrophe risk model developed specifically for the reasons, small island states have difficulty absorbing Facility. the financial impacts of disasters, including: i) lim- ited budgetary capacity prevents them from estab- Insured countries pay an annual premium commen- lishing sufficient financial reserves; ii) cross-regional surate with their own specific risk exposure and re- subsidization of recovery efforts is generally impos- ceive compensation based on the level of coverage sible due to their limited size and economic diversifi- agreed upon in the insurance contract upon the oc- cation; iii) high debt levels limit their access to credit currence of a triggering event. after disasters; and, iv) limited access to catastrophe insurance due to the high transaction costs resulting Outcome from the relatively small level of business brought into these markets. CCRIF is the first-ever multi-country risk pool. Six- teen Caribbean countries joined in 2007 and have CCRIF enables countries to pool their individual risks renewed their policies each year since. Seven pay- into a single, better diversified, joint reserve mecha- outs have been made to date (see below for CCRIF nism. Through risk pooling, CCRIF provides coverage members and payouts). The CCRIF has been well to countries at a significantly lower cost than individ- received by the reinsurance market, which has pro- ual governments would incur if they had to maintain vided capacity at a low rate to the Facility. A US$20 their own reserves or if they were to independently million cat swap between IBRD and CCRIF was the purchase insurance in the open market. first derivative transaction to enable emerging coun- Annex 10. Caribbean Catastrophe Risk Insurance Facility < 79 > tries to access the capital market to insure against (e.g., joint reserves and improved reinsurance rates), natural disasters. enough countries must participate in the Facility. Furthermore, the CCRIF carries administrative costs that are shared by participants; a significant number Lessons Learned of participants are required to maintain an afford- able average administrative cost per country. 1. The CCRIF addresses one disaster risk financing need of small island states: access to immediate 3. Dialogue on risk financing can enhance discus- liquidity in the aftermath of a disaster. The CCRIF sions with decision makers on more comprehensive does not cover all losses that a country may incur, in- disaster risk management. Risk modeling developed stead it covers estimated liquidity needs for the first for risk financing products can provide useful infor- three to six months after a major catastrophe. When mation on the risk exposure of the analyzed econo- designing a disaster risk financing strategy, it is im- my. This information and related dialogue on finan- portant to understand that each country requires a cial protection can help sensitize decision makers to tailored combination of disaster risk financing tools. the need for more comprehensive strategies to deal There is neither a “one size fits all” strategy nor a with increasing losses from adverse natural events, “silver bullet” disaster risk financing tool. including actions to try to avoid the creation of new risks (e.g., territorial planning, building standards) 2. A critical mass of country participation in CCRIF is and to reduce existing risks (e.g., protective mea- required for the Facility to benefit from risk pooling sures, strengthening of infrastructure). and diversification. In order for Caribbean countries to benefit from diversification through risk pooling CCRIF Member Countries Payouts to Date (USD in Millions) Anguilla Grenada 8.5 to Barbados (2010) Antigua and Barbuda Haiti 3.2 to St. Lucia (2010) Bahamas Jamaica 1.1 to St. Vincent & the Grenadines (2010) Barbados St. Kitts and Nevis 4.2 to Anguilla (2010) Belize St. Lucia 7.8 to Haiti (2010) Bermuda St. Vincent & the Grenadines 6.3 to Turks & Caicos Islands (20108) Cayman Islands Trinidad & Tobago 1 to Dominica (2007) Dominica Turks & Caicos Islands 1 to St. Lucia (2007) < 80 > Indonesia: Advancing a National Disaster Risk Financing Strategy - Options for Consideration Annex 11. Turkish Catastrophe Insurance Pool Bridging the contents of Europe and Asia, Turkey ernment of Turkey acts as a catastrophe reinsurer of is highly exposed to severe earthquakes. Despite last resort for claims arising out of an earthquake their common occurrence, Turkey’s private insurance with a return period of greater than 300 years. The market was previously unable to provide adequate full capital risk requirements for TCIP are funded by capacity for catastrophe property insurance against commercial reinsurance (currently in excess of US$1 earthquake risk. Without adequate commercial pro- billion) and its own surplus capital (about US$0.5 bil- tection of residential buildings, the Government lion). faced a significant contingent financial exposure in post-disaster reconstruction of private property. The TCIP policy is a stand-alone property earthquake policy with a maximum sum insured per policy of ■■ In the aftermath of the Marmara earthquake in US$65,000, an average premium rate of US$46 and 2000, the Government worked to limit its finan- a 