Building the Resilience of the Poor in the Face of Natural Disasters

Economic losses from natural disasters totaled 92 billion dollars in 2015. Such statements, all too commonplace, assess the severity of disasters by no other measure than the damage inflicted on buildings, infrastructure, and agricultural production. But 1 dollars in losses does not mean the same thing to a rich person that it does to a poor person; the gravity of a $92 billion loss depends on who experiences it. By focusing on aggregate losses—the traditional approach todisaster risk—we restrict our consideration to how disasters affect those wealthy enough to have assets to lose in the first place, and largely ignore the plight of poor people.This report moves beyond asset and production losses and shifts its attention to how natural disasters affect people’s well-being. Disasters are far greater threats to well-being than traditional estimates suggest. This approach provides a more nuanced view of natural disasters than usual reporting, and a perspective that takes fuller account of poor people’s vulnerabilities. Poor people suffer only a fraction of economic losses caused by disasters, but they bear the brunt of their consequences. Understanding the disproportionate vulnerability of poor people also makes the case for setting new intervention priorities to lessen the impact of natural disasters on the world’s poor, such as expanding financial inclusion, disaster risk and health insurance, social protection and adaptive safety nets, contingent finance and reserve funds, and universal access to early warning systems.Efforts to reduce disaster risk and poverty go hand in hand. Because disasters impoverish so many, disaster risk management is inseparable from poverty reduction policy, and vice versa. As climate change magnifies natural hazards, and because protection infrastructure alone cannot eliminate risk, a more resilient population has never been more critical to breaking the cycle of disaster-induced poverty.

The report was sponsored by the Global Facility for Disaster Reduction and Recovery (GFDRR) and the Climate Change Cross-Cutting Solutions Area of the World Bank, under the leadership of John Roome.
Building the resilience of the poor in the face of natural disasters.

OVERVIEW
E conomic losses from natural disasters totaled $92 billion in 2015, and average annual losses have been estimated at more than $300 billion a year." 1 Policy makers, analysts, and others are used to such statements, which measure the severity of disasters and their socioeconomic impacts using the value of the damages inflicted by disasters on buildings, infrastructure, equipment, and agricultural production.
Although these numbers are useful-they provide information on the trends and costs of disasters-they fail to detail how disasters affect people's well-being.Obviously, $1 in losses does not mean the same thing to a rich person and a poor person, and the severity of a $92 billion loss depends on who experiences it.The same loss affects poor and marginalized people far more because their livelihoods depend on fewer assets, their consumption is closer to subsistence levels, they cannot rely on savings to smooth the impacts, their health and education are at greater risk, and they may need more time to recover and reconstruct.A flood or earthquake can be disastrous for poor people, but have a negligible impact on a country's aggregate wealth or production if it affects people who own almost nothing and have very low incomes.
By focusing on aggregate losses, the traditional approach examines how disasters affect people wealthy enough to have wealth to lose and so does not take into account most poor people.

"
This shortcoming is not just a monitoring issue.When projects to reduce disaster risk are assessed on the basis of the value of damages that can be avoided, analyses favor projects that will protect or support richer areas or people.Imagine two flood protection projects with similar costs.The first would cover a wealthy neighborhood in a capital city.Because of the density of high-value assets, it would avert on average $10 million a year in damages.The second project would target poorer areas in a second-tier city and prevent just $5 million a year in losses.A traditional analysis would unambiguously select the first project.But a $5 million loss may matter more to poor people than a $10 million loss to richer people.If the second project benefits very poor people, it may generate greater benefits for well-being.And because well-being is the ultimate goal of public policy, the second project may be more attractive.
Moreover, not all risk management policies can be assessed using metrics that include only asset and production losses.Policies such as increasing access to financial services and expanding social safety nets make it easier for people to absorb, cope with, and recover from damages caused by natural disasters.Thus such policies can mitigate the impact of natural disasters on well-being even though they have no impact on direct damages from disasters.This report moves beyond asset and production losses and focuses instead on how natural disasters affect people's well-being.Here, natural disaster risk and losses are measured using a metric that can capture their overall effects on poor and nonpoor people, even if the economic losses of poor people are small in absolute terms.This metric can be used in the analysis of disaster risk management projects so that investments improve the well-being of all people and are not systematically driven toward wealthier areas and individuals.And this report proposes and uses a consistent framework to assess traditional approaches to reducing disaster risk (such as building dikes or reinforcing building regulations) and strengthening resilience (such as adopting adaptive social safety nets) to help design consistent risk management policies.
By examining well-being instead of asset losses, this report provides a deeper (and grimmer) view of natural disasters than does the usual reporting-indeed, this view takes better account of poor people's vulnerability.This analysis also identifies opportunities for action and policy priorities at the country level, with three main messages: 1. Efforts to reduce poverty and disaster risks are complementary.Estimates for 89 countries find that if all natural disasters could be prevented next year, the number of people in extreme poverty-those living on less than $1.90 a day-would fall by 26 million.The impact on poverty is large because poor people are exposed to hazards more often, lose more as a share of their wealth when hit, and receive less support from family and friends, financial systems, and governments.In fact, disasters can push people into poverty, and so disaster risk management can be considered a poverty reduction policy.And since poverty reduction policies make people less vulnerable, they can be considered part of the disaster risk management toolbox.

