80584 EAP DRM KnowledgeNotes Working Paper Series No. 27 Disaster Risk Management in East Asia and the Pacific COST BENEFIT STUDIES ON DISASTER RISK REDUCTION IN DEVELOPING COUNTRIES Shyam KC1 The focus of development actors working in the area of disaster management has shifted substantially from disaster recovery to disaster risk reduction over the past decade, coinciding with the decade of the Hyogo Framework for Action (HFA) 2005-20152. Amidst this strategic shift, there is now the need to work towards ensuring that investments made to reduce disaster risks are cost-effective and that the benefits reach all mem- bers of the population including the poor and vulnera- ble, who are often “affected disproportionately” (Global Assessment Report 2009, The Sendai Report 2012). The losses from natural disasters to mankind are undoubtedly massive—on average, globally every year over 100,000 people were killed and some 246 million people affected by natural disasters during the period 2002-2011 and the estimated average economic loss was US$131 billion per year (EMDAT 2013). The purpose of this note is to briefly survey existing evidence in developing countries with regard to the ben- efits and costs of various disaster risk reduction interventions so as to provide some general lessons for disaster risk reduction (DRR) practitioners on the strengths and limitations of such existing work. In doing so, this note examines evidence on the economics of DRR in developing countries3. The note begins by providing a compara- 1 Disaster Risk Management Specialist, East Asia and Infrastructure Unit, World Bank. 2 A Plan adopted by 168 nations in 2005 to “detail the work that is required from all different sectors and actors to reduce disaster losses” (UNISDR). 3 Such as those available in Department of International Development (DFID)’s desk review of costs and benefits of Natural Disaster and Disaster Risk Reduction Measures (Environmental Resources Management 2005); GTZ Manual on Cost-Benefit Analysis of Natural Disaster Risk Management in Developing Countries (Mechler 2005); Guidance Note No. 8 prepared by ProVention on Economic Analysis, Tools for Mainstreaming Disaster Risk Reduction (Benson, Twigg and Rossetto 2006); the Information Note No. 3 by The Global Platform for Disaster Risk Reduction (GPDRR/ISDR) on Costs and Benefits of Disaster Risk Reduction (Moench, Mechler and Stapleton 2007); and the WB and UN (2011) joint publication Natural Hazards, UnNatural Disasters: Economics of Effective Pre- vention. This working paper series is produced by the East Asia and Pacific Disaster Risk Management Team of the World Bank, with support from the Global Facility for Disaster Reduction and Recovery (GFDRR). The series is meant to provide just-in-time good practice examples and lessons learned from projects and programs related to aspects of disaster risk management. 2 Disaster Risk Management in East Asia and the Pacific tive guideline for analysis. This is followed by a sum- duction measure? If more than one option has been mary diagnostic of seventeen case studies4 along five analyzed, what are they? key dimensions comprising the guideline as follows: ■■ What are the data used and how is the analysis (1) Metric and methodology, (2) Sources of uncer- carried out? How are the risks estimated? Do the tainty, (3) Measuring fatalities and injuries, (4) Results analyses look at limited past data to project future obtained and, and (5) Disaggregated impacts. In the losses? Or do they have sufficient established haz- concluding section that follows, the note discusses the ard and vulnerability data base to estimate risk? overall trends in the field of performing cost and benefit analysis of DRR measures and offers some recommen- ■■ What are the discount rates adopted to estimate the dations for ways forward. present values of future benefits and costs? 2. Sources of uncertainty: What are the sources of COMPARATIVE GUIDELINE uncertainty? How have the issues of uncertainty been handled? The literature on assessing the socio-economics of di- saster risk reduction (DRR) in developing countries is 3. Measuring fatalities and injuries: Have the fa- an evolving one, and is limited compared to evidence talities and injuries been considered in measuring from developed economies. Detailed economic analy- losses? If so how have they been evaluated? ses are few (Benson, Twigg and Rossetto 2006), are not 4. Results obtained: What results have been ob- comprehensive (Environment Resources Management tained? Do the results suggest cost-effectiveness of 2005) and the use of tools such as cost-benefit analysis the DRR measure explored? (CBA) has declined over time (WB and UN, 2011). The studies examined here differ greatly in their inclu- 5. Disaggregated analysis: To what extent are the sion of detail, making it difficult to compare and con- distributional aspects of disaster risk reduction ben- trast them along exact dimensions. In order to make a efits and costs analyzed? In other words have the useful comparison, we have identified five broad dimen- cost and benefits been disaggregated for different sions and key questions for which we have some infor- socio- economic groups such as on the poor, women mation for most of the studies. These are: and children? 1. Methodology and Metrics: ANALYSIS ■■ What is the unit of analysis for the study? For ex- ample, are the costs and benefits calculated at the This section summarizes the case-studies along the structure, project, community, district or a national above dimensions. Table 1, Table 2 and Table 3 pro- level? vide the summary. Of the seventeen disaster cases that ■■ What is the major economic analysis? For example were examined, eight are from Asia and Pacific (Fiji, In- what does it compute? Benefit-Cost Ratio, Cost- dia, Indonesia, Philippines, Taipei City), four are from effectiveness Ratio, Net Present Value, Internal the Caribbean (Dominica, Jamaica), two from East- Rate of Return, etc.? Or is it a qualitative study ern Europe (Romania, Turkey) and three from South that does not quantify the benefits and costs at all? America (Argentina, Bogota City/Colombia, Peru). Hazard wise these studies deal with floods (Argentina, ■■ Is more than one option for disaster risk reduction Fiji, India, Indonesia, Peru, Philippines and Romania), explored? Or is the study focused on a single risk re- hurricane/typhoon/cyclone (Dominica, India, Jamaica, Vietnam), and earthquakes (Bogota City Istanbul, and Taipei City). 4 The selection of the 16 studies does not represent a scientific sample, and are a simple compilation of CBA studies. Addi- tional 3 studies have been included that discuss disaggregated impacts. 1. Methodology and Metrics Cost Benefit Studies on Disaster Risk Reduction in Developing Countries 3 Units of analysis study (Thomalla and Schmuck 2004 for India) employs Seven of the studies take structures as the units on purely qualitative methods. which to report the cost-benefit metric. Thus Pereira (1995) reports the incremental cost of mitigation mea- Metrics sure for different types of buildings in Jamaica while Benefit-Cost Ratio (also Cost-Benefit Ratio in some Ghesquiere et al (2006) present the analysis for dif- cases), Internal Rate of Return, Net Present Value5 are ferent structure types (schools, hospitals, fire stations, the most commonly computed economic metrics across other public buildings) in Bogota, Colombia. Smyth the quantitative case studies. Four of the studies provide et al (2004a) analyze for various retrofitting measures the incremental cost and incremental benefits of the of buildings in Turkey. Hochrainer-Stigler et al (2010) proposed mitigation measures (Pereira 1995 for Jamai- make the analysis for replacing buildings to reduce ca; Kay and Wilderspin 2004 for Vietnam; Vermeiren flood damage in Uttar Pradesh, India, and for improv- et al, 2004 for Jamaica and Dominca) while one esti- ing flood resilience in buildings