2 percent of sum insured deductible. Premium cial exposure to the residential housing market rates are based on the construction type (2 types) through the establishment of the Turkish Catas- and property location (differentiating between 5 trophe Insurance Pool (TCIP). The pool enables earthquake risk zones) and vary from less that 0.05 the Government of Turkey to ensure that owners percent for a concrete reinforced house in a low who pay property taxes on domestic dwellings risk zone to 0.60 percent for a house located in the can purchase affordable and cost effective cover- highest risk zone. age. In doing so, the government’s contingent fiscal exposure to earthquakes is decreased by The TCIP sold more than 3 million policies at market- the transferring of risk to the international rein- based premium rates (i.e., 23 percent penetration) surance markets, which reduces pressure to pro- in 2009, compared to 600,000 covered households vide post disaster housing subsidies. when the pool was established. To achieve this level of penetration, the government invested heavily in TCIP is a public sector insurance company which is insurance awareness campaigns and made earth- managed on sound technical and commercial in- quake insurance compulsory for home-owners on surance principles. The Pool operates as a genuine registered land in urban centers. The legal frame- public-private partnership with most, if not all, op- work for the program envisages compulsion en- erational functions outsourced to the private sector. forcement mechanisms in urban settings, while cov- TCIP purchases commercial reinsurance and the Gov- erage is voluntary for homeowners in rural areas. Figure A11.1 Operational Structure of the TPIC Board Governance and key operating decissions GDI Pool Manager TCIP Treasury policy, oversight, and Information systems and Risk assumption and revenue accumulation implementation reinsurance claims Insurers Distribution Annex 12. Disaster Risk Management Roles in Indonesia < 81 > Annex 12. Disaster Risk Management Roles in Indonesia Table A12.1. Disaster Management Responsibilities & Duties of Central & Local Governments Government Responsibilities Authorities Central Ex Ante ■■ Create and ensure execution of a disaster risk Government ■■ Integrate disaster risk management within the management framework within national develop- national development program ment planning ■■ Transfer sufficient funds to state budgets to ex- ■■ Set the rules for the declaration of natural disaster ecute risk mitigation activities at the national and local level ■■ Ensure adequate ‘on call’ funds are available to ■■ Develop policies to enable cooperation in disaster BNPB and line ministries for emergency response risk management with other countries, agencies efforts and international parties ■■ Develop and enforce land use planning framework ■■ Implement policies that prevent the depletion of natural resources, which increases vulnerability to ■■ Develop and enforce building code disasters Ex Post ■■ Provide support to communities and refugees impacted by disasters ■■ Transfer sufficient funds, based on requests made by local governments, to the Provincial govern- ments, who transfer to local governments for recovery/reconstruction Local Ex Ante ■■ Create and ensure execution of a disaster risk Government ■■ Integrate disaster risk management within local management framework within local development development planning planning ■■ Allocate funds for disaster risk management in ■■ Establish and execute policies to enable coopera- local expenditure budget provisions tion and coordination in disaster management and response with other provincial district and ■■ fulfillment of the rights of the community affected municipal governments by disaster, to protect them from disaster impacts and reduce disaster risks by utilizing the funds ■■ Implement policies that prevent the depletion of allocated in the Regional Revenue and Expenditure natural resources, which increases vulnerability to disasters Budget Ex Post ■■ Provide disaster response assistance to communi- ties impacted by disaster ■■ Reallocate budget resources for emergency and recovery efforts – prior to the receipt of Central Government funds, should the needs exceed the resources available at the local level Source: Law 24/ 2007. < 82 > Indonesia: Advancing a National Disaster Risk Financing Strategy - Options for Consideration Table A12.2. Role of National and Regional Disaster Management Agencies Institution Responsibilities Duties National Ex Ante ■■ Provide guidance and direction on disaster Disaster ■■ Create and implement a disaster risk manage- reduction efforts, including: disaster mitigation Management ment framework within national development emergency response; rehabilitation; and, recon- Agency program struction BNPB24 ■■ Disseminate disaster risk management informa- • Risk Identification tion to local communities • Vulnerability mapping ■■ Report status of disaster risk management efforts • Potential disaster impact assessment to the President on a monthly basis during normal • Early warning systems times and at all times during emergencies • Community outreach/awareness ■■ Execute and account for disaster