Natural disasters affect well-being more than what traditional estimates
suggest.Poor people suffer only a small share of the economic losses caused by disasters, but they suffer disproportionately.Based on estimates of socioeconomic resilience in 117 countries, and including in the analysis how poverty and lack of capacity to cope with disasters magnify losses in well-being, the effects of floods, wind storms, earthquakes, and tsunamis on well-being are equivalent to a $520 billion drop in consumption-60 percent more than the widely reported asset losses.The design of disaster risk management should, then, not rely only on asset losses.Targeting poorer people with disaster risk reduction interventions-such as dikes and drainage systemswould generate lower gains in avoided asset losses but larger gains in well-being.

3.
Policies that make people more resilient-and so better able to cope with and recover from the consequences of disasters that cannot be avoidedcan save $100 billion a year.Action on risk reduction has a large potential, but not all disasters can be avoided.Expanding financial inclusion, disaster risk and health insurance, social protection and adaptive safety nets, contingent finance and reserve funds, and universal access to early warning systems would also reduce wellbeing losses from natural disasters.If all countries implemented these policies in the proposed "resilience package," the gain in well-being would be equivalent to a $100 billion increase in annual global consumption.

Efforts to reduce poverty and disaster risks are complementary
Natural disasters keep or move people back into poverty and are one reason that eradicating poverty is so difficult.Between 2006 and 2011, 45 percent of poor households in Senegal escaped poverty, but 40 percent of nonpoor households fell into it, leaving the poverty rate almost unchanged.Natural risk contributed to this lack of progress: households affected by a natural disaster were 25 percent more likely to fall in poverty during the period (Dang, Lanjouw, and Swinkels 2014).Among Guatemalan households hit by tropical storm Agatha in 2010, per capita consumption fell 5.5 percent, increasing poverty by 14 percent (Baez et al. 2016).After Ethiopia's 1984-85 famine, it took a decade for most asset-poor households to restore livestock holdings to pre-famine levels (Dercon 2004).
Poor people suffer disproportionately from natural hazards.Natural disasters hit poor people particularly hard for five reasons: Overexposure.Poor people are overexposed to floods in many countries, such as in Panama and Zimbabwe, where they are greater than 50 percent more likely than the average to be flooded.Such overexposure is also true for drought and high temperatures in most countries.More important, poor people are often exposed to frequent, low-intensity events, such as the recurrent floods that affect many cities with insufficient drainage infrastructure.These events do not attract media interest and are poorly documented, but they can have significant cumulative impacts, especially through their effects on health.
Higher vulnerability.People's vulnerability-that is, how much they lose when they are hit-is also a critical determinant of the impacts of natural disasters.When poor people are affected, the share of their wealth lost is two to three times that of the nonpoor, largely because of the nature and vulnerability of their assets and livelihoods.A global analysis suggests that poor people are nearly twice as likely to live in fragile dwellings.
Less ability to cope and recover.The impact of natural disasters on well-being also depends on how well people cope and recover, which depends on the support they receive.Coverage of poor people by social protection is often low.And after they are hit by a shock, poor people receive less postdisaster support than do nonpoor people.For example, in response to the floods and landslides in Nepal in 2011, only 6 percent of the very poor sought government support, compared with almost 90 percent of the well-off (Gentle et al. 2014).

Permanent impacts on education and health. Disasters force poor households
to make choices that have detrimental long-term effects, such as withdrawing a child from school or cutting health care expenses.In such cases, children are often the main victims (Kousky 2016).In Guatemala, Storm Stan increased the probability of child labor by more than 7 percent in areas hit by the storm (Bustelo 2011).
In Ethiopia, children under 3 at the height of the 1984 famine were less likely to eventually complete primary school, leading to income losses of 3 percent (Dercon and Porter 2014).And in Peru, the impacts of the 1970 Ancash earthquake on educational attainment can be detected even for the children of mothers affected at birth, demonstrating that the effects of large disasters can extend even to the next generation (Caruso and Miller 2015).Irreversible effects on education and health can reinforce the intergenerational transmission of poverty.

