economic crises and natural disasters: coping strategies and policy implications

Economic Crises and Natural Disasters:

Coping Strategies and Policy Implications

EMMANUEL SKOUFIAS *

Inter-American Development Bank, Washington, DC, USA

Summary. — This paper reviews 12 studies presented at a conference examining two broad themes:(a) the interplay between household coping strategies and the impact of crises and natural disasterson various dimensions of well-being (e.g., consumption and child nutrition); and (b) some of the ex-ante and ex-post strategies that public agencies can adopt so they can be more effective in protectinghouseholds and their members from the potentially adverse impacts of aggregate shocks.� 2003 Elsevier Science Ltd. All rights reserved.

1. INTRODUCTION

Economic crises and natural disasters havebeen a recurrent phenomenon in the developingworld. Recent examples of these type of ag-gregate shocks abound: the 1995 and 2002 fi-nancial crisis in Argentina, the 1994–95 peso(or tequila) crisis in Mexico, the effects of ElNi~nno on Philippines and Indonesia, and the1998 currency crises in East Asia.

In addition, the frequency and severity ofthese events seems to have been increasing infrequency and severity (e.g., Von Braun, Vlek,& Wimmer, 2002). The Center for Research onthe Epidemiology of Disasters (CRED), col-lecting data on the number of natural disasters,reports a clear increase in the number of naturaldisasters reported probably due to changes inglobal climate. In the Latin American and Ca-ribbean region alone, during 1980–99 there were38 major droughts, floods, hurricanes, tropicalstorms, landslides, earthquakes, volcano erup-tions and El Ni~nno episodes (IADB, 2000). Thesame general trend seems to be present foreconomic crises. In spite of government effortsto reform and improve their efficiency, eco-nomic crises arising more frequently due to ex-ternal rather than internal factors continue tostall the efforts of most developing countries toachieve stable GDP growth. For example, dur-ing 1980–98, there have been over 40 episodeswhere GDP per capita fell by 4% or more inLatin America and the Caribbean (IADB,2000). Such adverse shocks can lead to verysharp increases in poverty at the country level asevidenced by the numbers presented for selectedcountries in Table 1.

There is little doubt that the poverty ratesreported in Table 1 above provide a one-dimensional view of how household welfare isaffected as a result of these adverse economicevents. In an effort to generate and disseminateknowledge on economy-wide shocks a confer-ence was organized in November 2001 by theInternational Food Policy Research Institute(IFPRI) and the Inter-American DevelopmentBank (IADB) with support from the WorldBank and the United States Agency for Inter-national Development (USAID). The confer-ence titled Crises and Disasters: Measurementand Mitigation of their Human Costs broughttogether a variety of papers examining two

World Development Vol. 31, No. 7, pp. 1087–1102, 2003� 2003 Elsevier Science Ltd. All rights reserved

Printed in Great Britain0305-750X/03/$ - see front matter

doi:10.1016/S0305-750X(03)00069-Xwww.elsevier.com/locate/worlddev

*The author would like to thank Cesar Bouillion,

Lawrence Haddad, Carole Levin, Nora Lustig, and

Michael Walton for their help and support throughout

this project, Lourdes Hinayon for her excellent admin-

istrative and secretarial support and the various col-

leagues who served as discussants at the November 2001

conference on: ‘‘Crises and Disasters: Measurement and

Mitigation of their Human Costs’’ where these papers

were first presented. The conference was organized by

the Inter-American Development Bank (IADB) and the

International Food Policy Research Institute (IFPRI)

with support from the World Bank and the United

States Agency for International Development (USAID).

The list of discussants at the conference includes: H.

Alderman, J. Behrman, W. Cunningham, S. Duryea, P.

Fallon, M. Ferrantino, A. Foster, M. Goldstein, E.

Gustaffson-Wright, D. King, A. Kochar, R.E.B. Lucas,

D. Mazumdar, T. Marchione, L. Pritchett, and G.

Sedlacek.

1087

broad themes. The first one relates to the in-terplay among the ex-ante (mitigating) and ex-post (coping) strategies of households and theimpact of crises and natural disasters on vari-ous dimensions of well-being (e.g., consump-tion and child nutrition). The second coverssome of the ex-ante and ex-post strategies thatpublic agencies can adopt so they can be moreeffective in protecting households and theirmembers from the potentially adverse impactsof economy-wide shocks and natural disasters.

A particular concern with the nature anddesign of government responses at times ofcrises is how programs can play effectively thedual role of providing immediate and effectiverelief to the households affected by the crisiswhile at the same time contributing to povertyalleviation in the long run (e.g., Lustig, 2000;Holzmann & Jorgensen, 2001). Poorer house-holds are typically less equipped to deal withshocks, and informal insurance arrangementsare likely to have serious limitations especiallyfor shocks that are common to all members ofthe formal or informal insurance group. In theabsence of an effective public safety net systempoorer households may use coping strategiesthat ultimately prevent households from everescaping from poverty or from reaping thebenefits of future economic growth. In order tosurvive during a time of crisis, for example,poor households may be forced to sell theirproductive assets, such as draft animals. To theextent that households have limited access tocredit and other sources of financing there is astrong possibility that these households may

never be able to replenish their stocks of pro-ductive assets, thus remaining in poverty per-manently or for years after the passing of thecrisis. In addition, during a period of crisis,households may be forced to decrease theirinvestments in the human capital of their chil-dren through their lowered ability to provideproper nutrition and health care for their chil-dren or by reallocating the time children de-voted in school toward earning income.

One unfortunate consequence of these cop-ing strategies is that they may transmit povertyfrom the current generation to the future gen-eration. Insofar as countries manage to developsocial safety nets that effectively insure poorhouseholds from the most adverse effects ofeconomic crises and natural disasters, thencrisis management programs may also be con-sidered as contributing significantly to povertyalleviation and the economic development of acountry.