vs. elevating property mates metrics such as Pure Risk Premium and Probable in Jakarta. Vermeiren et al (2004) estimate ex-post Maximum Loss (Ghesquiere et al 2006 for Colombia). the incremental cost of mitigation which would have avoided the damage faced by a Seaport in Jamaica and Mechler (2005b for Indonesia), Smyth et al (2004a), by a School in Dominica. Two of the studies, flood Ghesquiere et al (2006) for Colombia, The World Bank protection projects in Argentina (World Bank 1996) for Argentina (1996) and Romania (2004), Hochrain- and Romania (World Bank 2004), analyze the benefits er-Stigler et al (2010) for India and Indonesia) generate and costs at the project level. loss exceedance curves based on past hazard and vulner- ability data to estimate losses associated with varying Venton and Venton (2004) analyze the disaster miti- disaster probabilities. Mechler (2005a) considers the gation and prevention programs in Bihar and Andhra approach taken in his Peru study “back-ward looking” Pradesh, India at the village level. Hung and Chen because it assesses limited past disaster impacts to get a (2007) make their analysis at an administrative unit lev- rougher understanding of flood protection in Peru. This el for Shihlin district in Taipei City. Dedeurwaerdere approach is “less rigorous” and “less data-intensive” (1998) analyzes the benefit and costs for protecting the compared to the approach taken in his Indonesian study City of Angels in Philippines against floods and lahars which is “forward looking, risk based” (Mechler 2005b) (mud/debris flow), just as Mechler (2005b) analyzes and provides a more rigorous framework because it the benefits and costs of protecting the Indonesian city combines information on historic data on hazard and of Semerang from floods, Holland (2008) looks at the vulnerability to estimate risk and the risk reduced. De- benefit and costs of a flood early warning system in Nu- deurwaerdere (1998), on the other hand, draws a prob- ava of Fiji. Mechler (2005a) makes the analysis at the ability tree-based on past hazard and vulnerability data river basin level in Piura, Peru and Kay and Wilderspin to estimate the avoided losses from the disaster. Hung (2002) analyze the benefits and costs plantation wide in and Chen (2007) use available land use maps, surveys the coastal areas of Vietnam. and an earthquake loss estimation system (HAZ-Tai- wan) to make an earthquake “risk-benefit” analysis. Quantitative vs. Qualitative Sixteen of the studies employ quantitative methods to demonstrate the benefits of the respective mitigation measures carried out. Of these, three studies (Dedeur- Simply put, NPV gives an estimate of the net benefit of 5 waerdere, 1998 for Philippines, Holland 2008 for Fiji a proposal; the internal rate of return describes the dis- and Venton and Venton, 2004 for India) first assess count rate at which the present value of costs equals the benefits in a qualitative manner using interviews, focus present value of benefits; and CBR provides the relative group discussions and archives. As such they employ a size of the costs compared to benefits. Users should be mixture of qualitative and quantitative techniques. One aware of the specific context of the analysis in interpret- ing these indicators for decision making. 4 Disaster Risk Management in East Asia and the Pacific Options analyzed discount rate. For five of the remaining quantitative Typical to the cost-benefit analysis methodology, sev- studies the discount rate is either not available or not eral of the studies analyze the “with” and “without” sce- employed. nario which means that they examine the loss associ- ated with the disaster with the mitigation measure, and 2. Sources of Uncertainty without the mitigation measure. Some studies discuss a single mitigation option while some analyze multiple Lack of hazard and vulnerability data, difficulty in eval- options for DRR. For example, Pereira (1995) inves- uating indirect losses, and dealing with climate change tigates ex-post, three mitigation options—bearing wall impacts are cited as major sources of uncertainty in systems, building frame systems and moment resisting performing the economic analysis for disaster risk re- frame systems—that could have potentially prevented duction measures. Historic data on hazard (magni- the Gilbert (1988) scale hurricane damage on Jamaican tude, frequency and duration) and vulnerability are not buildings. On the other hand, Ghesquiere et al (2006) known in most of the studies which makes it difficult examine the benefits of retrofitting structures against to perform cost-benefit analysis of the DRR measures not retrofitting in Colombia for an earthquake vulner- on hazards of different scale. In this context, sensitivity ability reduction program. Smyth et al (2004a) ana- analysis assists in examining the reliability of the com- lyze the benefit and cost of various retrofitting options puted metrics. Eight of the cases carry out sensitivity (original, braced, partial and full shear wall). Hung and analysis of the results by varying one or more of the Chen (2007) in their application of seismic risk-benefit following: (i) the discount rate (ii) the benefits (iii) the analysis are more focused on examining benefits and costs (iv) project start date (v) duration/frequency of costs associated with three hypothetical earthquakes of hazard, (vi) life of a structure or (vii) a combination of varying magnitudes and frequencies for different kind the above. Sensitivity analysis is not available in the of land uses. Holland (2008) examines the “with” and other quantitative studies. “without” scenario of having an early warning system In addition, the cases examined fare poorly when it for floods. comes to accounting for indirect losses resulting from Dedeurwaerdere (1998) examines ex-ante, three op- a disaster, such as losses in terms of business interrup- tions for protecting a Philippino city from the disaster tions, reduction in crop yields, reduction in revenue risks of floods/ lahars, namely through watershed resto- from reduced tourist inflow, losses to livelihoods, losses ration by rain forestation, river channel improvements due to service closures, etc. Finally, what are rarely ad- and bamboo plantation. On the other hand The World dressed in the examined cases are the climate change Bank (2004) explores flood defense projects; large de- impacts and the uncertainties thereof. Particularly in fense projects and small dam safety projects designed to infrastructure design, climate change-related uncertain- prevent flood disasters in Romania. ty can increase “the probability of either under-adapta- tion or over-adaptation” (Hallegatte, 2006). Discount Rates The use of discount rates is aimed at bringing the future 3. Valuation of fatalities and injuries benefits and future costs to the present value. Discount One of the most contentious issues related to perform- rates ranging from 3 to 20% have been used to con- ing economic analysis of disaster risk reduction mea- vert future values to the present value. Dedeurwaerdere sures is the value assigned to deaths and injuries. Values (1998) uses a “discretionary” figure of 20%. Leaving running in the millions of dollars have been assigned aside this outlier, all other discount rates fall between in studies carried out in the developed country context 5% and 12%. Venton and Venton (2004) use 10% for (for example FEMA assigns US$ 3 million per fatal- India, based on local lending rates from moneylender ity) while some studies do not take into account fatali- and banks. All four of the Mechler (2005a, 2005b) and ties and injuries at all. Three of the studies examined the World Bank (1996, 2004) studies use 12% as the accounted for the fatalities and injuries explicitly. For Cost Benefit Studies on Disaster Risk Reduction in Developing Countries 5 Colombia, Ghesquiere et al (2006) use US$ 500,000 nica in 1979, whereas only an additional 1.9% would