cash transfers in emergencies: evaluating ... - odi hpn · network coordinator would appreciate...

June 2001

ABSTRACT

In terms of both theory and practice, thereappears to be a strong case for cash-basedresponses to food emergencies where thesupply and market conditions areappropriate. Amartya Sen’s work onentitlements offers a solid theoretical basefor cash transfers, and the practicalexperience so far, limited though it is,provides evidence that direct cashdistribution, in the right circumstances andwith careful planning and monitoring, canbe more timely, less costly and moreempowering to local communities thantraditional food distribution. Nevertheless,there appears to be a reluctance within thehumanitarian relief system to include cash-based responses in emergency responseportfolios.

This paper reviews the theoreticalunderpinnings of a cash-based approach tofood emergencies, and presents case-studies

of cash distribution. These examples, whichare drawn from Africa, South Asia and theBalkans, highlight both the risks and thebenefits of cash-based responses as againsttraditional food aid. On the one hand, cashis more cost-effective because its transactioncosts are lower; it is more easily convertible,allows for greater beneficiary choice andcan stimulate local markets. On the otherhand, cash can be used in ways not intendedby the donor, can contribute to localinflation and poses security risks notnormally associated with food aid. Thepaper concludes by setting out theconditions under which cash aid might bean appropriate response, and highlights howits associated risks can be minimised. Therecan be no ‘blueprint’ for the use of cashacross all emergencies and in allcircumstances; instead, agencies need toweigh the benefits against the risks on acase-by-case basis.

Cash transfers in emergencies:evaluating benefits and assessing risksby David Peppiatt, John Mitchell and Penny Holzmann

35

Humanitarian Practice Network (HPN)Overseas Development Institute111 Westminster Bridge RoadLondon, SE1 7JDUnited Kingdom

Tel: +44 (0)20 7922 0331/74Fax: +44 (0)20 7922 0399Email: <[email protected]>Website: <www.odihpn.org>

Layout and production: Rebecca Lovelace, HPNPrinting: DS Print & ReDesign, Enfield, London

ISBN: 0 85003 523 6Price per copy: £5.00 (excluding postage and packing).© Overseas Development Institute, London, 2001.

Photocopies of all or part of this publication may be made providing that the source is acknowledged. Requestsfor the commercial reproduction of HPN material should be directed to the ODI as copyright holders. TheNetwork Coordinator would appreciate receiving details of the use of any of this material in training, researchor programme design, implementation or evaluation.

Notes on the Authors

Comments

This Network Paper does not pretend to be the definitive treatment of what is an emerging area of debateand discussion. The HPN invites comments concerning this paper, and particularly encourages readers totell us of any practical experiences with cash-based interventions.

Acknowledgements

This paper first appeared as a British Red Cross Working Paper entitled Buying Power: The Use of CashTransfers in Emergencies, published in November 2000.

The authors gratefully acknowledge the assistance of all the individuals and agencies who contributedthroughout the writing of this paper, in particular Stephen Devereux, Mark Lawrence, Oxfam, SCF(UK),UNHCR and UNICEF. The authors would also like to thank Matthew Foley and Margie Buchanan-Smith atthe ODI for their editorial support.

David Peppiatt is a technical advisor on Red Cross and Red Crescent disaster-mitigation and preparednessprogrammes, including food security, in Africa and Asia.

John Mitchell is Senior Humanitarian Advisor at the British Red Cross.

Nutritionist and freelance editor Penny Holzmann has worked with Save the Children on developing a newmethodology for vulnerability assessment and risk-mapping.