The papers included in this special issue ofWorld Development are directly related to theabove issues. A number of the papers analyzerigorously the effects of recent macroeconomicshocks such as the peso crisis in Mexico, nat-ural disasters such as El Ni~nno in Philippines,the recent adjustment process in Jamaica,and the effects of AIDS in Africa, amongothers. The list of the main questions addressedby the conference papers is both diverse andlong. Here is a list of the questions that may beof interest to most readers: Do communitycharacteristics play a role in how a shock affectshouseholds? Does a rapid increase in the in-flation rate affect the nutritional status of chil-dren? Do households have a more difficult timecoping with covariant as opposed to idiosyn-cratic shocks? What are some of the strategiesthat households use in the case of aggregateshock? Does social capital make a difference incoping with shocks? Domacroeconomic shocks,even if short-lived, have adverse effects in thelong run? What types of programs and gov-ernment actions are required to prevent ormitigate the adverse long-term effects of somehousehold coping strategies? Is a shift of gov-ernment interventions to an ex-ante risk-reduction strategy to be preferred over ex-postmitigation and coping programs? Is it possibleto design a targeting system that serves the dualroles of alleviating poverty and insurancesimultaneously? Are cash transfer programslikely to be less effective at maintaining caloricavailability during periods of higher infla-tion than in normal times? Are cash transfer

Table 1. Economic crises and poverty in selected coun-tries (headcount ratios)

Country Before crisis Year of crisis

Argentina 16.8 24.8

(1993) (1995)

Indonesia 11.3 18.9

(1996) (1998)

Korea 2.6 7.3

(1997) (1998)

Malaysia 8.2 10.4

(1997) (1998)

M�eexico 36

(1994) (1995)

Thailand 9.8 12.9

(1997) (1998)

Venezuela 41.4 53.6

(1993) (1994)

Source: IADB (2000), and World Bank (2001a).

WORLD DEVELOPMENT1088

programs also effective at maintaining mi-cronutrient availability?

Section 2 is devoted to summarizing the im-pact of economic crises and natural disastersand the various strategies households adopt inorder to cope with these aggregate shocks.Section 3 is devoted to the second theme ofthe conference concerning the public actionsthat can minimize the impact of and exposureto such shocks. Section 4 concludes with asummary of the policy lessons that can ex-tracted from the papers presented at the con-ference.

2. IMPACT OF CRISES AND NATURALDISASTERS AND HOUSEHOLD COPING

STRATEGIES

In order to obtain a better perspective on thefindings of the papers presented at the confer-ence it is useful to go over briefly how economiccrises and natural disasters may affect house-hold welfare. Natural disasters such as of floods,droughts, earthquakes, and other weather-related phenomena can affect household welfarethrough the destruction of physical and humancapital stock. In contrast, economic crises canaffect household welfare through a variety ofadditional channels: 1 (a) a slowdown in eco-nomic activity that usually translates to a de-crease in demand for labor services, a decreasein the probability of finding new employment,an increase in the unemployment rate and adecrease in the level of earnings of individualsalready employed (e.g., see Fallon & Lucas,2002); 2 (b) changes in relative prices or the re-moval of price subsidies for staple foods such asrice or wheat. A devaluation of the local cur-rency, for example, is likely to affect the relativeprice of tradable commodities. Along relatedlines, an increase in the price of staple foods islikely to lead to a large reduction in the pur-chasing power (real income) of landless ruraland urban poor households who are typicallynet consumers of food and spend a large shareof the budget on these items; (c) cutbacks in thelevel of public transfers (e.g., Ravallion, 2002);and (d) changes in the value of and returns toassets.

In addition to all the different ways economiccrises and natural disasters can affect house-holds, one must also take into consideration thevariety of risk management arrangements andrisk management strategies that are availablefor protection (see, for example, World Bank,

2001a). Risk management arrangements in-clude all the informal, and formal (market-based and public) arrangements that haveevolved for the purposes of providing someprotection for households in the event of acrisis. Risk management strategies, on the otherhand, typically cover all the prevention andmitigation strategies that households may im-plement prior to the crisis event, as well as allthe coping arrangements that are availableto households after the realization of a crisis(Alderman & Paxson, 1994). Table 2 adaptedfrom World Bank (2001a, 2001b) and fromHolzmann and Jorgensen (2001) provides amore detailed exposition of the mechanisms formanagement of risk from crises and naturaldisasters. 3 For the purposes of designing safetynet programs as well as for quickly imple-menting public interventions during times ofcrises, it is critical to know the main copingstrategies households rely upon at times ofcrisis.

A distinguishing feature of economic crisesand natural disasters from other types ofshocks experienced by households is that theyaffect many households simultaneously. Theaggregate nature of these shocks means thatmany of the informal mechanisms for mitigat-ing and coping with risk become ineffective.This is particularly the case for the informalmechanisms that are group-based (Morduch,1999). For example, when aggregate demandand employment falls, being a member of anoccupational association can be of little helpwhen most of the members of the same occu-pation are also affected negatively by the shock.Along similar lines, natural disasters affectingwhole villages or even regions may put signifi-cant strains on local group-based insuranceschemes that in more normal circumstancesmay be quite effective in providing some in-surance. In fact, aggregate shocks may alsostrain market-based coping mechanisms suchas borrowing from formal financial institutions.For example, rural financial institutions, themajority of whose deposits are from individualsengaged in agricultural activities, are likely tobe unable to serve the role of a lending insti-tution when the majority of the deposits arewithdrawn during a period of a harvest failureor a flood (Binswanger & Rosenzweig, 1986).

Considering that safety net programs areeither weak or malfunctioning in many devel-oping countries, it is of particular interest toknowwhat are the effects of these adverse eventson various dimensions of welfare. Simple as that

ECONOMIC CRISES AND NATURAL DISASTERS 1089

Table 2. Mechanisms for managing risks

Strategies for Arrangements using

Informal mechanisms Formal mechanisms

Individual and household-

based

Group-based Market based Publicly provided

Reducing risk Preventive health practices Collective action for infrastructure, dikes,

terraces

Sound macroeconomic policy

Migration Common property resource management Environmental policy

More secure income sources Education and training policy

Public health policy

Infrastructure (dams, roads)

Active labor market policies

Mitigating risk

Diversification Crop and plot diversification Occupational associations Savings accounts in fi-

nancial institutions

Agricultural extension

Rotating savings and credit associations Microfinance Liberalized trade

Income source diversification Protection of property rights

Investment in physical and

human capital

Insurance Marriage and extended family Investment in social capital (networks,

associations, rituals, reciprocal gift giving)

Old age annuities Pension systems

Sharecropper tenancy Accident, disability, and

other insurance

Mandated insurance for

unemployment, illness,

disability, and other risks

Buffer stocks

Coping with

shocks

Sale of assets Transfers from networks of mutual support Sale of financial assets Social assistance

Loans from moneylenders Loans from financial

institutions

Workfare

Child labor Subsidies

Reduced food consumption Social funds

Seasonal or temporary

migration

Cash transfers

Source: Adapted from World Bank (2001b).