per fatality while for India, Venton and Venton (2004) have been sufficient to mitigate the losses incurred had computed the cost using the average daily wage rate— it been incorporated into initial reconstruction. Also about Rs. 35 (Approx. US$ 0.6 in May 2013 exchange 40.7 % of the original construction cost was spent in rate) per day. For Peru, Mechler (2005a) used 150,000 reconstructing a school compared to only an additional sols (Approx. US$ 57,000 in May 2013 exchange rate) 11.5% that would have been sufficient for mitigating per fatality. Three of the studies do mention fatality the 1988 Gilbert scale hurricane (Vermeiren, Stichter losses but the value assigned is not available. The re- and Wason 2004). Smyth et al (2004a) consider four maining nine, including the qualitative studies, do not damage levels associated with damage to an apartment take into account fatalities. In benefit-cost analyses by building in Turkey in the event of an earthquake. Smyth et al. (2004a, 2004b) of earthquake strengthen- In Taipei City, higher losses in industrial, education and ing measures for apartment buildings and schools, none commercial type land uses and lower losses in agricul- of the strengthening measures considered passes the tural type land use were predicted by Hung and Chen benefit-cost test unless the value of lives saved is includ- (2007). Expected average annualized earthquake losses ed in the analysis. In the analysis for Istanbul, Turkey were estimated to decrease by approximately NT$ 1 M two of the strengthening measures pass the benefit cost (approx. US$ 30,000 in May 2013 values) per year. In test when avoided deaths are valued at US$1 million Bogota city, Pure Risk Premium was expected to decline each. In the case of the school building strengthening from US$ 7.4 per million to US$ 1.5 per million when program, strengthening measures pass the benefit-cost the buildings are retrofitted (Ghesquiere et al 2006). test when lives saved are each valued at US $400,000. Similarly Probable Maximum Loss for a 1 in 1000 year Studies such as by Cropper and Sahin (2009) sug- earthquake for a retrofitted school would be 4% of the gest methods of estimating the value of mortality risk asset value compared to 30% without the retrofits. reductions (i.e., of the Value of a Statistical Life or An Internal Rate of Return (IRR) was estimated to be VSL) and the value of avoided injuries associated with between 12% and 79% for the mitigation measures for disasters in performing DRR CBA analysis for devel- various subprojects proposed in the Argentina Flood oping countries where often reduced death and injury Protection Project carried out by the World Bank are not monetized. They provide a literature survey on (1996). The overall project internal rate of return falls published VSL values for developing countries and a from 20.4% to 7.5% if the project is initiated 5 years framework for judging whether these values are ap- later making a case not to delay the project. Dedeur- propriate for CBA. waerdere (1998) computed a benefit cost ratio of 30 for rain forestation farming, 14.7 for bamboo plantation 4. Results and 3.5 for river channel improvement for protecting the City of Angeles in the Philippines. Despite the All of the studies demonstrate higher benefits of disaster high BCR, the forestation project had the lowest Net risk reduction measures compared to the costs incurred. Present Value. For example, the incremental cost of including earth- quake resistant features in building design in Jamaica In Vietnam, planning and protection of 12,000 hect- would have been less than 3% compared to the losses ares of mangroves cost US$ 1.1 million compared to that amounted to 2% to 100% (Pereira 1995) of the the reduction of dyke maintenance by US$ 7.3 mil- building costs. Such cost would have been a small 1% lion per year. Deaths reduced to zero and livelihoods of the building costs compared to a benefit of 35-40% if 7750 families were positively affected. In Romania, for mitigating hurricane disaster of the Gilbert scale in an IRR between 13% and 31% and a BCR between Jamaica (Pereira 1995) had the DRR measures been in- 1.5 and 4.4 for dam sub-projects, and an IRR between tegrated in the initial building design. Similarly 4.2% 13.2% and 42.3% and a BCR between 1.1 and 2.1 for of the original construction cost was spent on recon- the flood defense sub-projects were computed for the structing the Seaport from Hurricane David in Domi- Hazard Risk and Emergency Preparedness Project 6 Disaster Risk Management in East Asia and the Pacific (World Bank 2004). The overall project had an IRR studies examined explicitly disaggregate data by various of 26% and a BCR of 2.2. Venton and Venton (2004) socio-economic groups (by gender, caste-ethnicity, race, computed for Bihar, India a BCR of 3.8 and a NPV of able and disabled, etc.) to examine the differential ben- Rs. 3.7 Million (Approx US$ 100,000) and in Andhra efits and costs that DRR measures can have upon dif- Pradesh a BCR of 13.4 for the respective disaster miti- ferent groups. Thus, it is not known the extent to which gation and preparedness interventions in the selected the marginalized/vulnerable groups have been or will be villages. Similarly for the Rohini river basin in Uttar served by the proposed mitigation measures. Pradesh, Hochrainer-Stigler et al (2010) find that with Although not directly CBAs, there are researches which the commonly used discount rate of 12%, only two of explore how different groups are impacted by disasters, the six measures against flooding appear to have high which allows us to speculate who would benefit the economic returns (building new mud or brick homes on most from DRR measures. For example, Pradhan et al a plinth); and that it does not pay to build a new house (2007) examine the risk of mortality due to floods that in brick instead of mud if the only benefit is reducing occurred in 1993 in Southern Nepal. Their examination flood. of over 41,000 children and adults revealed more flood- In Peru, a BCR of 3.8 and an IRR of 1% was computed related deaths amongst girls and women compared to for flood protection (Mechler 2005a) while in Indo- boys and men. They also found flood related-deaths nesia, a BCR 2.5 and an IRR of 23% was computed were associated with lower socio-economic status and against floods and tidal inundations (Mechler 2005b). having a thatched roof. Similarly Hochrainer-Stigler et al (2010) find that the Similarly in a qualitative study conducted after the 1998 results show benefit-cost ratios substantially higher floods, Rashid and Michaud (2000) found adolescent among mixed-wall structures than among masonry females in Bangladesh, already at a disadvantage due ones as flood protection measures, while elevating the to cultural restrictions, to be even more vulnerable and property by 1m also has mostly favorable results, with at a greater disadvantage in a post- flood environment. the benefit cost ratios ranging up to 6.7. In Fiji, the Lack of latrines and private places translated to them early warning system would have a benefit cost ratio of not being able to remain “secluded” as is expected by so- 3.7 to 7.3 at a 10% discount rate for 1 flooding event of cietal norms, which in turn added to their stress. Also, a 1 in 20 year return period (Holland 2008). Winchester’s (2000) statistical analysis for South India shows that having an established income and asset base 5. Distributional analysis are important variables in reducing one’s vulnerability A few of the studies make reference to vulnerable to disasters. Such income and asset base enables them groups in their analysis. For example, although briefly, to earn “social and economic credit worthiness” in the the Romania analysis (World Bank 2004) makes ref- local economy. erence to how the vulnerable population are expected to benefit from the proposed projects. Some 124,000 DISCUSSION AND CONCLUSION of the 453,364 