Table of ContentsExecutive summary 1

Chapter 1 Cash distribution in theory 3

Chapter 2 Cash distribution in practice 5

Cash distribution and food emergencies 5

Cash distribution and rehabilitation 7

Cash distribution and conflict 8

Cash distribution and social interventions 9

Chapter 3 Assessing the effectiveness of cash-based responses 11

A mixed picture 11

The benefits of a cash-based response 12

The risks of a cash-based response 14

Balancing benefits and risks 16

Conclusion 19

References 21

List of boxes and tablesBox 1: Defining ‘entitlement’ 3

Box 2: Cash distribution in western Sudan, 1984 5

Box 3: Spending patterns in the Ethiopian intervention 6

Box 4: Cash distribution in Bangladesh, 1998 8

Box 5: The role and capacity of local traders 12

Table 1: Food versus cash 17

Glossary

CFF Cash-For-Food

CFW Cash-For-Work

DFID Department for International Development (UK)

EU European Union

FAO Food and Agriculture Organisation

GAPVU Gabinete de Apoio à População Vulneravel

LRA Lord’s Resistance Army

SCF(UK) Save the Children Fund (UK)

UNHCR UN High Commissioner for Refugees

UNICEF UN Children’s Fund

UNMIK UN Interim Mission in Kosovo

WFP World Food Programme

Executive Summary

1

The idea of distributing cash relief in famines iscontroversial. Although it is increasingly recognisedthat food aid serves not purely as a nutritionalintervention, but also as a transfer of an economicresource, there is considerable reluctance todistribute cash in place of food. Indeed, since themonetisation of food aid is now accepted asimportant in strengthening livelihood strategies, itis surprising that the idea of direct cash transfers isso rarely thought of as a practical alternative by therelief community. Nonetheless, there appears to bea growing willingness to at least consider the use ofcash as an alternative to either direct food aiddelivery or the delivery of non-food relief items,and as the medium of support in safety nets.

This paper tackles some of the key questions posedby cash distribution. Its focus is on distribution innatural disasters, and/or in relatively peaceful settings.(Cash distribution in the midst of a complex politicalemergency – where there may be no functioningstate – throws up very different issues, such as thebreakdown of markets and of a working bankingsystem.) Although the paper is pr incipallyconcerned with straight cash transfers, it also drawssome lessons from Cash-For-Work (CFW)programmes. While its starting-point and emphasislie in the use of cash as a substitute for food, it alsogoes a step further and considers cash as a substitutefor non-food items, particularly in rehabilitationprogrammes, and its use in safety-net programmes.

The paper summarises the theoretical rationale forusing cash, reviews agency experiences of cashdistribution, and outlines the main benefits and risksinvolved in this type of response. It identifies anumber of advantages that cash distribution has over

traditional food distribution. These include thepotential for faster delivery and lower transactioncosts, and the possible beneficial impact of a cashinjection on local markets and trade. Delivering cashrather than food also addresses the problem ofidentifying requirements, since beneficiaries are ina position to determine these themselves. The rangeof food items that can be purchased may be widerand more appealing than the standard food-aidbasket. Finally, there may also be benefits to be hadin terms of livelihood security.

Against these advantages, one must weigh the risksof cash responses. The key potential danger stemsprecisely from money’s flexibility and fungibility:how can donors ensure that their aid is going whereit is intended? Targeting can also become moredifficult, since cash is of inherent value to everyone,and does not allow for self-selection. While theimpact of an infusion of cash may stimulate a localeconomy, it may also lead to inflation and increasedprices, potentially penalising people not includedin the programme. There are also potentialproblems to do with security: even in relatively stableenvironments, agencies distributing large amountsof cash face the risk of theft; this risk is heightenedin conflict-related emergencies, where beneficiariesof a cash distribution may also be targeted bybelligerents for that very reason.

Although there is clearly more work to be done inthis area, this paper suggests that, in the right circum-stances, cash distributions can be a viable alternativeto food aid. However, there is no general ‘blueprint’for agencies to follow; ultimately, the decision todistribute cash or food depends on a case-by-caseassessment of the benefits and risks involved.