WORLD

DEVELOPMENT

1090

might sound, the ideal way of addressing thatquestion is very difficult to implement in prac-tice. The aggregate nature of these shockscombined with relative scarcity of longitudinaldata on households observed before and afterthe crisis eliminates the possibility of using thedifference-in-differences estimator that is thepreferred estimator in the nonexperimentalevaluation literature (e.g., Heckman, La Londe,& Smith, 1999). 4 Given these constraints mostresearch on the impact of economic crises andnatural disasters has to rely either on cross-sectional surveys conducted during or after thecrisis event, or on short panel surveys thathappened to have at least two observations onhouseholds with one being before the crisis andthe other during or soon after the crisis (e.g.,Fallon & Lucas, 2002).

The two papers related to impact employsome new ways of evaluating the impact of acrisis on a variety of dimensions of householdwelfare. The paper by Datt and Hoogeveen,uses cross- sectional household survey data for1998, to assess the distributional impact of therecent economic crisis in the Philippines. Adistinguishing feature of that survey is that italso includes direct questions on the crisis,where households were asked if they were ad-versely affected by the crisis in five differentways including loss of job, reduced wages anddrought or El Ni~nno. The answers to thesequestions allowed the authors to distinguishbetween households experiencing three mutu-ally exclusive categories of shock: (i) labormarket shock alone; (ii) El Ni~nno shock aloneand (iii) joint labor market and El Ni~nno shock.Keeping in mind that the answers to the ques-tions are subjective, their results suggest thatthe impact of the crisis was modest, leading to5% reduction in average living standards and a9% increase in the incidence of poverty, withhigher increases indicated for the depth andseverity of poverty. Moreover, the largest shareof the overall impact on poverty appeared to beattributable to the El Ni~nno shock as opposed toshocks mediated through the labor market.Both household and community characteristicsmattered to the differential impact of the crisis,with the poorer households being less able toprotect their consumption.

The paper by Handa and King investigatesthe effects of the September 1991 liberalizationof the exchange rate in Jamaica on the weightfor height of children, which is an indicator of achild�s nutritional status that is insensitive toshort-term fluctuations in living conditions.

Using eight years of nationally representativerepeated cross-sectional survey data for 1989–96, they control for a number of observablechild and parental characteristics and thencarefully disentangle the impacts of a child�sage, date of birth, and measurement date onchild wasting (or weight for height). Their es-timates indicate that in the aftermath of theliberalization of the exchange rate, children,and particularly those in the urban areas,weighed significantly less than comparablechildren few months later. The authors alsofind that the elasticity of weight for height z-score with respect to food price inflation wasvery high, which suggests that adjustment pol-icies which bring about sudden and largechanges in the inflation rate can have real ef-fects on children�s nutritional status.

As discussed above, the extent to which ag-gregate shock impact on household welfare isintimately related to their capacity to cope withshocks. The paper by Carter and Maluccio usespanel data from children in South Africa toexplore in more detail how successfully house-holds cope with shocks. Carter and Maluccioexamine the effects of shocks on child nu-tritional status (measured by height-for-age z-scores) by paying particular attention towhether households have a more difficult timecoping with covariant as opposed to idiosyn-cratic shocks and whether access to ‘‘bridgingsocial capital’’ facilitates household capacity todeal with idiosyncratic shocks. Their findingsreveal that households are unable to insurefully against idiosyncratic risk and that com-munity-level shocks have little or no directeffect on household level outcomes. Instead,community-level shocks appear to affect house-holds indirectly by straining the effectivenessof informal sharing mechanisms. For example,the damage to child nutritional status fromhousehold-level losses appears to be greaterin communities that experienced large losses.Households in communities with more socialcapital, however, seem better able to cope withidiosyncratic shocks.

The overall picture emerging from the lasttwo studies suggests that macroeconomicshocks, even if short-lived, may also have ad-verse consequences on the prospects of futuregenerations. 5 A number of studies provideevidence that child malnutrition is correlatedwith lower school achievement, and attainment,lower health as an adult and lower wages andproductivity as an adult (e.g., Glewwe, Jacoby,& King, 2000). To the extent that such links are


strong and prevalent then public actions thatprevent deterioration in the nutritional statusof children and maintain access to health ser-vices during periods of crisis may not only servethe role of a safety net or insurance but alsocontribute to poverty alleviation and inequalityreduction in the future.

The findings that economic crises have anadverse impact on child nutrition naturally raisethe question as to whether household copingstrategies also have an impact on other forms ofhuman capital such as child schooling andwork. The available evidence on this questionsuggests that this is also the case. One of the firststudies on this topic by Jacoby and Skoufias(1997) finds that child school attendance de-creases as a consequence of the shocks ex-perienced by poor rural households in India.Duryea (1998) for Brazil, and Skoufias andParker (2002) for urban Mexico, both findsimilar negative effects on school attainmentrather than just attendance. Finally, Flug,Spilimbergo, and Wachtenheim (1998), usingaggregate crosscountry panel data find a sig-nificant negative correlation between secondaryschool enrollment rates and income or employ-ment volatility.

In contrast to these earlier findings, the evi-dence presented in two of the conference paperssuggests that the effects of crisis on householdeducation spending and schooling seem to bemuch milder. McKenzie, for example, findsthat school attendance rates actually roseamong 15–18 year olds during the Mexicancrisis. One possible explanation for such afinding is that aggregate shocks give rise toopposing income and substitution effects de-termining children�s school and work behavior.On the one hand, decreases in real householdincome, usually following a depreciation of thedomestic currency and increases in the domesticinflation rate, lead households to shift theirchildren out of schooling activities towardwork. On the other hand, decreases in realwages during periods of crisis lower the price(or the opportunity cost) of schooling inducinghouseholds to keep their children in school.Within this framework, the question of whetherit is the negative effect of income or the positivesubstitution effect of wages that dominates canonly be determined empirically. 6

The paper by Duryea and Arends-Kuenningexamines these issues in more detail. Usinglarge repeated cross-sectional surveys fromBrazil over a 20-year period (1977–98) theyexamine the determinants of the decision to

work and/or attend school. Controlling forhousehold income and other child and parentalcharacteristics, they find that for 14–16 year oldboys in urban Brazil, the marginal effect ofwages on the probability of attending school(being employed) is negative (positive) duringthe years of ‘‘normal’’ economic growth. Thusduring normal years children are more likely toleave school when labor market conditionsimprove and wages increase. But during peri-ods of an economic crisis, the size of the mar-ginal effect of wages on the probability ofattending school is significantly lower whichsuggests that during a crisis period there areforces at work tending to decrease the negativerelationship between wages and schooling. 7 Infact, during crisis years the ‘‘normally’’ positiveeffects of wages on children�s work seem to becompletely nullified. 8 The recent findings sug-gest that the question of whether economiccrisis and disasters lead to decreased child-schooling remains to be settled. It would not besurprising if both the direction and the magni-tude of the effect of aggregate shocks on childschooling and work turns out to vary fromcountry to country depending on the level ofurbanization and the financial and economicdevelopment.