recovery and ■■ Coordinate disaster risk management activities reconstruction donations and other international ■■ Deliver research, education and training assistance ■■ Develop an Early Warning System ■■ Report on use of funds received from the national budget ■■ Encourage community participation ■■ Develop guidelines for the establishment of Ex Post regional disaster management agencies ■■ Support communities impacted by disasters through coordination of recovery efforts Regional Ex Ante ■■ Provide guidance and direction on disaster relief Disaster ■■ Create and implement a disaster risk manage- efforts, including disaster prevention, emergency Management ment framework within national development response, rehabilitation, and reconstruction in Agency program accordance with local government policies and BPBD National Disaster Management Agency • Risk Identification ■■ Collect data and develop risk maps for disaster • Vulnerability mapping prone areas to inform local development • Potential disaster impact assessment ■■ Create and implement operational procedures for • Early warning systems disaster management • Community outreach/awareness ■■ Report status of disaster risk management efforts ■■ Coordinate disaster risk management activities to the Regional Head on a monthly basis dur- ing normal conditions and at all times during ■■ Encourage community participation emergencies Ex Post ■■ Manage collection and disbursement of money ■■ Support communities impacted by disasters and goods to disaster affected populations through coordination of recovery efforts ■■ Report on the use of budget revenues received from local expenditure budget Source: Law 24/ 2007. The National Disaster Management Agency is a minister-level non-departmental Government Institution. 25 Annex 12. Disaster Risk Management Roles in Indonesia < 83 > Table A12.3. Post-Disaster Risk Financing Activities Source of Financing Emergency ■■ Search, rescue, and evacuation Authority ■■ Provision of basic needs Emergency disaster response is under the authority of local/regional governments. Central Government ■■ Food provides assistance for events declared ‘national ■■ Water disasters’ and takes full responsibility through the ■■ Clothing BNPB, which leads the emergency activities. ■■ Health Funding ■■ Psychosocial services “On Call” funds from Central Government’s budget ■■ Provision of emergency shelter as well as line ministries and the BNPB, in the case of a national disaster. Within the line ministries, dur- ing emergency phase, only the Ministry for Social Af- fairs - which provides emergency goods and support – is active. Local and provincial governments provide emergency financing through their contingency bud- get line (they do not have “on call” facilities). Recovery ■■ Environmental rehabilitation Authority ■■ Emergency repairs to public infrastructure Recovery activities are under the authority of local/ regional governments. Central Government provides ■■ Housing rehabilitation assistance (not the full assistance if required and takes full responsibility for cost of rebuilding) national disasters. ■■ Psychological and social recovery ■■ Provision of health services Funding ■■ Establishment of security and order If disaster damage is less than 20% of the municipal budget, recovery is financed at the local level. If dam- ■■ Resumption of government services age exceeds 20% of municipal budget, assistance is ■■ Emergency resumption of utility services provided by the provincial government. If additional (electricity, communication, water, sanitation, funds are required, they are provided by the Central etc) Government through the Disaster Fund. If further ■■ Conflict resolution, in case of social disaster, ie funds are required, they are provided through real- inter-community conflict or terrorism. location of line ministry budgets. Local governments can also ask for assistance in the form of equipment to help rebuild infrastructure. Reconstruction ■■ Reconstruction of infrastructure Authority ■■ Reconstruction of education, health and Reconstruction activities are under the authority of community facilities local/regional governments. Central Government provides assistance if required and takes full respon- ■■ Full restoration of public and community sibility for national disasters. services ■■ Ensure proper disaster resistant design of Funding reconstruction works Funding for infrastructure is transferred from the Central Government to line ministry regional offices. It is usually budgeted in next fiscal years as capital expenditures. Source: Summarized from Law no. 24/2007, and Government Regulation no. 21 and 22/2008. < 84 > Indonesia: Advancing a National Disaster Risk Financing Strategy - Options for Consideration Annex 13. BNPB Database on Natural Disasters Figure A13.1. Number of people affected by natural disasters, by peril, 1800-2009 1,000,000 Number of people affected 100,000 10,000 1,000 100 10 1 1800 1850 1900 1950 2000 Floods Earthquake and Tsunamis All other perils Figure A13.2. Number of buildings affected by natural disasters, by peril, 1800-2009 1,000,000 Number of buildings damaged or 100,000 destroyed (log scale) 10,000 1,000 100 10 1 1800 1850 1900 1950 2000 2050 Floods