Effects of risk on saving and investment behavior.
The losses the poor suffer are not the only way in which disasters and natural risks keep them in poverty.Sometimes, the impact exists even before the disaster hits (ODI and GFDRR 2015).For example, smallholders tend to plant low-return, low-risk crops because they cannot afford to lose one year of production in case of bad weather, so their income is reduced even when the weather is good (Cole et al. 2013).And people are less likely to invest in their house or production equipment if these investments are likely to be washed away by a flood.

Natural disasters increase global poverty
Poverty is thus a factor in the vulnerability to disasters.Similarly, disasters are a driver of poverty.Although it remains impossible to quantify the full effect of natural disasters on the number of impoverished, it is possible to assess the short-term impacts of income losses (see Rozenberg and Hallegatte forthcoming).To do so, a counterfactual scenario was built of what people's income would be in developing countries in the absence of natural disasters.This scenario uses surveys of 1.4 million households, which are representative of 1.2 billion households and 4.4 billion people in 89 countries.The analysis concludes that if all disasters could be prevented next year, 26 million fewer people would be in extreme poverty-that is, living on less than $1.90 a day.Although this estimate is subject to large uncertainties and cannot capture all impacts, including those on health, education, and savings, it still shows how severely natural hazards affect poverty.
Vulnerability to natural hazards and disasters can be reduced through development and poverty reduction efforts that enable people to settle in safer places, make their livelihoods and assets less vulnerable, and provide them with the tools and support needed to cope with shocks.Thus policies that help reduce poverty can be considered part of the disaster risk management toolbox.But the connection between poverty and disaster risk goes both ways: disasters make it harder for poor people to escape poverty.
Disaster risk management can thus also be considered a poverty reduction policy.

Natural disasters affect well-being more than most people think
For hazards such as floods, storms, tsunamis, and earthquakes, risk assessment typically focuses on: » Hazard-the probability of an event occurring » Exposure-the population and assets located in an affected area.
» Asset vulnerability-the value lost when an asset is affected by a hazard.These three factors constitute the risk to assets-that is, the average monetary value of the damages that disasters inflict on assets (often measured as replacement or repair value).But the risk to assets is an incomplete metric.
This report extends risk assessment to measure the well-being losses caused by natural disasters (figure O.1).To do so, risk assessment was conducted separately for poor and nonpoor people, defined as the bottom 20 percent and the top 80 percent in terms of consumption in each country.The analysis takes into account the various dimensions of inequality of poor and nonpoor people in the face of disasters and the distribution of losses across individuals.Indeed, losses concentrated on fewer or poorer individuals have a larger impact than the same losses affecting richer people or shared across larger populations.

Socioeconomic resilience
Specifically, the analysis considers the different abilities of poor and nonpoor people to cope with asset losses by modeling the effects of asset losses on income (accounting for capital productivity and diversification of income sources) and consumption (accounting for savings, remittances and social protection, and postdisaster transfers).Consumption losses are translated into well-being losses, taking into account the different impacts of a $1 loss on poor and nonpoor individuals.Well-being loss at the country level depends on the distribution of impacts within the population, but it is expressed as the equivalent loss in national consumption.Thus a finding that a disaster causes $1 million in wellbeing losses means that the impact of a disaster on well-being is equivalent to a $1 million decrease in country consumption, perfectly shared across the population.
Socioeconomic resilience measures an economy's ability to minimize the impact of asset losses on well-being.It can be defined as the ratio of asset losses to well-being losses: If socioeconomic resilience is 50 percent, then well-being losses are twice as large as asset losses-that is, $1 in asset losses from a disaster is equivalent to $2 in consumption losses, perfectly shared across the population.Socioeconomic resilience can be considered a driver of the risk to well-being, along with the three usual drivers of risk assessment: The impacts of natural disasters on well-being are larger than asset losses In all of the 117 countries studied, well-being losses from natural disasters are larger than asset losses (Hallegatte, Bangalore, and Vogt-Schilb, forthcoming).According to the United Nations Global Assessment Report on Disaster Risk Reduction-the so-called GAR (UNISDR 2015)-total asset losses from natural disasters in these countries average $327 billion a year. 2 Because disaster losses are concentrated on a small share of country populations, imperfectly shared, and affect more poor people (who have limited ability to cope with them), this report estimates that well-being losses in these countries are equivalent to consumption losses 60 percent larger than asset losses, or about $520 billion a year.Globally, poor people are disproportionately affected by these losses: people in the bottom 20 percent experience only 11 percent of total asset losses but 47 percent of well-being losses.Thus poor people experience asset losses that are only half of the average but well-being losses that are more than twice as large.
Maps O.1 and O.2 show this report's estimates of socioeconomic resilience and risk to well-being.Risk to well-being decreases with country income (figure O.2b).This decrease is driven mostly by better protection against floods, higher-quality buildings, and widespread early warning systems in wealthier countries, but resilience also matters.
The average global socioeconomic resilience is 62 percent, ranging from 25 percent in Guatemala to 81 percent in Denmark-meaning that $1 in asset losses in Guatemala has the same impact on well-being as a $4 reduction in national