Additional coping mechanisms are investi-gated in the paper of McKenzie who uses therepeated cross-sectional surveys from Mexico.The coping mechanisms examined (in addi-tion to the changes in child schooling men-tioned above) include adjustments in householdstructure, changes in fertility, changes inhousehold labor supply, and interhouseholdtransfers. McKenzie�s findings confirm thathouseholds rely on a portfolio of strategiesrather than one single strategy to cope with risk.In the case of Mexico it seems that the primarycoping strategies used were declines in fertility(with approximately 1 in 20 households post-poning having a child), and interhouseholdtransfers (transfers received from abroad in-creased). Only small changes were detected inhousehold structure and in household laborsupply.

One of the first investigations of householdcoping strategies to the AIDS epidemic inUganda and other similarly affected countriesin the region is the paper by Deininger, Garcia,and Subbarao. As the authors document, theAIDS crisis has swelled the number of orphansat risk, and led to a large increase in the mag-nitude of child fostering by households. Thesecoping strategies, in turn, have changed the


nature of risks faced by households and com-munities, and may threaten the care-giving ca-pacity of communities. An examination of theimpact of orphanhood on child education andhealth outcomes using household data revealedthat although foster children in Uganda did notexperience a serious disadvantage in education,they were at a disadvantage in terms of healthoutcomes including immunization health. Theauthors also show that households fosteringorphan children consume less, save less andinvest less. These observed trends are likely tohave serious macroeconomic impacts on ag-gregate savings and investment in the economy.

Undoubtedly, there is a lot of room forpublic interventions to minimize exposure andimpact of economic crises and natural disasters.The paper by del Ninno, Dorosh, and Smithserves as a bridge to the next topic of the con-ference regarding the role of public actions attimes of crises. del Ninno, Dorosh, and Smithexamine not only the coping strategies ofhouseholds but also the role of public policiesand markets following the 1998 flood in Ban-gladesh. At their peak, the 1998 floods coveredtwo-thirds of Bangladesh, causing severe dam-age to the major rice crop and threatening thefood security of tens of millions of households.Their study highlights how the combinationof suitable government policies including anearlier trade liberalization, well-functioningprivate markets, public, and NGO interven-tions and effective private coping strategies wassuccessful at preventing a major post-disastercrisis. The impacts of the floods on flood-exposed households are examined using a paneldata set covering 750 households in threerounds over a 13-month period. The study findsthat private sector borrowing, a major house-hold coping strategy, played a key role inhelping households to maintain consumption.

3. EX-ANTE AND EX-POST PUBLICACTIONS TO MINIMIZE EXPOSURE TO

AND IMPACT OF SHOCKS

The recent experiences of a number of Asianand Latin American countries with finan-cial crises has generated a large and growingliterature on safety nets particularly withinmultilateral agencies. The survey of Blomquist,Cordoba, Verhoeven, Moser, and Bouillon(2002) focusing on the lessons learned fromhow various governments in Asia and LatinAmerica responded to the crises identifies at

least two key principles that should guidethe design, formulation and implementation ofpublic responses. In the first place, public re-sponses are likely to be more effective if they arebased on programs and mechanisms that are inplace before a crisis occurs. When safety netprograms are not in place prior to the inci-dence of a crisis the difficulty of public effortsto protect households in the midst of crisisrises exponentially. During a crisis, constraintsmultiply due to the severe scarcity of fiscal re-sources, the lack or weakness of institutionalcapacity to act quickly, the lack of instruments,and severe information problems. Second, it isessential that the programs are targeted, pro-vide adequate protection to the poor, avoidcreating a culture of dependency among bene-ficiaries, and are consistent with economic in-centives and overall targets of fiscal policy.

Table 3 presents some of the instrumentsavailable to governments in the event of aneconomic crisis or a natural disaster. The list ofinstruments includes cash transfer and publicwork programs, unemployment assistance, wageand commodity price subsidies, targeted humandevelopment or cash transfer programs condi-tioned on school attendance and regular visitsto health centers, service fee waiver, food andnutrition programs, micro-finance and socialfund programs. Table 3 also describes the tar-geting method commonly used with each par-ticular type of intervention, discussed in a bitmore detail below, and the advantages anddisadvantages associated with the intervention.Clearly, the variety of instruments availablesuggests that governments have to make somevery difficult choices. A number of the paperspresented at the conference provided some ex-cellent illustrations of how governments canmake more informed choices on the type andnature of the safety net programs they imple-ment and how to help households cope withcrises and natural disasters.

As a start, the paper by Owens, Hoddinott,and Kinsey provides an interesting analysis ofdrought shocks in Zimbabwe focusing on thepotential effects of shifting government inter-ventions from ex-post mitigation and copingprograms to ex-ante risk reduction programs.The ex-post response is assumed to mirror theprincipal policy of the government that pro-vided targeted grain loans augmented by asmall supplementary feeding program for chil-dren under five. The ex-ante strategy is assumedto consist of additional capital and extensionservices provided to households in the survey


Table 3. Public sector interventions in response to crises and natural disasters

Intervention type

(typical programs)

Beneficiaries Common targeting methods Advantages Disadvantages

Cash transfers (family

allowance, poor unem-

ployed and elderly assis-

tance, disability

assistance)

––Poor families, women and

children

––Means and proxy means

and/or

––Do not distort prices ––Can distort incentives to

labor market participation

––Working poor including

informal sector

––Categorical ––Transfers are fungible,

can directly meet critical

household needs

––Transfers are fungible,

subject to unintended

household uses

––Disabled ––Implementation is infor-

mation intensive

––Poor elderly

––Other vulnerable groups

Public works (labor-

intensive, usually infra-

structure development

projects)