2

c a s h t r a n s f e r s i n e m e r g e n c i e s

3

t h e p o l i t i c a l e c o n o m y o fw a r

The rationale for a cash-based response derivesfrom Amartya Sen’s explanation of contemporaryfamines. Sen identified the key problem as lack ofaccess to food, rather than failures in food supply(Sen, 1981). He reanalysed the food productionand availability data from several famines andshowed that, for nearly all his case studies, theoverall food supply within the country wassufficient to support the population for the periodin question, and that groups or communitiesbecame vulnerable to under-nutrition becausethey lacked access and purchasing power. In hisstudies of the Bangladesh famine of 1974, forexample, Sen discovered that the areas most severelyaffected had, in fact, enjoyed the highest availabilityof food in that year, compared to other areas. Atthe same time, such areas typically experienced amajor decline in purchasing power due to loss ofwork during flooding, and an increase in foodprices as traders predicted shortages. In such a case,famine is caused not so much by a decline in thefood available, but in people’s access to it – whatSen terms ‘entitlement’.

If famine is caused partly by a decline inentitlement, it follows that an economic responseaimed at boosting purchasing power and increasingfood entitlement can be an appropriate, andperhaps preferable, alternative to general fooddistribution. Injecting cash into a market increasesdemand, which in turn can generate supply. AsWilson (1991) explains:

in situations of regional famine there tend to be pocketsof food surplus, which are not otherwise redistributeddue to costly transport and weak purchasing power.Cash disbursement drives demand through ‘market-based entitlement’ that mobilises food by increasingdemand (and hence prices) and thus encouragingtraders to locate and transport food.

Sen’s entitlement approach emphasises the linksbetween poverty and famine, with the implicationthat famine might be mitigated by protectingpeople’s purchasing power. Income transfer is adirect way of doing this.

The theory of entitlements is now widely accepted(although Sen’s minimisation of the magnitude andimportance of overall food supply failure in Africanfamines is disputed by many). But it appears tohave had only limited influence on famine-reliefprogrammes: famines may be understood from aneconomic point of view, but the responses to themare rarely based on economics. Typically,

Box 1: Defining ‘entitlement’

Sen defined ‘entitlement’ as ‘the command overcommodities that people have’. Devereuxsummarises: ‘the entitlement approach recognisesfour legal ways of acquiring food: growing it(‘production-based entitlement’), buying it (‘trade-based entitlement’), working for it (‘own-labourentitlement’) and being given it (‘transferentitlement’). Individuals face starvation if their‘entitlement set’ does not provide them withadequate food. Famine scales this up: a famineoccurs when occupationally or geographicallyrelated groups of people experience sharpdeclines in their entitlements simultaneously.Entitlement failure can be direct – a loss of accessto production-based entitlements, for instanceduring a crop- and livestock-destructive drought– or exchange-related – a fall in trade- or own-labour-based entitlement due to unfavourableshifts in prices (livestock prices fall, food pricesrise) or incomes (nominal or real wages fall, wagesare lost due to unemployment)’. (Devereux,2000a).

Cash distribution in theory

1

4


interventions are designed on the premise that theavailability of food has declined, and they aimsimply to increase supplies through fooddistribution. This is also reflected in the way thatmany needs assessments are carried out – focusingmore on food supply than on effective demand.

Why has there been such reluctance to putentitlements theory into practice? The first reasonis fear of the unknown. There are very few well-documented cases of direct income transfers and cashdistributions. Key questions remain unanswered,including the effectiveness of the intervention (willsufficient food be attracted to the market to satisfybeneficiaries’ needs?); the effects on the market(will the injection of cash exacerbate inflation?);and the capacity of beneficiaries to use the moneyas intended. This lack of experience requiresexper imentation, which agencies tend to becautious about.

Donor policy has also played a – perhaps the – keyrole in determining the type of aid provided infamines. Traditionally, donor surpluses of food havebeen donated ‘in kind’, rather than as untied cash.Donating agricultural surpluses has certain neteconomic benefits for the donor. Earmarkedfunding has also placed constraints on innovativeresponses, such as cash distribution. The donorclimate is now changing due to declining globalfood stocks and a more strategic approach aimedat strengthening food security. The new EuropeanUnion (EU) Food Aid Regulation, for example,advocates a shift away from the distribution offood aid towards assistance programmes thatfocus on household food security. This means thatthe context in which food aid is used haswidened, and this aid may be monetised morefrequently than hitherto (Mitchell, 1996).Furthermore, as tied food aid declines, untied cashnow represents a larger share of international donorresponse. By way of an example of this trend, in

June 1999 the major donors approved the new FoodAid Convention (FAC). The total volume ofcommitments is 4.89m tonnes in wheat equivalent,compared with 5.35m tonnes under the previousFAC, agreed in 1995. The difference is accountedfor by the EU’s pledge of 130m euros in cash,equivalent to about 588,000 tonnes (FAO, 2000).