beneficiaries of flood defense areas were classified as poor and estimated to represent 90% This survey of case studies on benefits and costs of of the most vulnerable groups in Romania. The Ven- DRR reveals three trends. First, there is now endorse- ton and Venton (2004) study looks at the impact upon ment that investing in DRR is equally—if not more— the vulnerable in general and also in relation to how necessary than post-disaster operations. In all the cases they can be effectively evacuated. The options used by analyzed, DRR measures provided higher benefits than Hochrainer-Stigler et al (2010) for the flood protection costs incurred which reinforces the message that gov- measure in the Rohini Basin of Uttar Pradesh, India ernments, donors and development agencies need to considers both mud houses and brick houses, thereby intensify their efforts towards, and increase investments making the cost-benefit analysis inclusive of the poor in risk reduction. The consensus is that there are high (who generally have mud houses). However none of the economic and social returns of DRR actions—both by Cost Benefit Studies on Disaster Risk Reduction in Developing Countries 7 DRR projects on their own or when integrated with de- vulnerability is a necessary first step to perform a sound velopment projects. It must be noted here that despite economic analysis. the strategic shift towards ex-ante risk reduction, it is The question that arises next, however, is whether more estimated that only about 3.5% of total international data collection would address the “deep uncertainties” development assistance was allocated for disaster-relat- related to highly complex phenomena such as climate ed activities during the period 1991-2010, and of that change, mega-disasters, uncontrolled urbanization, only about an eighth on disaster prevention and pre- changes in land-use patterns, and unprecedented de- paredness (GFDRR and ODA, 2013). cline/growth in population. Uncertainty then poses Second, there is some degree of ambivalence about the a bigger challenge to performing CBAs, only part of appropriateness of the commonly-employed cost-bene- which can be resolved by improving data. Hallegatte fit analysis (CBA) method to analyze costs and benefits and Przyluski (2010) argue that uncertainties in cost- of DRR. That is, the CBA is prone to “critical limi- ing disasters come both from insufficient data and inad- tations” (Moench, Mechler and Stapleton 2007), the equate methodologies. While some uncertainties arise true benefit-cost ratio is never known (Dedeurwaerdere from our inability to fully assess indirect losses, others 1998), and the results appear dependent on parameters arise from differences in values and preferences (such on which there is no consensus (Hallegate et al 2012). as in the use of discount rates, valuation of human life) The key limitations are: difficulty in monetizing the in- which can be measured in a consensual way. Cropper tangible benefits; the need to make too many assump- and Sahin (2009) note that death and injury are often tions regarding hazard and vulnerability; lack of histori- not monetized for developing country context, and cal hazard data to predict loss in a probabilistic manner; suggest methods for estimating the value of mortality discretionary discounting of future costs to present val- risk reductions (i.e., of the Value of a Statistical Life or ues, etc. VSL) and the value of avoided injuries associated with disasters in performing DRR CBA in such context. The cost-benefit analyses presented are based on real cost savings or monetary benefits, and are found to be The limitations in terms of monetizing the intangible inadequate when it comes to assessing benefits of an benefits, the need to make assumptions regarding haz- intangible nature. Many of the studies have avoided ard and vulnerability, lack of historical hazard data to counting for the reduced fatalities and indirect benefits predict loss in a probabilistic manner, and the need which are social in nature. This limits the true cost- to account for uncertainty in making sound DRR in- benefit analysis as the results obtained does not take vestment decisions are well reflected in the case stud- into account the full range of costs and benefits associ- ies examined here. Despite these limitations, the CBA ated with the DRR activities. tool remains powerful (Dedeurwaerdere 1998, Mechler 2005, Venton and Venton 2004), provides good evi- The wide variation found in the methodologies, dence for analyzing the benefits and costs of DRR and assumptions, discount rates and sensitivity analysis sug- are useful when “issues are complex and there are sever- gest that economic analysis of DRR measures is highly al competing proposals” (World Bank and UN, 2011). context sensitive. Following Mechler’s (2005) classifi- Indeed, irrespective of the limitations, the CBA is more cation, many of the cases examined adopt a “backward useful as a process in itself than outcome (Hallegate et looking, impact-based” approach. In the absence of re- al 2012; Kull et al 2013) where stakeholders, if enabled, liable historical hazard and vulnerability data, it is dif- can participate in sharing information and opinion, ob- ficult to perform cost-benefit analysis in a probabilistic serving what constitutes benefits or costs and how the manner or in a “forward looking, risk-based manner”. results are achieved. Most of the cost-benefit studies acknowledge that data on hazard and vulnerability is limited. This is a big Given the limitations, the analysis suggests that there challenge faced by the DRR community in conduct- is a need to develop clear protocols and guidelines as ing comprehensive economic studies of proposed DRR to what constitutes reliable CBA of DRR interven- measures. Investing in data collection on hazard and tions, taking into account the development in the 8 Disaster Risk Management in East Asia and the Pacific field thus far. Such guidelines should make recom- REFERENCES mendations on executing each step of the CBA— Benson, Charlotte, John Twigg and Tiziana Rossetto. 2007. option identification, stakeholder identification and “Economic Analysis Guidance Note No. 8” in Tools consultation, costs and benefit estimates over a pe- for Mainstreaming Disaster Risk Reduction. Geneva: riod of time, choice of discount rate, computing net ProVention Consortium. present value of project options, performing sensi- Centre for Research on the Epidemiology of Disasters. 2013. tivity analysis, and offering recommendations—as it CRED Crunch 31 Newsletter, Issue No 31. applies to DRR interventions. Related to this, alter- Cropper, Maureene L. and Sahin, Sebnem. 2009. Valuing native ways of weighing costs and benefits need to be Mortality and Morbidity in the Context of Disaster Risks, explored and used to handle “deep uncertainties”. One The World Bank, Development Research Group, Sus- tainable Rural and Urban Development Team. such approach is the Robust Decision Making (RDM) methodology which was developed by RAND Corpo- Dedeurwaerdere, A. 1998. Cost-Benefit Analysis for Natural Disaster Management: A Case- study in the Philippines. ration over the last decade and has been employed in Universite Catholique De Louvain: Centre for Research areas of water management, renewable energy and flood on the Epidemiology of Disaster. risk management, amongst others. The RDM meth- Donner, William R. 2007. “The Political Ecology of Disas- odology - rather than predicting a best-estimate future- ter: An Analysis of Factors Influencing U.S. Tornado Fa- iterates models to “determine how plans perform in a talities and Injuries. Demography 44(3): 669-685. range of plausible futures”, and thereby help decision Environmental Resources Management. 