––Poor unemployed and

underemployed including

informal sector

––Self selection (by setting

program remuneration be-

low the minimum wage) and

––Can be implemented or

adapted quickly after crisis

onset provided capacity ex-

ists

––Can distort incentives to


––Poor agricultural workers

during off seasons

––Geographic ––Program size can be easily

reduced once the crisis is

over

––Substantial leakage to

nonpoor depending on pro-

gram design and targeting

methods

––Needed infrastructure is

created or maintained

––Difficult to administer,

tradeoff between infrastruc-

ture development and pov-

erty alleviation objectives

Unemployment assistance

(unemployment benefits,

severance payments)

––Formal sector unem-

ployed

––Coverage determined by

eligibility and employer/em-

ployee contributions

––Provides immediate assis-

tance to eligible beneficiaries

in the event of a crisis

––Can distort incentives to


––Has automatic counter-

cyclical financing character-

istics

––Difficult to adapt quickly

due to qualification and

contribution requirements

––Biased to urban formal

sector

Wage subsidies ––Formal sector unem-

ployed, working age youth,

usually poor

––Targeting by firm type,

industrial category, firm

size, and/or age of the

worker

––Can be implemented

quickly after crisis onset

––Substantial negative in-

centive effects for employers

WORLD

DEVELOPMENT

1094

––Can reach individuals

with variety of skills and

experience

––Biased to urban formal

sector

Commodity price subsidies

(food, energy, housing)

––Poor and extreme poor

families, especially the ur-

ban working poor

––Self-selection (by subsi-

dizing only basic staples)

––Potentially low adminis-

trative costs, depending on

delivery mechanism

––Distorts commodity

prices and use

––Can be implemented or

expanded quickly after crisis

onset

––Substantial leakage to

nonpoor depending on

commodity consumption

patterns

––Often biased to urban

populations

––Difficult to remove once

established due to interest

group pressure

Targeted human

development (conditional

transfers such as school

attendance or preventa-

tive health care receipt

linked to cash transfers)

––Poor students ––Geographic and/or ––Can improve school at-

tendance and/or health care

use

––Effectiveness influenced

by existing education/health

infrastructure

––Poor families with access

to health services

––Categorical and/or ––Supports income of the

poor

––Extensive monitoring and

compliance costs

––Means or proxy means

and/or

––May promote human

capital development

––Community (together

with one of above)

Service fee waivers (school

fees, scholarships, health

care)

––Poor students ––Geographic and/or ––May promote human

capital development

––Effectiveness influenced

by existing education/health

infrastructure

––Poor families with access

to health services

––Categorical and/or ––Limited evidence of long-

term impact on school at-

tendance or health

––Means or proxy means

and/or

––Community (together

with one of above)

ECONOMIC

CRISESAND

NATURALDISASTERS

1095

Table 3 (continued)

Intervention type

(typical programs)

Beneficiaries Common targeting methods Advantages Disadvantages

Food and nutrition (school

feeding, supplemental

feeding and nutrition

for young children and

women)

––Small children, pregnant

and lactating mothers

––Geographic ––Can be effective in allevi-

ating hunger, increasing

school attendance for poor

children

––Limited beneficiary group

––Children attending

schools in poor communities

––Means or proxy means ––May promote human

capital development

––Resource intensive

––Self-targeting ––Substantial benefit leak-

ages depending on targeting

method

––Often biased to urban

populations

Microfinance

(microenterprise credit,

seasonal rural, and emer-

gency credit for the poor)

––Poor microentrepreneurs ––Means and proxy means

and/or

––Promotes physical capital

accumulation in poor com-

munities

––Limited beneficiary group

––Poor women ––Geographic and/or ––May increase household

income

––Administratively costly

––Individual project quality ––Benefits of public re-

sources may be enhanced by

multiplier investment effect

––Biased to rural popula-

tions

––Limited application to

economy-wide crises be-

cause of procyclical demand

for microcredit

Social funds (small scale

infrastructure develop-

ment, microenterprise

support, community-

based social services)

––Poor families, women and

children

––Geographic ––May promote human and

physical capital accumula-

tion in poor communities

––Difficult to implement or

adapt quickly after crisis

onset

––Poor unemployed and

under-employed

––High degree of commu-

nity involvement in project

selection and implementa-

tion

––Often biased to rural

populations

Source: Adapted from Annex II of Blomquist et al. (2002).

WORLD

DEVELOPMENT

1096

only two years prior to the incidence of the1994–95 drought. Using their estimated effectsof additional extension services and capitalstocks on net crop income, the authors thenestimate the effect of higher crop incomes onthe holdings of agricultural tools and livestock(assuming that net private transfers are notcrowded out). As shown in the paper, the re-allocation of funds from an ex-post response toshocks to an ex-ante intervention is successfulat reducing poverty in nondrought years whileat the same time allowing households to buildup additional stocks of livestock that helpbuffering consumption in the aftermath of the1994–95 drought. Thus shifts to ex-ante policyresponses to potential shocks promise to bewelfare enhancing as well as poverty reducing.

One of the main problems of designing tar-geting mechanisms during a crisis or naturaldisaster is that identifying the ‘‘newly’’ poorcould be difficult and very costly. Householdincome is known to be very volatile, and thusan important percentage of the population of acountry move in and out of poverty periodi-cally, even in noncrisis environments (Baulch &Hoddinott, 2000). This continuous inflow andoutflow of households from poverty places animportant challenge to targeting systems basedon proxy means testing, categorical and geo-graphic targeting, and even community-basedtargeting (e.g., see Grosh, 1994). Most of thesetargeting systems are designed to identify thestructurally poor and thus fail to identify thetemporary poor generated by crisis.

The paper by Sumarto, Suryahadi andPritchett on Indonesia highlights clearly theimportance of adequate targeting systems whendesigning programs that help households copewith crises. The two primary contributions ofthe paper are the formalization of the conceptof dynamic benefit incidence and the evidencepresented on the difference that program designhas on both static and dynamic benefit inci-dence. The static benefit incidence of a programis typically measured by examining the pro-portion of the population of households orpeople covered by the program as one movesfrom lower to higher deciles (or quintiles) of thedistribution of household current consumptionexpenditures. If one were to use a graph, thencoverage may be represented by the height andtargeting by the slope as one moves to higherquantiles of the distribution of consumption.According to the authors, dynamic benefit in-cidence can be defined along similar by addinga third dimension (or axis) to the static inci-

dence graph consisting of deciles or quintilesof the distribution of changes in consumptionfrom period to period (or consumption ex-penditure shocks). Using these concepts theauthors analyze the static and dynamic benefitincidence of two social safety net programs(collectively known as JPS) in Indonesia: anin-kind targeted transfer program that sellssubsidized rice to households based on admin-istrative criteria and a set of public employmentschemes based on self-targeting. The timing ofthe 100-village survey and its panel featurespermit the authors to compare the static anddynamic benefit incidence of these two pro-grams during the period of financial crisis andincreased poverty in Indonesia (see Table 1).