Earthquake and Tsunamis All other perils Annex 13. BNPB Database on Natural Disasters < 85 > Figure A13.3. Buildings damaged and destroyed reported by BNPB, all perils, 1970-2009 600,000 500,000 400,000 300,000 200,000 100,000 0 1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 Source: BNPB. < 86 > Indonesia: Advancing a National Disaster Risk Financing Strategy - Options for Consideration Annex 14. Post-Disaster Operational Phases Emergency response/relief operations include society’s functions are restored, such as re-opening emergency assistance provided to the affected of schools, businesses, etc, even if only in tempo- population to ensure basic needs, such as the need rary shelters. They include, among other things, the for shelters, food and medical attention. This is the emergency restoration of lifeline infrastructure (e.g., provision of emergency services and public assis- water, electricity and key transportation lines), the tance during or immediately after a disaster in order removal of debris, the financing of basic safety nets, to save lives, reduce health impacts, ensure public and the provision of basis inputs (e.g., seeds, fertil- safety and meet the basic subsistence needs of the izers) to restart agricultural activities. It is also during people affected. This phase aims at stabilizing the this phase that engineering firms can be mobilized society, with termination of further loss. Such costs to start the design of infrastructure works that will can be difficult to estimate ex-ante, as they depend take place during the reconstruction phase. Govern- on the specific characteristics of the catastrophic ment may also have to subsidize the basic restora- event (location, intensity, time of the year (winter or tion of private dwellings, particularly for low-income summer), time of day (day or night), etc.), but are families, before the reconstruction phase starts. relatively small compared to the subsequent recov- ery and reconstruction operations. While relief costs Reconstruction operations generally center on are limited, they need to be financed in a matter of the rehabilitation or replacement of assets damaged hours after a disaster event. The capacity of govern- by a disaster. They include repair and rebuilding of ments to mobilize resources for relief operation at housing, industry, infrastructure and other physical short notice should be a key component of its risk and social structures that comprise that community financing strategy. or society. These include public building and infra- structure which are the direct responsibility of the Recovery operations following the initial relief ef- state. National or local authorities generally have to forts are crucial to limit secondary losses and ensure face obligations that go beyond their own assets. In that reconstruction can start as soon as possible. most cases, government will have to subsidize the They are the restoration and improvement, where reconstruction of private assets and, in particular, appropriate, of facilities, livelihoods and living con- housing for low-income families who could not oth- ditions of disaster-affected communities, including erwise afford to rebuild their homes. efforts to reduce disaster risk factors. That is, the Figure A14.1. Timing of post-disaster operational phases Resource requirement ($) Relief Recovery Reconstruction Time Annex 15. Disaster Risk Management Framework in Indonesia < 87 > Annex 15. Disaster Risk Management Framework in Indonesia Law 24/2007 provides a comprehensive framework for the implementation of disaster risk management ac- tivities in Indonesia, defining roles and responsibilities, institutional relationships, and funding guidelines. The Law provides clear definition of the government’s responsibility to allocate sufficient funds and, in particular, the allocation of on-call funds for emergency response. Law 27/2007 only provides the principles of funding allocation for disaster management in Indonesia, which includes: a. The definitions of responsibility for national, provincial and local governments for the allocation of suf- ficient funding for disaster management and emergency response in national and local budgets (APBN and APBD), as stipulated in Articles 6 (e and f) and 8 (d); b. The importance of public participation in funding provisions, as stipulated in Article 60 (2); c. The requirement that national, provincial and local governments allocate sufficient funds to cover their respective areas of responsibility, as stipulated in Article 61; d. The directive that during emergency response, BNBP can use ready to use funds (also known as ‘on-call’) that are allocated in BNPB’s annual budget, as stipulated in Article 62; and, e. The statement that further details on the management of disaster funding should be included in a gov- ernment regulation, as stipulated in Article 63. Government Regulation 22/2008 on Funding and Management of Disaster Assistance was enacted the fol- lowing year to provide more detailed guidance on the Disaster Management Fund. The Disaster Risk Man- agement Fund was allocated IDR 3.8 trillion in 2010 and IDR 4 trillion in 2011. This regulation defines the uses of the Disaster Management Fund, which