What matters is not only how much benefit a project generates, but also who benefits
To assess the potential benefits of projects that protect populations against hazards, consider two similar interventions.The first would reduce by 5 percent the share of the population exposed to natural hazards, but target only the poorest 20 percent of people in each country.If the entire world implemented this intervention, avoided asset losses would be $7 billion a year-but global gains in well-being would be $40 billion because the intervention would benefit poor and highly vulnerable people.
The second intervention would also reduce the share of the population exposed to natural hazards by 5 percent, but target only the top 80 percent.Because richer people have so many more assets than do the poor, avoided asset losses would be much larger-about $19 billion.But gains in well-being would be smaller-$22 billion.

Note:
The figure shows avoided annual average losses from a 5 percent reduction in exposure, achieved by reducing the exposure of the poorest 20 percent of people, expressed in absolute terms (millions of U.S. dollars per year, adjusted for purchasing power parity) and relative terms (percentage of current average asset and well-being losses).
These results highlight the trade-offs between monetary gains and well-being gains.
If a disaster risk reduction budget is allocated based only on avoided asset losses and monetary benefits, most investments will go to rich areas.Instead, investments in disaster risk management need to balance the need for economic efficiency with the imperative to protect the most vulnerable.Measuring benefits in terms of increased well-being instead of avoided asset losses is a way to do so.
The same approach can also be applied at the subnational level to identify regional priorities within countries.For example, it can help prioritize between two similar risk reduction projects in two different provinces in Vietnam.A project that would prevent $1 million a year in asset losses in Binh Dinh province, which has an estimated resilience of 69 percent, would generate well-being benefits valued at $1.4 million a year ($1 million divided by 69 percent).By contrast, a project that would prevent $1 million a year in asset losses in Kien Giang province, which has estimated resilience of 29 percent, would increase well-being by $3.4 million a year ($1 million divided by 29 percent).Thus the project in Kien Giang would do far more to increase well-being.

Increasing resilience is good economics
Despite efforts to reduce people's exposure to natural hazards or make their assets less vulnerable to hazards, natural risk cannot be cut to zero.Disasters will continue to occur, and they may even become more frequent because of climate change, urbanization, and increasing population densities in coastal areas.Thus it is critical to supplement actions on exposure and vulnerability with improvements in people's ability to cope with unavoidable shocks.Such efforts require a flexible, holistic risk management strategy that uses different tools for different types of disasters and populations (figure O.4).
Revenue diversification.Diversifying revenue and receiving remittances or cash transfers from social programs help households at all income levels cope with small shocks (Bandyopadhyay and Skoufias 2012).People suffer less from a local disaster if some of their income comes from outside the area through government transfers or remittances.

Financial inclusion. Financial inclusion helps poor people save in forms less
vulnerable to natural hazards than in-kind savings like livestock and housing, which diversifies risk.It also enables the poor to access credit, thereby accelerating and improving recovery and reconstruction.But improving poor people's access to formal financial instruments is a long-term challenge in many developing countries and is insufficient for larger shocks.
Market insurance.Market insurance can protect against larger losses, but efforts to provide universal access to insurance face multiple obstacles, including weak institutional and legal capacity, affordability issues, and high transaction costsespecially for poor people.
Adaptive social protection.For poor households-and to cover the largest shocks-easily scalable social safety nets are needed.Although social safety nets always improve resilience, a growing body of evidence reveals that such instruments are even more efficient when their targeting and delivery are flexible enough to transfer resources to disaster victims in a timely fashion.Postdisaster transfers have a benefit-cost ratio above 1.3 in the 117 countries studied.And in 11 countries-Angola, Bolivia, Botswana, Brazil, Central African Republic, Colombia, Honduras, Lesotho, Panama, South Africa, and Zambia-every $1 spent on postdisaster transfers yields well-being benefits of more than $4.