Sumarto, Suryahadi and Pritchett find thatalthough the employment creation schemes hadsmaller average participation than the riceprogram, they had a much better static as wellas dynamic benefit incidence than the sales ofsubsidized rice. Thus the employment-creationschemes distributed benefits to the householdsthat were poorer initially and households thatsuffered negative shocks, thus fulfilling both atransfer and an insurance role. As discussed bythe authors in more detail, these findings havepotentially serious implications on the natureand design of social safety net programs.

The two roles of transfer and insurance pro-vided by natural disaster relief funds are ex-amined in more detail in the paper of Morrisand Wodon. As the authors argue, the alloca-tion of natural disaster relief funds can playonly a limited role in helping the poorer house-holds more. The reason for this lies in the natureof emergency aid typically provided at a time ofa natural disaster. The bulk of emergency aidconsists of food, clothing and medicine, whilethe need for these goods is relatively similarbetween households. As a consequence there isonly limited scope for providing more relief tothose who suffered greater losses or who arepoorer following a disaster. Morris and Wodonuse household survey data collected in thepoorest areas of Honduras six to nine monthsafter Hurricane Mitch to examine how reliefeffort was allocated. They find that the targetingof the relief program was primarily based on thesize of asset losses experienced by householdsand much less on the level of household assetsprior to the hurricane. Thus the allocation ofrelief funds is found to be motivated by insur-ance considerations. But when it comes tothe amount of relief received, they show that itis not significantly related to the size of asset


losses experienced by households, thus con-firming the authors� hypothesis that the na-ture of emergency aid limits ones ability toadjust aid according to the size of the loss ex-perienced.

The papers summarized above do not con-sider cash transfers as a possible alternativefor public interventions for coping at times ofcrisis. Yet cash transfers at times of crisis andnatural disasters targeted toward the mostvulnerable households may offer the least costlyand most rapid mechanism for helping house-holds. The paper by Skoufias addresses thequestion of whether the calorie income elasti-city, an important parameter summarizing thesensitivity of caloric availability to changes inhousehold income, remains the same or chan-ges during periods of radically different relativeprices. Using data from the 1996 and 1999SUSENAS surveys in Indonesia he finds thatthe income elasticity of the demand for totalcalories is slightly higher during the year of thecrisis compared to its level in 1996. This sug-gests that the potential effectiveness of eithercash transfer programs aimed at protectingcaloric availability within households at a timeof crisis are not likely to diminish as a result ofchanges in the price environment faced byhouseholds. At a broader level this finding alsosuggests that structural parameters estimatedusing cross-sectional data from a normal eco-nomic environment continue to be very usefulin describing economic behavior even at timesof crises and higher inflation.

Caloric availability, however, is not equiva-lent to micronutrient availability. Householdsfor example, may shift toward foods that arerich in calories, such as cassava and other rootcrops, at the expense of micronutrient-richfoods such as fruits and vegetables (e.g., seeBlock et al., 2002). In an effort to shed somelight on the sensitivity of micronutrient avail-ability to changes in income, Skoufias also in-vestigates the relationship between income andcalories from cereals (which are rich in caloriesand low in micronutrients) and calories fromother foods (excluding cereals and root crops).The latter analysis reveals that the incomeelasticity for cereals as a group increases whilethe calorie income elasticity for other foodsdecreases. Thus as the purchasing power ofincome decreases, pushing households be-low the minimum level of calories requiredfor subsistence, households tend to allocate ahigher proportion of a marginal increase intheir income to cereals. This finding suggests

that while cash transfer programs targeted topoor households near subsistence levels are stilleffective in protecting caloric intake, they mustbe complemented with programs that protectbasic micronutrient intake.

The last paper in this group by Quisumbingconsiders two additional and closely relatedfactors determining the relative effectivenessof public interventions. The first one concernsthe source of food aid and whether differentforms of food aid have different impacts on thenutritional status of children. The second oneconcerns the identity of the food aid recipient.A growing body of empirical literature suggeststhat the effect of public transfers such as foodaid may differ depending on the identity ofthe transfer recipient. In fact, a number ofhuman capital investment programs such as thePROGRESA program in Mexico have delib-erately targeted cash transfers to women on thegrounds that resources controlled by womenare associated with better educational and nu-tritional outcomes of children (e.g., Skoufias,2001).

Using a panel data set from Ethiopia, Qui-sumbing examines the determinants of partici-pation in and receipts of food aid through twoprograms: free distribution (FD) and food-for-work (FFW). FD receipts, which are notconditioned on work effort, can be consideredanalogous to an increase in nonlabor income.FFW opportunities, on the other hand, may beconsidered as reflecting a change in the wagerate as well as improvements in women�s out-side options according to the increasingly ac-cepted collective model of the household (e.g.,see Haddad, Hoddinott, & Alderman, 1997).The effects of FD and FFW on child nutritionalstatus are found to differ depending on themodality of food aid, and the gender of thechild. Both FFW and FD have a positive effecton weight for height of children. In particular,households appear to invest proceeds from FDin girl�s nutrition, while earnings from FFW aremanifested in better nutrition for boys.

4. CONCLUDING REMARKS ANDLESSONS FOR POLICY

The aggregate nature of economic crises andnatural disasters implies that many of the in-formal mechanisms for mitigating and copingwith risk, and in particular those that arecommunity-based, may become less effective.Under such circumstances households may be


forced to rely on self-insurance strategies thatare particularly costly in terms of current aswell as future welfare.