2005. Natutral makers “identify conditions under which their plans Disaster and Disaster Risk Reduction Measures. London: will perform well or poorly” (RAND 2013). DFID. GFDRR and ODA. 2013. International Financing for Di- Finally, most of the analyses we examined do not look saster Risk Management: The 20 Year Story (1991-2010). into the distributional aspects of DRR costs and ben- Ghesquiere, Francis, Luis Jamin and Olivier Mahul. 2006. efits. That is, to what extent DRR interventions benefit Earthquake Vulnerability Reduction Program in Colom- differentially the socially and economically vulnerable— bia: A Probabilistic Cost-benefit Analysis. Washington such as the poor, women, children, the old, disabled and D.C.: The World Bank. people of disadvantaged caste/ethnicity—needs more Ghesquiere, Francis, Prashant, Robert Reid, Jan Kellett, research. Conventional CBA focus on overall returns Shyam KC and Jack Campbell. 2012. The Sendai Re- port: Managing Disaster Risks for a Resilient Future. The and not how costs and benefits are distributed (Kull et World Bank. al 2013) and whether CBAs can appropriately capture Hallegatte, Stéphane. 2006. A Cost-Benefit Analysis of distributional impacts is a contested issue, in particular the New Orleans Flood Protection System. Joint Cen- if different weights can be attributed to different group ter AEI-Brookings Joint Center for Regulatory Studies. of people to account for distributional impacts. There Regulatory Analysis. are differing views on this topic, for example, Joahns- Hallegatte, Stephane & Przyluski, Valentin, 2010. The Eco- son-Stenman (2005) argues that it is difficult to defend nomics of Natural Disasters: Concepts and Methods. both the proposition that distributional concerns should Policy Research Working Paper Series 5507. The World always be used in CBA, and that they should never be Bank. used. Increased research in this direction would en- Hallegatte, S., Ankur Shah, Robert Lempert, Casey Brown hance not only our ability to understand who benefits and Stuart Gill. 2012. Investment Decision Making Un- der Deep Uncertainty: Application to Climate Change. from DRR measures and how resources can be used to Policy Research Working paper 6193. The World Bank. benefit those most vulnerable to disasters, but also what Hochrainer-Stigler, S., Howard Kunreuther, Joanne Lin- critical roles DRR interventions can play in reducing nerooth-Bayer, Reinhard Mechler, Erwann Michel-Ker- poverty, and prioritizing measures to be taken. jan, Robert Muir-Wood, Nicola Ranger, Pantea Vaziri, and Michael Young. “The Costs and Benefits of Reduc- ing Risk from Natural Hazards to Residential Structures in Developing Countries.” 2010. Risk Management and Cost Benefit Studies on Disaster Risk Reduction in Developing Countries 9 Decision Processes Center, The Wharton School, Univer- Rand Corporation, Research highlights. “Making Good De- sity of Pennsylvania. Working Paper # 2010-12-01. cisions Without Predictions: Robust Decision Making Holland, Paula. 2008. Fiji Technical Report, An Economic for Planning Under Deep Uncertainty” http://www.rand. Analysis of Flood Warning in Fiji. EU EDF 8 – SOPAC org/content/dam/rand/pubs/research_briefs/RB9700/ Project Report 122 Reducing Vulnerability of Pacific ACP RB9701/RAND_RB9701.pdf States. Rashid Sabina F. and Stephanie Michaud. “Female Ado- Hung, Hung-Chih and Liang-Chun Chen. 2007. “The Ap- lescents and Their Sexuality: Notions of Honor, Shame, plication of Seismic Risk-Benefit Analysis to Land Use Purity & Pollution during the Floods.” Disasters 24(1): Planning in Taipei City.” Disasters 31(3): 256-276. 54- 70. Kay, Rohan and Ian Wilderspin. 2002. “Mangrove Planting Smyth, Andrew W., M. Eeri, Gulay Altay, George Deodatis, Saves Lives and Money in Vietnam”. P. 95 in World Di- Mustafa Erdik, Guillermo Franco, Polat Gulkan, How- saster Report Focus on Reducing Risk. Geneva: Interna- ard Kunreuther, Hilmi Lus, Esra Mete, Nano Seeber, tional Federation of Red Cross and Red Crescent Societ- and Ozal Yuzugullu. 2004a. “Probabilistic Benefit-Cost ies (IFRCRCS). Analysis for Earthquake Damage Mitigation: Evaluating Measures for Apartment Houses in Turkey.” Earthquake Kull, Daniel, Reinhard Mechler and Stefan Hochrainer-Sti- Spectra 20(1): 171-203. gler. 2013. Probabilistic Cost–Benefit Analysis of Disaster Risk Management in a Development Context. Disasters Smyth, Andrew W., G. Deodatis, G. Franco, Y. He, and 37(3): 357-553. T, Gurvich. 2004b. “Evaluating Earthquake Retrofitting Measures for Schools: A Demonstration Cost-Benefit Mechler, Reinhard. 2003. “Natural Disaster Risk and Cost- Analysis. “Columbia University, New York. Benefit Analysis.” Pp 45-55 in Building Safer Cities: The Future of Disaster Risk, edited by A. Kreimer, M. Arnold The World Bank. 1996. Staff Appraisal Report Argentina and A. Carlin. The World Bank: 2003. Flood Protection Project. Washington D.C.: The World Bank. ___________. 2005a. “Case Study Piura, Peru.” Pp. 45-61 in Cost Benefit Analysis of Natural Disaster Risk Manage- The World Bank. 2004. Project Appraisal Document to ment in Developing Countries Manual. GTZ. Government of Romania for a Hazard Risk Mitigation and Emergency Preparedness Project. Washington, D.C.: ___________. 2005b. “Case Study Semerang, Indonesia.” Pp. The World Bank. 62-76 in Cost Benefit Analysis of Natural Disaster Risk Management in Developing Countries Manual. GTZ. The World Bank and United Nations. 2011. Natural Haz- ards, UnNatural Disasters: Economics of Effective Pre- Moench, Marcus, Reinhard Mechler and Sarah Stapleton. vention. 2007. Information Note No. 3: Costs and Benefits of Di- saster Risk Reduction. Geneva: United Nations Interna- Thomalla and Schmuck. 2004. ““We All Knew that a Cy- tional Strategy for Disaster Reduction. clone was Coming”: Disaster Preparedness and the Cy- clone of 1999 in Orissa, India.” Disasters 28(4): 373-387. Multihazard Mitigation Council (MMC). 1995. Hazard Mitigation Saves: An Independent Study to Assess the United Nations. 2009. Global Assessment Report on Di- Future Savings from Mitigation Activities Volume I: saster Risk Reduction. Risk and Poverty in a Changing Findings, Conclusions, and Recommendations. Washing- Climate. ton, D.C., National Institute of Building Sciences (NIBS). Venton, Courtney Cabot and Paul Venton. 2004. Disaster Olof Johansson-Stenman. 2005. Distributional Weights in Preparedness Programmes in India: A Cost Benefit Anal- Cost-Benefit Analysis—Should We Forget about Them? ysis. London: Humanitarian Practice Network at ODI. Land Economics 81 (3): 337–352. Vermeiren, Jan, Steven Stichter and Alwyn Wason. 2004. Pereira, John. 1995. Costs and Benefits of Disaster Mitigation Costs and Benefits of Hazard Mitigation for Building in the Construction Industry. Caribbean Disaster Miti- and Infrastructure Development: A Cass Study in Small gation Project, available at http://www.preventionweb. Island Developing States. Caribbean Disaster Mitigation net/files/1177_CDMPCostsandBenefits.pdf Acessed on Project, available at http://www.oas.org/CDMP/docu- May 12, 2013. ment/papers/tiems.htm Acessed on May 12, 2013. Pradhan, Elizabeth K., Keith P. West, Joanne Katz, Steven Winchester, Peter. 2000. “Cyclone Mitigation, Resource Al- C. Le Clerq, Subarna K. Khatry and Sharada R. Shrestha. location and Post Disaster Reconstruction in South India: 2007. “Risk of Flood-related Mortality in Nepal.” Disas- Lessons from Two Decades of Research.” Disasters 24 (1): ters 28(4): 373-387. 18-37. 