Another reason why cash-based responses havebeen ignored is the long-standing assumption thatfamine victims are peculiarly unable to handle cashdistributions within the household (Keen, 1992).This attitude seems partly rooted in the Westernconcept of charity, which assumes that recipientsdo not necessarily know what is in their bestinterests. On the one hand, development rhetoricstresses ‘empowerment’, and urges that beneficiariesshould take the lead in their own relief andrehabilitation; on the other, their needs and interestsare determined by outside agencies and donors. Itis crucial that we begin to challenge the assumptionthat ‘we’ always know what is best for ‘them’, andthat we recognise the inherent contradiction (andhypocrisy) between theory and current practice.Numerous studies on the coping strategies offamine victims, refugees and other vulnerablegroups have confirmed that affected communitiesare more than capable of determining their bestneeds and interests. But they rarely have theeconomic power to do so.

Other constraints include questions of gender andaccess, leakage, diversion and corruption andsecur ity. But many of these risks are equallyapplicable to the delivery of direct food aid: thechallenge remains to explore ways to mitigate them.Neither cash nor food distribution offers a panaceafor the long-term problem of food insecurity; butboth offer the potential to avert human suffering.The following chapter examines in detail some ofthe practical experiences gained in the use of cashtransfers since the mid-1980s.

5


The international relief community’s experienceof cash distributions is limited. This does not meanthat the idea is particularly novel; in 1948, forinstance, the Br itish colonial administrationdistributed cash, coffee and train tickets to faminevictims in Sudan. In India, the British realised thathunger was caused as much by lack of income asby scarcity of food, and responded to famines byproviding waged labour (Keen, 1992). Cash reliefinterventions were implemented in famines inTanganyika, Rhodesia and colonial China (Drèzeand Sen, 1989). There is also a long history of cashrelief in Bangladesh, particularly in response to thecyclone disasters of the 1970s.

Cash distribution and food emergencies

UNICEF in Ethiopia, 1984–85The best-documented case of cash aid is the ‘Cash-For-Food’ (CFF) programme administered by the

UN Children’s Fund (UNICEF) in the 1983–85famine in Ethiopia. In this case, many drought-stricken areas were characterised primarily by alack of access to food, rather than lack of availability.Typically, these areas were identified as famine‘pockets’ cut off from the major distr ibutioncentres, but situated close to areas of food surplus.The people who lived in these areas had severelyweakened purchasing power. It was recognised that,even if sufficient food aid was received from abroad,this might not reach all of the population in needof assistance in time. The aim of the CFFprogramme was to enable the affected populationto secure food supplies from areas of surplus.

Almost 95,000 people compr ising 18,900households were targeted with cash transfers at 14sites in seven of Ethiopia’s administrative regions.Sites were selected according to a number ofcriteria, including accessibility, a settled and non-scattered population, and the availability of amarketable surplus. Special attention was paid tovulnerable categories, such as female-headedhouseholds and large families. Cash was paidmonthly to the beneficiar ies throughrepresentatives and peasants’ associations, enablingthem to obtain food from neighbouring markets,rather than from more distant food-aid distributionsites. While the programme included acommunity-development component (allrecipients were expected to participate regularlyin community-based work schemes), the emphasiswas on relief, rather than on work. The CFFprogramme was planned to last for eight monthsuntil the next harvest, but at some sites it was inplace for two years.