Knowledge and better understanding of themain coping strategies of households is veryuseful for setting the priorities for public pro-grams and safety nets. Many of the papersin this special conference issue contribute inthis effort by describing the variety of copingstrategies used by households in differentcountries and by examining how successful theywere at protecting household welfare. Mexicanhouseholds, for example, decreased their fer-tility in response to the tequila crisis (McKen-zie). Rural households in Bangladesh borrowedmore soon after the 1998 floods (del Ninno,Dorosh and Smith); Ugandan households re-sorted to fostering orphan children of rela-tives dying from AIDS (Deininger, Garcia andSubbarao), while South African householdswere found to cope with the crisis by relying onlocal support networks (Carter and Maluccio).Yet most of these coping strategies do not ap-pear to be completely effective at protectinghouseholds. Poorer households in the Philip-pines, for example, were found to be less able toprotect their consumption as a result of therecent economic crisis and El Ni~nno shock (Dattand Hoogeveen). At least two of the papers inthe conference (Handa and King; Carter andMaluccio) provide strong empirical evidencethat the self-insurance strategies used at timesof crises are associated with lower child nutri-tion in the short-term (measured by weight forheight) and in the long-term (measured byheight for age). Households fostering orphanchildren in Uganda are found to consume, saveand invest less while orphan children were at adisadvantage in terms of health outcomes andimmunizations. Yet surprisingly little evidencewas found regarding the potentially adverseeffects of aggregate shocks on the schooling ofchildren in Brazil (Duryea and Arends-Kuen-ning). Even in Uganda were the AIDS epidemichas increased dramatically the number of or-phans and the incidence of fostering, fosterchildren do not appear to experience a seriousdisadvantage in education (Deininger, Garciaand Subbarao).

In combination these findings suggest thataggregate shocks, even if short-lived, are alsolikely to have irreversible consequences on theprospects of future generations. Public actionsthat prevent deteriorations in the nutritionalstatus of children and maintain access to healthservices for poor and vulnerable households

deserve top priority. Should such interventionsalso include incentives for families to keepchildren in school? While there is scatteredevidence pointing to the affirmative, the ab-sence of any solid evidence from the paperspresented at the conference can only suggestthat this issue deserves further investigation.

Having shed some light on the priorities ofinterventions at time of crises and natural dis-asters the conference papers also addressed thedeeper issues related to the timing and theguiding principles that can increase the effec-tiveness of public interventions in protectinghouseholds welfare. A number of importantfindings stand out. First, ex-ante risk reductionprograms offer ‘‘a bigger bang for the dollar’’relative to ex-post mitigation and copingprograms since they appear to be welfare en-hancing as well a poverty reducing (Owens,Hoddinott and Kinsey). Second, governmentled reforms such as an earlier trade liberaliza-tion combined with well functioning privatemarkets, public and nongovernmental organi-zations (NGOs) interventions have the poten-tial of being successful at preventing majorcrises as was the case in Bangladesh after the1998 floods (del Ninno). Insofar as govern-ments have the foresight and the discipline toadopt such strategies there are serious oppor-tunities for reducing the adverse effects of cri-ses.

Third, programs that target the structurallypoor do not necessarily reach households thatcan be affected by economic crises and disas-ters. If poverty alleviation programs are toalso serve the role of providing insurance, itis necessary to adopt appropriate targetingmechanisms. Targeting based on administrativecriteria, proxy means tests, or categorical vari-ables do a relatively better job at identifying thehouseholds that are ‘‘permanently’’ poor thanthose who are actually hurt by the crises. Incontrast, self-targeting systems typically asso-ciated with employment generation and infra-structure maintenance and creation programsappear to be possess the advantage of provid-ing both insurance and transfer (Sumarto,Suryahadi and Pritchett).

Fourth, the nature of the aid that a programprovides may also constrain the extent to whicha program can help poorer or more needyhouseholds. Emergency aid, for example, con-sists mainly of in-kind transfers such as food,clothing and medicine. To the extent that theneeds for these goods are the same acrosshouseholds, there is only limited scope for


providing ‘‘more’’ relief to those who sufferedgreater losses or who are poorer following adisaster (Morris and Wodon). Targeted cashtransfers, on the other hand, seem to offer aquick and flexible alternative that could bedistributed at a lower administrative cost thanin-kind transfers, and may be easier to tailoraccording to the losses of the affected house-holds. At least for the instances where the ob-jective is to prevent a deterioration of thecaloric availability at the household level, thehigher calorie income elasticity among poorerhouseholds and the insensitivity of the elasti-city to changes in relative prices ensure thata well-targeted cash transfer program will havea greater effect on the caloric availability ofpoorer households (Skoufias). To the extentthat the availability of micronutrients is also ofprimary consideration then cash transfer pro-gram may need to be complemented by pro-grams that protect micronutrient intake.

Finally, the design of crisis relief and socialsafety net programs directed to specific mem-bers of households, such as children or pre-gnant women, must also take into considerationthe intermediary role of families (Quisumbing).Person-specific transfer programs run the risk ofhaving a lower effect on the intended individualsdue to reallocations of resources within familiesover which policymakers have no direct control.

In recent years a number of countries, in-cluding Bangladesh, Brazil, Colombia, Mexico,Honduras, Jamaica, and Nicaragua, have shif-ted their national poverty alleviation strategiestoward cash transfer programs targeted to poorhouseholds and conditioned on householdsinvesting in the nutrition, health and educationof their children. 9 Moreover, the transfers as-sociated with these programs are deliberatelytargeted to women on the grounds that re-sources controlled by women are associatedwith better educational and nutritional out-comes of children. The preceding findings sug-gest that programs of this type provide a goodfoundation toward the creation of social safetynet systems that could be quite effective at timesof crises. Governments, for example, can ensurethat child nutrition, child health and childschooling of the already poor households areminimally affected by the crisis, by changingthe amount of the cash transfer to the poorhouseholds already in the program. How-ever, these programs do not have the built-inflexibility to expand coverage to householdsfalling below the poverty line during timesof crises. It is hoped that the papers in thisspecial issue make a useful contribution towardthe effort to combine long-run poverty alle-viation with effective relief during periods ofcrises.

NOTES

1. A more detailed discussion of these issues can be

found in Chapter 9 of the 2001 WDR, in Ferreira,

Prennushi, and Ravallion (1999) and in Baldacci, de

Mello, and Inchauste (2002).

2. It is important to keep in mind that in the case of a

decrease in the value of the domestic currency, producers

exporting goods and services are likely to increase

production and demand for labor as demand for their

products from abroad increases.

3. It should be noted that the mechanisms for manag-

ing risks listed in Table 2 are general enough as to be

also applicable to other types risk such as those related

to health, crime and violence, and the environment.

4. For an example of a study using the double

difference estimator to evaluate the impact of financial

crises see Baldacci et al. (2002). It should be noted,

however, that the double difference estimator in this

paper relies on country-level time series observations

from a number of countries some of which experienced a

financial crisis, instead of panel data at the household

level.