10 Disaster Risk Management in East Asia and the Pacific Table 1: Economic Studies of Disaster Risk Reduction SN Title of Research Year Disaster type Country 1 Costs and Benefits of Disaster in the Construction Industry (Pereira) 1995 Hurricane/Earthquake Jamaica 2 Staff Appraisal Report Argentina Flood Protection Project (LAC, The World 1996 Floods Argentina Bank) 3 Application of the Selected Cost-Benefit Model to Natural Disaster 1998 Floods/Lahar Philippines Management Case-Study of Floods/Lahars in the Pampanga Region (Dedeurwaerdere) 4 Mangrove Planting Saves Life and Money in Viet Nam 2002 Typhoon (Wukong) Vietnam (Kay and Wilderspin /IFRCRCS) 5 Costs and Benefits of Hazard Mitigation for Building and Infrastructure 2004 Hurricane (David), 1979 Dominica Development: A Case Study in Small Island Developing States (Vermeiren, Stichter and Wason) 6 Costs and Benefits of Hazard Mitigation for Building and Infrastructure 2004 Hurricane (Gilbert), Jamaica Development: A Case Study in Small Island Developing States 1988 (Vermeiren, Stichter and Wason) 7 Project Appraisal Document for a Hazard Risk Mitigation and Emergency 2004 Floods Romania Preparedness Project: Economic Analysis of Flood Risk Reduction Investments (ECA, The World Bank) 8 Disaster Preparedness Programs in India-A cost Benefit Analysis (Venton 2004 Floods (Bihar); India and Venton) Floods/droughts (AP) 9 “We All Knew that a Cyclone was Coming”: Disaster Preparedness and 2004 Cyclones India the Cyclone of 1999 in Orissa, India (Thomalla and Schmuck) 10 Probabilistic Benefit-Cost Analysis for Earthquake Damage Mitigation: 2004 Earthquake Turkey Evaluating Measures for Apartment Houses in Turkey. (Smyth et al) 11 Case Study Piura, Peru, Cost-benefit Analysis of Natural Disaster Risk 2005 Floods Peru Management in Developing Countries (Mechler) 12 Case Study Semarang, Indonesia, Cost-benefit Analysis of Natural Disaster 2005 Floods and tidal Indonesia Risk Management in Developing Countries (Mechler) inundation 13 Earthquake Vulnerability Reduction Program in Colombia: A Probabilistic 2006 Earthquake Colombia Cost Benefit Analysis (Ghesquiere, Jamin and Mahul) 14 The Application of Seismic Risk-Benefit Analysis to Land Use Planning in 2007 Earthquake Taiwan Taipei City (Hung and Chen) 15 An Economic Analysis of Flood Warning in Navua, Fiji (Holand) 2010 Flood Fiji 16 The Costs and Benefits of Reducing Risk from Natural Hazards to Residential 2010 Flood Indonesia Structures in Developing Countries (Hochrainer-Stigler et al): Flood Risk in Jakarta 17 The Costs and Benefits of Reducing Risk from Natural Hazards to Residential 2010 Flood India Structures in Developing Countries (Hochrainer-Stigler et al): Flood Risk within the Rohini River Basin in Uttar Pradesh Cost Benefit Studies on Disaster Risk Reduction in Developing Countries 11 Table 2/1: Economic Study of Disaster Risk Reduction-Metrics, Options, Units, Discount Rate and Fatality Reporting Fatality/ Unit of Discount injury SN Research Year Metric computed Options Analysis Rate Reporting 1 Costs and Benefits 1995 Incremental cost For Earthquake: Bearing wall Structure N/A Not of Disaster in the for mitigation systems/Building frame considered Construction Industry measures at systems/ Moment Resisting (Pereira) the design/ frame systems. For Hurricane: construction stage Alusteel or galvalume of the building. sheeting/ Zinc sheeting 2 Staff Appraisal Report 1996 (Economic) Benefits computed under Project 10% Not Argentina Flood Internal conditions with and without considered Protection Project (LAC, Rate of Return, the implementation of the The World Bank) Net Present Value subproject. 3 Application of the 1998 Benefit Cost Watershed restoration (rain City 20% Not Selected Cost- Ratio, Net forestation) considered Benefit Model to Present River channel improvements Natural Disaster Value, Cost- Bamboo Plantation Management Case- effectiveness Study of Floods/ analysis Lahars in the Pampanga Region (Dedeurwaerdere) 4 Mangrove Planting 2002 Cost Single option of Mongrove Plantation N/A N/A Saves Life and Effectiveness forests Money in Viet Nam discussed. (Kay and Wilderspin/ IFRCRCS) 5 Costs and Benefits 2004 Incremental cost Single option of cost of Structure Price Not of Hazard Mitigation restoration/reconstruction. (Sea Port) inflation considered for Building and The question asked: What per year Infrastructure mitigation measure would 7.9% Development: A Case have been required during the Study in Small Island design and construction of Developing States each project, to avoid losses (Vermeiren, Stichter and from the particular extreme Wason) event that affected the projects? 6 Costs and Benefits 2004 Incremental cost Single option of cost of Structure Inflation Not of Hazard Mitigation restoration/reconstruction. (School) 7.9%/year considered for Building and The question asked: What Infrastructure mitigation measure would Development: A Case have been required during the Study in Small Island design and construction of Developing States each project, to avoid losses (Vermeiren, Stichter and from the particular extreme Wason) event that affected the projects? 12 Disaster Risk Management in East Asia and the Pacific Table 2/2: Economic Study of Disaster Risk Reduction-Metrics, Options, Units, Discount Rate and Fatality Reporting Fatality/ Unit of Discount injury SN Research Year Metric computed Options Analysis Rate Reporting 7 Project Appraisal 2004 Internal Rate of Flood defense projects/ Project N/A Fatality rate Document for Return, Benefit Large dam safety projects/ of 1% of a Hazard Risk Cost Ratio Small dam safety projects; affected Mitigation and For Flood defense projects, population Emergency do nothing alternative/ 5% expected, Preparedness probability protection/ 1% but value not Project in Romania: probability protection; estimated Economic For dam safety do Analysis of Flood nothing/one-stage Risk Reduction rehab/two stage rehab/ Investments (ECA, complete replacement The World Bank) options analyzed. 8 Disaster 2004 Benefit Cost What would have been the Community 10% Rs 35 per Preparedness Ratio, Net Present impact of the hazard on the day (daily Programs in India-A Value, Cost- community before the DMP average cost Benefit Analysis effectiveness intervention had taken place? wage rate) (Venton and Venton) analysis What is the impact now that the DMP has taken place? 9 “We All Knew that a 2004 Qualitative Shelters, early warning Community N/A N/A Cyclone was Coming”: Disaster Preparedness and the Cyclone of 1999 in Orissa, India (Thomalla and Schmuck) 10 Probabilistic Benefit- 2004 Probabilistic No retrofit, braced retrofit, Building 3% $1 to $4 Cost Analysis for Analysis partial shear wall retrofit, full million Earthquake Damage shear wall retrofit Mitigation: Evaluating Measures for Apartment Houses in Turkey. (Smyth et al) 11 Case Study Piura, 2005 Net Present Value, Installing an artificial retention River basin 12% 150000 Sols Peru (Cost-benefit Benefit Cost Ratio, system encircled by a dam in per fatality; Analysis of Natural Internal Rate of the upstream area (Polder): health Disaster Risk Return Case with and without a injuries not Management in polder examined; Project life considered Developing Countries) of 30 years. (Other options, (Mechler) namely elevating the existing dikes and creating an exit for Rio Piura to the sea also proposed.) Cost Benefit Studies on Disaster Risk Reduction in Developing Countries 13 Table 2/3: Economic Study of Disaster Risk Reduction-Metrics, Options, Units, Discount Rate and Fatality Reporting Fatality/ Unit of Discount injury SN Research Year Metric computed Options Analysis Rate Reporting 12 Case Study 2004 Net Present Value, Integrated management of City 12% N/A Semarang, Indonesia Benefit Cost Ratio, flooding and water supply; (Cost-benefit Analysis Internal Rate of project life of 50 years of Natural Disaster Return assumed. (Other options, Risk Management in namely dam protecting Developing Countries) harbor and installation of (Mechler) more drainage pumps also proposed.) 