comprises three financing vehicles, including: i) a contingency fund for disaster preparedness and mitigation activities; ii) a ready fund (known as ‘on call’ funds) to be used in emergency response; and iii) grant patterned social assistance funds to be used for livelihood recovery and reconstruction of critical public services. The majority of these resources are earmarked for the grant pat- terned social assistance fund, which provides monies to households impacted by disasters. Figure A15.1 below shows the schematic diagram of articles in the Law 24/2007 and Government Regulation 22/2008 that govern existing disaster funding mechanisms. Law 24/2007 establishes overall guidelines on a comprehensive disaster risk management strategy while Government Regulation 22/2008 governs the Di- saster Management Fund. The GR22/2008 leaves room for additions to the Disaster Management Fund that would enable the Government to enter contingent financing agreements and purchase disaster risk transfer mechanisms such as parametric insurance, CAT bonds and Alternative Risk Transfer Vehicles. Opportunities to incorporate a comprehensive disaster risk financing strategy are proposed in Figure A15.1. < 88 > Indonesia: Advancing a National Disaster Risk Financing Strategy - Options for Consideration Figure A15.1. Schematic Diagram of articles on disaster funding mechanism Law 24/2007 On Disaster Management General Provision GR22/2008 on Disaster Funding Principles and Objectives Roles and Responsibilities Article 3: Scope, source, usage, management, reporting Institution Article 4: Sources of funds Rights and Obligation of the Community Insurance coverage Roles of Business and International Organizations Article 5: Funding Allocation” contingency, on-call, social grant Conduct of Disaster Management Insurance premium payment Article 6: Contingnecy fund only Disaster Funding and Asistance Management for pre-disaster • Funding resposibility: Articles 6 (e, f) and 8 (d) • Public participation: Article 60 (2) Contingency can be used for RR • Allocation of Sufficient Fund: Article 61 • Use of ’on-call’ budget for emergency: Article 62 Article 23: Fun usage for pre, during, • Further stipulations through Government Regulation: and post disaster Article 63 Supervision Other Provisions Insurance payout can be used for all 3 stages of DM cycle : Stipulation currently not existed BNPB may establish and/or appoint public body to manage funding assistance Annex 16. Borrowing Capacity of Indonesia < 89 > Annex 16. Borrowing Capacity of Indonesia The level of debt of Government of Indonesia has dramatically reduced over the last ten years. This low level debt gives GoI room to self-finance through debt not only the post-disaster reconstruction activities but also the short-term recovery activities (if debt issuance is fast). See Figure A16.1. Figure A16.1. General gross debt of GoI, as percent of GDP Note: 2011(f): forcast for 2011 Source: IMF (2011) The major natural disasters that occurred in Indonesia did not seem to significantly affect the cost of bor- rowing nor the government’s ability to access the capital markets. As of April 2011, the government bonds issued by Indonesia had a spread of around 200 basis points over US Treasury bonds. Note that the peak observed in 2008 was caused by the financial crisis. See Figure A16.2. Figure A16.2. Emerging market Bond Global Index for Indonesia, spread over US Treasury bonds (basis points) 1,400 EMBIG Indonesia 1,200 1,000 800 600 400 200 0 5/31/2004 7/31/2004 10/1/2004 12/1/2004 2/1/2005 4/1/2005 6/1/2005 8/1/2005 10/1/2005 12/1/2005 2/1/2006 4/1/2006 6/1/2006 8/1/2006 10/1/2006 12/1/2006 2/1/2007 4/1/2007 6/1/2007 8/1/2007 10/1/2007 12/1/2007 2/1/2008 4/1/2008 6/1/2008 8/1/2008 10/1/2008 12/1/2008 2/1/2009 4/1/2009 6/1/2009 8/1/2009 10/1/2009 12/1/2009 2/1/2010 4/1/2010 6/1/2010 8/1/2010 10/1/2010 12/1/2010 2/1/2011 4/1/2011 EMBIG Indonesia Global Facility for Disaster Reduction and Recovery 1818 H Street, NW Washington, DC 20433, USA Telephone: 202-458-0268 E-mail: GFDRR@worldbank.org Facsimile: 202-522-3227 Australia Bangladesh Belgium Brazil Canada China Colombia Denmark Egypt Finland France Germany Haiti India Indonesia Ireland Italy Japan Luxembourg Malawi Malaysia Mexico The Netherlands New Zealand Nigeria Norway Portugal Saudi Arabia Senegal South Africa South Korea Spain Sweden Switzerland Turkey United Kingdom United States Vietnam Yemen Special thanks and appreciation are extended to the partners who support GFDRR’s work to protect livelihood and improve lives: ACP Secretariat, Arab Academy for Science, Technology and Maritime Transport, Australia, Bangladesh, Belgium, Brazil, Canada, China, Colombia, Denmark, Egypt, European Commission, Finland, France, Germany, Haiti, India, Indonesia, International Federation of Red Cross and Red Crescent Societies, Ireland, Islamic Development Bank, Italy, Japan, Luxembourg, Malawi, Malaysia, Mexico, the Netherlands, New Zealand, Nigeria, Norway, Portugal, Saudi Arabia, Senegal, South Africa, South Korea, Spain, Sweden, Switzerland, Turkey, United Kingdom, United Nations Development Programme, United States, UN International Strategy for Disaster Reduction, Vietnam, the World Bank, and Yemen.