5. Related evidence on the adverse effects of droughts

and civil wars on child height-for-age z-scores can be

found in Hoddinott and Kinsey (2001) and Alderman,

Hoddinott, and Kinsey (2002), respectively.

6. The discussion above presumes that schooling is a

normal good and a simple static model of time alloca-

tion between productive and nonproductive (school)

activities. If one is willing to acknowledge that schooling

involves intertemporal tradeoffs, then one must also take

into account the role of credit markets and insurance

(e.g., Jacoby & Skoufias, 1997).

7. Interestingly, Duryea and Arends-Kuenning find

that the marginal effect of household income on the

probability that a child attends school and/or works is

the same during a period of crisis as in normal economic


conditions. This is consistent with the findings of

Skoufias on the sensitivity of the calorie–income elasti-

city in Indonesia.

8. In fact, Schady (2001) also finds that children

exposed to the 1988–92 economic crisis in Peru com-

pleted more years of school.

9. The list of such programs includes PROGRESA

(recently renamed Oportunidades) in Mexico, Bolsa

Escola and Bolsa Alimentacao in Brazil, Programa de

Asignacion Familia (PRAF) in Honduras, Red de Pro-

teccion Social (RPS) in Nicaragua, Food for Education in

Bangladesh, Familias en Accion and Empleo en Accion in

Colombia and PATH in Jamaica.

REFERENCES

Alderman, H., Hoddinott, J., & Kinsey, B. (2002). Longterm consequencies of early childhood malnutrition.Unpublished manuscript, IFPRI, Washington, DC.

Alderman, H., & Paxson, C. (1994). Do the poor insure?A synthesis of the literature on risk and consumptionin developing countries. In E. Bacha (Ed.), Econom-ics in a Changing World. New York: St. Martin�s.

Baldacci, E., de Mello, L., & Inchauste, G. (2002).Financial crises, poverty, and income distribution.IMF Working Paper WP/02/4.

Baulch, B., & Hoddinott, J. (2000). Economic mobilityand poverty dynamics in developing countries. Jour-nal of Development Studies (Special Issue), August.

Binswanger, H., & Rosenzweig, M. R. (1986). Behav-ioral and material determinants of production rela-tions in agriculture. Journal of Development Studies(April), 503–539.

Block, A. S., Keiss, L., Webb, P., Kosen, S., Moench-Pfanner, R., Bloem, M. W., & Timmer, C. P. (2002).Did Indonesia’s crises of 1997/98 affect child nutrition?A cohort decomposition analysis of National NutritionSurveillance data, Fletcher School of Law andDiplomacy, Tufts University.

Blomquist, J., Cordoba, J. P., Verhoeven, M., Moser, P.,& Bouillon, C. (2002). Social safety nets in responseto crisis: lessons and guidelines from Asia and LatinAmerica. In OECD, Towards Asia�s SustainableDevelopment: The Role of Social Protection (pp.297–332). Paris: OECD.

Duryea, S. (1998). Children’s advancement through schoolin Brazil: the role of transitory shocks to householdincome. Inter-American Development Bank, Officeof the Chief Economist, Working Paper No. 376,July.

Fallon, P., & Lucas, R. E. B. (2002). The impact offinancial crises on labor markets, household incomesand poverty: a review of evidence. The World BankResearch Observer, 17(1), 21–45.

Ferreira, F., Prennushi, G., & Ravallion, M. (1999).Protecting the poor from macroeconomic shocks: anagenda for actions in a crisis and beyond. Mimeo.

Flug, K., Spilimbergo, A., & Wachtenheim, E. (1998).Investment in education: do economic volatility andcredit constraints matter? Journal of DevelopmentEconomics, 55, 465–481.

Glewwe, P., Jacoby, H., & King, E. (2000). Earlychildhood nutrition and academic achievement: alongitudinal analysis. Journal of Public Economics,81, 345–368.

Grosh, M. (1994). Administering targeted social pro-grams in Latin America: from platitudes to practice.Washington, DC: The World Bank.

Haddad, L., Hoddinott, J., & Alderman, H. (Eds.).(1997). Intrahousehold resource allocation in develop-ing countries: methods, models, and policy. Baltimore:Johns Hopkins University Press for the InternationalFood Policy Research Institute.

Heckman, J. J., La Londe, R., & Smith, J. (1999). Theeconomics and econometrics of active labor marketprograms. In O. Ashenfelter, & D. Card (Eds.),Handbook of labor economics, vol. 3A. Amsterdam,The Netherlands: North Holland.

Hoddinott, J., & Kinsey, B. (2001). Child health in thetime of drought. Oxford Bulletin of Economics andStatistics, 63, 409–436.

Holzmann, R., & Jorgensen, S. (2001). Social riskmanagement: a new conceptual framework for socialprotection and beyond. International Tax and PublicFinance, 8(4), 529–556.

Inter-American Development Bank (2000). Social pro-tection for equity and growth. Washington, DC:IADB.

Jacoby, H., & Skoufias, E. (1997). Risk, financialmarkets, and human capital in a developing coun-try. Review of Economic Studies, 64(3), 311–335.

Lustig, N. (2000). Crises and the poor: socially respon-sible macroeconomics. Economia, 1(1), 1–19.

Morduch, J. (1999). Between the state and the mar-ket: can informal insurance patch the safety net?The World Bank Research Observer, 14(2), 187–207.

Ravallion, M. (2002). Who is protected? On the incidenceof fiscal adjustment. World Bank Working Paper,February.

Schady, N. (2001). The (positive) effect of macroeco-nomic crises on schooling and employment deci-sions taken by children in a middle-income country.Paper presented at the Crises and Disasters Con-ference, IFPRI, Washington, DC, November 13–14.

Skoufias, E., 2001. The impact of PROGRESA on thewelfare of rural households in Mexico ResearchReport. Washington, DC: IFPRI.

Skoufias, E., & Parker, S. W. (2002). Labor marketshocks and their impacts on work and schooling:evidence from urban Mexico. IFPRI-FCND Discus-sion Paper #129, March.


Von Braun, J., Vlek, P. L. G., & Wimmer, A., (2002).Disasters, conflicts and natural resource degradations:multidisciplinary perspectives on complex emergencies.ZEF Bonn. Center for Development Research.Annual Report 2001/2002.

World Bank (2001a). World Development Report 2000/01attacking poverty. New York: Oxford University Press.

World Bank (2001b). Social protection sector strategy:from safety net to springboard. Washington, DC:World Bank.


economic crises and natural disasters: coping strategies and policy implications

Documents