13 Earthquake 2006 Catastrophe Risk Retrofitting structures versus Structures 12% US$ 500000 Vulnerability Modeling: not retrofitting (Buildings, per fatality Reduction Program Probabilistic BCR, fire stations, in Colombia: A Loss Exccedance hospitals) Probabilistic Cost Curve Benefit Analysis (Ghesquiere, Jamin and Mahul) 14 The Application of 2007 Risk-benefit ratio Benefits and costs associated Li (basic N/A Four level Seismic Risk- calculated to with three hypothetical unit of admin. (severity 1 Benefit Analysis to express earthquakes of different in Taipei City) for minor Land Use Planning in the level of seismic magnitudes, frequency and injuries to Taipei City (Hung and risk attached to locations. severity 4 Chen) different land use for instantly plans. killed) 15 An Economic Analysis 2008 Benefit Cost Ratio Single option Early 10% Not of Flood Warning in Warning considered Navua, Fiji (Holand) system 16 The Costs and 2010 Benefit Cost Ratio (1) Improve flood resilience Structure 5-12% Not Benefits of Reducing and resistance of the (building) considered Risk from Natural property. Hazards to Residential (2) Elevate the property by 1 Structures in meter. Developing Countries (Hochrainer-Stigler et al) :Flood Risk in Jakarta 17 The Costs and 2010 Benefit Cost Ratio Six options of demolishing Structure 5-12% Not Benefits of Reducing mud house/brick house and (building) considered Risk from Natural replace it with a mud house/ Hazards to Residential brick house built on a raised Structures in plinth/ replacement in year Developing Countries 1/end of lifetime. (Hochrainer-Stigler et al): Flood Risk within the Rohini River Basin in Uttar Pradesh 14 Disaster Risk Management in East Asia and the Pacific Table 3/1: Economic Study of Disaster Risk Reduction-Sensitivity, Data Source, Analysis, Results, Strengths and Limitations Sensitivity Data source and SN Title of Research Year analysis analysis Results Major strength Remarks 1 Costs and Benefits 1995 N/A Costs for changes The incremental Provides Indirect losses of Disaster in in roof cost for evidence in not taken into the Construction detailing including favor of making account. Industry (Pereira) understood based earthquake buildings on the impacts resistant features earthquake of Hurricane in to the building and hurricane Gilbert of 1988, design usually resistant. and the effect of costs less than the specifications 3% compared to on the costs 2-100% losses. determined; Costs Such cost is 1% of earthquake and benefits 35- resistant cost 40% for hurricane determined. mitigation. 2 Staff Appraisal 1996 Discount rate Historical flood Subproject IERRs May provide Loss calculation Report Argentina 8-12%; Cost data used for between 12% useful guidelines does not include Flood Protection increased by flood risk analysis; and 79%; When for similar analysis. losses due to Project (LAC, contingency Actual damage project delayed, employment, The World Bank) amount; Benefits costs calculated NPV fatality. increased by in probabilistic estimated at the indirect terms for various $510M, $478M benefit amount; water levels and $ 449M for Project delayed through modeling a discount rate by 5 years; and using GIS; of 8%, 10% and and Additional Only direct 12% respectively; analysis using costs (in terms IERR falls to 7.5% a mathematical of rehabilitation when project model. costs) computed, delayed by 5 indirect costs years--Provides not considered; a case for not Project life of 50 to postpone the years; project. 3 Application of 1998 Not carried out Probability BCR Rain Develops a Assumes the Selected Cost- based, modified forestation simple probability availability of Benefit Model to Asian Social farming: 30 tree for hazard composite hazard Natural Disaster Institute Model; Bamboo occurrence and and vulnerability Management Data on natural Plantation: 14.7 the losses avoided data;Analysis Case-Study disasters, damage River channel based on historic hindered of Floods/ estimates and improvement: 3.5 data. because of lack Lahars in the current indicators NPV higher for of data and/ Pampanga Region collected through forestation. or ambiguous/ (Dedeurwaerdere) key development inconsistent data. agencies. Uses a probability tree. Loss Exceedance curve Cost Benefit Studies on Disaster Risk Reduction in Developing Countries 15 Table 3/2: Economic Study of Disaster Risk Reduction-Sensitivity, Data Source, Analysis, Results, Strengths and Limitations Sensitivity Data source and SN Title of Research Year analysis analysis Results Major strength Remarks 4 Mangrove 2002 N/A N/A Planning and Planting Saves protection of Life and Money 12000 hectares of in Viet Nam (Kay mangroves has cost and Wilderspin/ US$ 1.1 million IFRCRCS) compared to reduce the maintenance by US$ 7.3 million per year; No lives lost; Submerged, coastal forests act as buffers against the sea, high devastating waves broken to harmless ripples. 5 Costs and 2004 N/A Damage assessment 4.2% of the original Provides evidence Indirect Benefits of carried out post construction for benefits of losses Hazard Mitigation David; designs cost was spent making structures not taken for Building and completed for repair on reconstruction, resistant to into account Infrastructure and reconstruction whereas only hurricanes; Shows Development: work necessary to additional 1.9% importance of A Case Study make port functional; would have analyzing correct in Small Island Mitigation costs been sufficient and full hazard Developing estimated for to mitigate the data during States (Vermeiren, additional structural losses incurred; design. Stichter and and nonstructural Failure was due Wason) elements to resist to use of incorrect “David- force” winds. or inadequate hazard information during design, and pressure on designers to maintain lowest possible costs. 6 Costs and 2004 N/A Damage assessment 40.7 % of Provides evidence Indirect Benefits of carried out post the original for benefits of losses Hazard Mitigation David; designs construction making structures not taken for Building and completed for repair cost spent on resistant to into account Infrastructure and reconstruction reconstruction, hurricanes; Shows Development: work necessary to but only additional importance of A Case Study make port functional; 11.5% would have analyzing correct in Small Island Mitigation costs been sufficient to and full hazard Developing estimated for mitigate the losses data during States (Vermeiren, additional structural incurred design. Stichter and and nonstructural Wason) elements to resist “David- force” winds. 16 Disaster Risk Management in East Asia and the Pacific Table 3/3: Economic Study of Disaster Risk Reduction-Sensitivity, Data Source, Analysis, Results, Strengths and Limitations Sensitivity Data source and SN Title of Research Year analysis analysis Results Major strength Remarks 7 Project Appraisal 2004 Project cost Flood damage curves IRR between 16% Makes a brief Loss Document for increased by 20%; determined by detailed and 31% for analysis by poor calculation a Hazard Risk Benefits reduced field surveys; Indirect dam projects; IRR and non- poor does not Mitigation and by 20% benefits estimated between 18% people; Different include Emergency as 50% of direct and42% and a probability losses due to Preparedness benefits; Incremental BCR between alternatives employment, Project in economic costs of the 1.1and 1.5; Chosen analyzed. fatality Romania: project derived from subprojects Economic the financial costs by represent good Analysis of Flood adjusting for taxes, economic returns. Risk Reduction credits and interests Investments (ECA, during construction, The World Bank) and shadow pricing forex. These costs include environmental costs and other costs such as land acquisition. 8 Disaster 2004 Discount rate FGD using Provides evidence Most Lessons Preparedness varied between Participatory for benefits of vulnerable learnt: Programs in 5% and 15%. Disaster Risk making structures identified to Need for India-A cost Drought and flood Assessment resistant to ensure effective robust data, Benefit Analysis period per year Methodologies in hurricanes; Shows evacuation; need for (Venton and also varied. non-DMP and DMP importance of Mixture of baseline data, Venton) villages; Triangulation analyzing correct qualitative and weakness of with village records; and full hazard data quantitative; FGD because local experienced during design. Beneficiaries it raises NGO’s help; Natural, involved. expectations. physical, human, social and economic impacts assessed at the micro/ internal level. 