The evaluation of the programme reports itsachievements as ‘considerable’ (UNICEF, 1988a).In most assisted areas, the CFF encouraged the

Box 2: Cash distribution in western Sudan, 1984

Cash was distributed to internally-displacedpeople in Nyala, western Sudan, during a severefamine in 1984 (Wilson, 1991). Immediatelyfollowing the distribution, local food pricesincreased dramatically because existing grainsupplies were unable to meet the sudden demand.However, the increased demand coupled withthe assurance that cash distributions wouldcontinue attracted traders from surplus areas, andled to a rapid and regular flow of grain into thearea, with prices soon returning to reasonablelevels. Wilson notes that this area had the best-developed grain market infrastructure in westernSudan. He also questions the effect of this tradeon grain availability and prices in the non-assisted rural areas, which remains unknown.

Cash distribution in practice

2

6


movement of grain from surplus areas to deficitmarkets, enabling the population to buy the foodthey needed. At many sites, the programmeprevented further distress migration and thesplitting up of families, and forestalled theformation of camps around food distribution sites.Some of those who had left their homes at thebeginning of the drought were able to return.

Price increases were a problem, although this wasthe result of lack of availability, rather than stemmingfrom the injection of cash itself. The overallevaluation found ‘no evidence that CFF has causedlocal price inflation’ (UNICEF, 1988a). In Siqewereda (district) in northern Shoa, the price of maizedoubled between October 1984, when the projectstarted, and April 1985. Thus, beneficiaries couldbuy only half of the intended amount of grain withtheir monthly payment (UNICEF, 1985a). Thispicture is repeated in Menzna Gishe, where theeffective value of the cash payment declined by athird in the first four months of the project, and toa half of its initial value after eight months (UNICEF,1985b). In Konso in Gamu Gofa, prices doubledover the course of the project. Although paymentswere increased in response, beneficiaries reporteda shortage of grain on the market (UNICEF, 1985c).In Yifatna Timuga in northern Shoa, high pricesfrom the outset meant that a typical family couldbuy only about six kilograms of grain per personper month – less than half its requirements – untilprices began to fall ten months later (UNICEF, 1986).

Those areas that did not suffer from rising priceswere either closer to surplus-producing areas, orwere included in the CFF programme from late1985, when the famine was past its peak. The valueof the payments in terms of food was stable in thesesites, and even rose once prices started droppingin the run-up to the main harvest at the end of1986. No shortages of grain in the market werereported (UNICEF, 1987; UNICEF, 1988b).

Perhaps the two major advantages of theprogramme over a relief distribution were the speedand relatively low cost of delivery. Despite delaysin payments in many weredas due to administrativeproblems, beneficiaries almost certainly receivedpayments more quickly than they would havereceived food relief. For example, the first paymentin Konso in Gamu Gofa, which was very hard hit,was received as early as July 1984, only a short timeafter the idea of the project was first mooted(UNICEF, 1985a; UNICEF, 1988a). In YifatnaTimuga, the first cash was issued in November 1984(UNICEF, 1986), and in Mamamidir the first paymentwas received in February 1985 (UNICEF, 1985b).The programme’s evaluation offers a comparativeanalysis of the costs involved. It estimates that, at atotal of $5.5m, the cost of the CFF programme wasroughly half that of World Food Programme(WFP)-supplied grain (UNICEF, 1988a).

Accounting and monitor ing were the mainoperational problems. The CFF programme was

Box 3: Spending patterns in the Ethiopian intervention

A curious characteristic of the Ethiopia project, given that these were famine-affected populations, wasthe way in which beneficiaries spent the cash they were given. In most cases, only 75 per cent of themoney was used to buy food, and up to a third was spent on non-food items such as clothes, animals, seedsand tools, as well as land taxes, dues to peasant associations and debt repayments. The evaluation teamsconcluded that beneficiaries were in fact able to buy food at prices lower than those quoted, and that theywere willing to keep food consumption to a minimum in order to put the cash to alternative uses. Thisraises questions as to how severely affected people at some of these sites really were. But it is consistentwith research on coping strategies, which shows that, in an acute food crisis, people cut back on foodconsumption first (and hence malnutrition rises) in order to preserve their productive assets for as long aspossible (de Waal, 1989). The UNICEF evaluation team looking at the two pilot sites, Tatch Gaint andMamamidir, both of which were badly affected by drought, admitted to finding this ‘confusing’:

most, if not all, of the beneficiaries have bought clothing and some livestock … Virtually all hadbought farm implements … it had never been expected that the recipients would be able to do morethan survive on the money … and market prices had increased. (UNICEF, 1984)