Financial inclusion
(savings, credit) Revenue diversification (social protection, remittances)  2015).This later estimate deviates from observations because the model and data are imperfect and because the average annual losses include the average losses from low-probability, high-impact events that have not occurred and the underreported losses from high-probability, low-impact events such as recurrent floods.All dollar amounts are U.S. dollars unless otherwise indicated.

Adaptive social protection
2. The average annual losses are slightly higher than the ones published in the 2015 GAR report because revised estimates of the stock of capital were used in this analysis.
"Economic losses from natural disasters totaled $92 billion in 2015." Such statements, all too commonplace, assess the severity of disasters by no other measure than the damage inflicted on buildings, infrastructure, and agricultural production.But $1 in losses does not mean the same thing to a rich person that it does to a poor person; the gravity of a $92 billion loss depends on who experiences it.By focusing on aggregate losses-the traditional approach to disaster risk-we restrict our consideration to how disasters a ect those wealthy enough to have assets to lose in the first place, and largely ignore the plight of poor people.
This report moves beyond asset and production losses and shifts its attention to how natural disasters a ect people's well-being.Disasters are far greater threats to well-being than traditional estimates suggest.This approach provides a more nuanced view of natural disasters than usual reporting, and a perspective that takes fuller account of poor people's vulnerabilities.
Poor people su er only a fraction of economic losses caused by disasters, but they bear the brunt of their consequences.
Understanding the disproportionate vulnerability of poor people also makes the case for setting new intervention priorities to lessen the impact of natural disasters on the world's poor, such as expanding financial inclusion, disaster risk and health insurance, social protection and adaptive safety nets, contingent finance and reserve funds, and universal access to early warning systems.
E orts to reduce disaster risk and poverty go hand in hand.Because disasters impoverish so many, disaster risk management is inseparable from poverty reduction policy, and vice versa.
As climate change magnifies natural hazards, and because protection infrastructure alone cannot eliminate risk, a more resilient population has never been more critical to breaking the cycle of disaster-induced poverty.

Figure O. 1 :
Figure O.1:This report moves beyond asset losses to estimate how natural disasters affect well-being to well-beingWhere would such interventions be the most attractive?In absolute terms, reducing the exposure of poor people to disasters would provide the most benefits in large and highrisk countries (figure O.3a).But in relative terms, reducing the exposure of poor people is more efficient in countries in which they have limited social protection and access to finance (figure O.3b).In such countries, resilience is low, magnifying the benefits of lower exposure.In Mali and Niger, for example, reducing exposure to natural disasters by 5 percent could cut asset losses by more than 10 percent and well-being losses by 25 percent-but only if such efforts target poor people.

Figure
Figure O.3: Reducing poor people's exposure to disasters could prevent large losses in well-being and assets a. Absolute terms

2: Socioeconomic resilience tends to increase with income, whereas risk to well-being decreases with income
consumption.Figure O.2a shows that, overall, resilience grows with GDP per capita.

Risk management should include a range of tools for different types of disasters and households
(Clarke and Dercon 2016)07) Fallavier 2016)rograms can be especially effective at preventing humanitarian emergencies and cutting intervention costs(del Ninno, Coll-Black, and Fallavier 2016).In 2015 Kenya's Hunger Safety Net Programme delivered support to more than 100,000 additional households in response to drought, and added a special transfer to 200,000 households in anticipation of expected droughts.In Ethiopia, rural farmers affected by drought in 2005 and 2011 and covered by the Productive Safety Net Programme had consumption losses 25 percent lower than those of other rural farmers(White and Porter 2016).These types of adaptive social protection programs create liabilities for governments, which may require them to draw on various tools such as reserve funds (for smaller disasters), contingency credit lines (such as World Bank's Cat-DDOs), regional risk pools (such as the Caribbean Catastrophe Risk Insurance Facility), or transfers of part of the risk to global reinsurance or global capital markets (such as with FONDEN bonds in Mexico)(Mahul and Ghesquiere 2007).Such tools make it possible for governments to support the affected population, and they improve the transparency and predictability of the postdisaster response(Clarke and Dercon 2016).Meanwhile, combined with institutional preparedness and contingent plans, they can accelerate recovery and reconstruction, reducing overall losses (de Janvry, del Valle, and Sadoulet 2016).
Hallegatte et al. 2016l.2016.Note: Instruments in blue target households; instruments in green protect governments' or local authorities' budgets.Quick

US$, millions per year
Note: U.S. dollars are measured in terms of purchasing power parity (PPP).Poor people are defined as the poorest 20 percent in terms of consumption in the country.The estimate for the 2015 losses is from Swiss Re.The estimate of average annual losses is from the United Nations Global Assessment Report on Disaster Risk Reduction (UNISDR Source: World Bank estimates.NOTES1.