9 “We All Knew 2004 N/A Rapid appraisals, Community based Provides tool Need for that a Cyclone was interviews, strategies such to make rapid verifying Coming”: Disaster information from media as Disaster Risk appraisals of the benefits Preparedness and Committees, benefits and and costs the Cyclone of information costs in the quantitatively. 1999 in Orissa, dissemination, absence of India (Thomalla shelters effective quantitative data. and Schmuck) for improving disaster preparedness. Cost Benefit Studies on Disaster Risk Reduction in Developing Countries 17 Table 3/4: Economic Study of Disaster Risk Reduction-Sensitivity, Data Source, Analysis, Results, Strengths and Limitations Sensitivity Data source and SN Title of Research Year analysis analysis Results Major strength Remarks 10 Probabilistic 2004 Uses 3% discount Fragility curves None of the Importance of Benefit-Cost rate but shows established analytically strengthening counting for Analysis for how to determine measures fatalities is well Earthquake max discount considered passes demonstrated. Damage rate for which the benefit-cost Mitigation: mitigation will test unless the Evaluating be most cost- value of lives saved Measures for effective is included in the Apartment Houses analysis. in Turkey. (Smyth et al) 11 Case Study Piura, 2005 Discount rate Past data at two time Base case: NPV Detailed Hazard Peru (Cost- benefit varied to 0% and points on damages, 268M Soles, stepwise case recurrence Analysis of Natural 20%;Costs vulnerability and BCR 3.8, IRR 31%; study for Cost based on Disaster Risk increased ad hazard used to assess Project remains Benefit Analysis two time Management hoc to 30%; Not risk; Risks determined economically points only; in Developing taking account of with and without viable in all other Additional Countries) loss of life; Not mitigation options; scenarios. benefits (eg (Mechler) taking account Project life of 30 years irrigation) not of indirect assumed. quantified; effects; Not Damages for taking account larger region of increases in scaled down hazard exposure. for the study. 12 Case Study 2005 Not taking A risk based “forward- Base case: NPV Detailed Need for Semarang, account looking” approach 414 Billion, BCR stepwise case verifying Indonesia (Cost- increased in building on a detailed 2.5 and IRR 23%; study for Cost the benefits benefit Analysis of hazard exposure; assessment of hazard Project is viable in Benefit Analysis and costs Natural Disaster Not taking and vulnerability. all other scenarios. quantitatively. Risk Management account of Assessment based in Developing increase of the on work by BGR/ Countries) subsidence GTZ, JICA, direct and (Mechler) problem; Not indirect damages taking account of assessed, project life both the above of 50 years assumed. (i.e., benefits constant over the life time) 18 Disaster Risk Management in East Asia and the Pacific Table 3/5: Economic Study of Disaster Risk Reduction-Sensitivity, Data Source, Analysis, Results, Strengths and Limitations Sensitivity Data source and SN Title of Research Year analysis analysis Results Major strength Remarks 13 Earthquake 2006 By reducing Makes a probability Pure risk premium Expands the The model is Vulnerability benefits cost benefit analysis. expected to standard cost- sophisticated Reduction Program reduced by up Computes Pure Risk decline from US$ benefit analysis and requires in Colombia: A to 40%. Premium, Probable 7.4 per million to to capture the hazard and Probabilistic Cost Maximum Loss, Loss US$ 1.5 per million uncertainty disaster data. Benefit Analysis Exceedance Curve. when the building is related to the (Ghesquiere, retrofitted; Probable catastrophic Jamin and Maximum Loss event while Mahul) for a 1 in 1000 conducting year earthquake cost-benefit for a retrofitted analysis. school would be 4% of the asset value compared to 30% without the retrofits. Shows that proposed investments are viable. 14 The Application 2007 N/A HAZ Taiwan Higher losses in Assist city Haz-Taiwan of Seismic Risk- earthquake loss industrial, education planners model is based Benefit Analysis to estimation system; and commercial to evaluate on census areas Land Use Planning Land use maps type land uses appropriateness and assumes in Taipei City and surveys; Higher loss avoided of their planning homogenous (Hung and Chen) Cluster analysis in agricultural type decisions and soil, demography for vulnerability land use; Expected to steer urban and building- distribution; OLS average annualized growth; Uses risk type conditions regressions. earthquake losses based approach throughout is are estimated to for decision each Li. decrease by approx. making in land- NT $ 1 M per year. use planning. 15 An Economic 2008 Discount rate Losses sustained in 3.7-7.3 (Most likely Analysis of Flood (3 to 10%) and the 2004 flood are scenario) Warning in Navua, flood duration calculated and used to Fiji (1 in 10 and estimate the potential 1 in 20 years); benefits from a flood worst, most warning system. likely and best response (to the warning) Cost Benefit Studies on Disaster Risk Reduction in Developing Countries 19 Table 3/6: Economic Study of Disaster Risk Reduction-Sensitivity, Data Source, Analysis, Results, Strengths and Limitations Sensitivity Data source and SN Title of Research Year analysis analysis Results Major strength Remarks 16 The Costs and 2010 Discount rate Analysis based on The results Addresses 60% Community- Benefits of (5to 12%) approximate flood show BC ratios of Jakarta’s level protection Reducing Risk and life of the extent maps and substantially higher structures. such as flood from Natural structure (10 limited depth estimates among mixed wall defenses Hazards to and 25 years). for two past floods in structures than or improved Residential January/February among masonry drainage Structures in 2002 and February ones. Elevating systems not Developing 2007. Hazard analysis the property by 1m. considered. Countries also uses a 30-year also has mostly Fatality and (Hochrainer-Stigler monthly rainfall time favorable results, assets not et al) :Flood Risk in series, observed at the with B/C ratios included in Jakarta Jakarta Observatory. ranging from 0.61 counting losses. to 6.73. 17 The Costs and 2010 Discount rate Analysis based on The results show Considers Fatality and Benefits of (5 to 12%) historical flood data that in all but different socio- assets not Reducing Risk and life of the and on a survey of one case it is not economic included in from Natural structure (10 losses from floods; advisable from level when counting losses. Hazards to and 25 years). data on housing losses a benefit-cost considering Residential from past floods, perspective to options (i.e., mud Structures in namely major floods demolish homes vs. brick house); Developing occurring in 1998 and and rebuild them on Case extended Countries 2007. a plinth. by including (Hochrainer-Stigler climate change; et al): Flood Risk within the Rohini River Basin in Uttar Pradesh ENDNOTE 1 The author would like to thank Sebnem Sahin, Alejandro de La Fuente and Nidhi Kalra (peer reviewers), Margaret Arnold, Rachel Cipryk, Stephane Hallegatte, Daniel Kull and Asmita Tiwari for their comments and feedback on the Note. East Asia and the Pacific Region The World Bank 1818 H St. NW, Washington, D.C., 20433 http://www.worldbank.org/eap Special thanks and appreciation are extended to the partners who support GFDRR’s work to protect livelihood and improve lives: ACP Secretariat, Australia, Austria, Belgium, Brazil, Canada, Chile, 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, Korea, Republic of, Luxembourg, Malaysia, Mexico, Morocco, the Netherlands, New Zealand, Nigeria, Norway, Portugal, Saudi Arabia, Senegal, Solomon Islands, South Africa, Spain, Sweden, Switzerland, Togo, United Kingdom, United Nations Development Programme, United States, UN International Strategy for Disaster Reduction, and the World Bank.