Notwithstanding this unexplained difference between the beneficiaries’ apparent needs and their purchases,one cannot but agree with the team’s conclusions that ‘rural people have a remarkable ability to managetheir own affairs and know very well the value of the little cash they ever see’ (UNICEF, 1984). As Kumar(1985) puts it: ‘cash distribution has enabled recipients to decide their own balance between consumptionand saving and it appears from this project that even poor households have a marked propensity to save’.

7


also considered to be less successful in its medium-and long-term goals of establishing developmentactivities (community-based activities and income-generation). The programme’s long-term economicimpact was minimal, largely because of problemsat the policy level, where there was a contradictionbetween the desire to use CFF as short-termemergency response, and as a means to longer-term rehabilitation. It was recommended that CFFshould be viewed as purely a form of relief, andshould last no longer than six months. UNICEF(1988a) concluded that cash distribution is ‘first andforemost … emergency assistance, an ad hoc actionwhich is directly addressed to the populationaffected, and gets to them rapidly and at lower costs’.

ActionAid in Ghana, 1994ActionAid used cash distribution in 1994 in Ghana’sBawku West District in response to widespreadfood insecur ity stemming from decliningagricultural productivity, high population densityand irregular rainfall. The agency’s emergency‘Food Security Programme’ involved cash transfersto 1,000 households identified as the mostvulnerable within the population.

Households were targeted and registered througha process of community self-targeting whichidentified tarims, or the poorest and most vulnerableindividuals. These included disabled people, thesick and elderly, widows, members of female-headed households and people without livestockand poultry. The identification and registration oftar ims were car r ied out by zonal targetingcommittees and cross-checked by villagecommittees. Identity cards were distributed to allthe registered beneficiaries, and cash was disbursedover five days. Each household received a one-offpayment of 10,000 cedis.

According to the review of the emergencyprogramme (Buchanan-Smith et al., 1995), theprogramme achieved its primary objective in thatcash aid alleviated the hunger of individual tarims,and often whole households. Tarims living alonewere able to live off the food they bought with thecash payment for three to four months. The grantalso reduced the pressure on family food resourcesand, in particular, on women responsible forfeeding tarims in their household. Most recipientsspent at least half the money on food immediately.Some male recipients bought livestock or investedin other income-generating activities, whereas allof the female tarims spent the entire cash grant onfood. All recipients claimed they had kept controlof the cash, although food purchased had benefitedthe entire household.

ActionAid’s evaluation of its programme reportedthat most female tarims rated the value of a cashgrant as equal to that of the grain bank, but thattheir male counterparts saw cash as more valuablethan grain. The three main benefits of cash overgrain distribution for the tarims were:

(i) cash allowed the tarims to purchase othersupplementary food items;

(ii) cash enabled the tarims to purchase grain whenand as they needed;

(iii) the tarims required cash to have their grainground into flour.

Another benefit of the cash distribution was theempowering effect it had on the beneficiaries.Firstly, as recipients of income tarims were able tocontribute to household security. Secondly, theywere given the power to control money and theautonomy to decide how to spend it. With increasedpurchasing power, tarims were able to participateto a greater extent in community life, and to do sowith greater dignity, for example as buyers atmarkets and grain banks.

As in the Ethiopia case, there is little evidence tosuggest that the cash distribution improved long-term food security in the region. Nonetheless, theprogramme achieved its primary objective, whichwas to alleviate the hunger of the most vulnerablegroups through one-off income transfers. In theprocess, it enabled the recipients to determine theirown needs and priorities.

Cash distribution and rehabilitationThe Red Cross in Guatemala and Nicaragua,1998The Red Cross Hurricane Mitch AgriculturalSupport Programme is one example of the use ofcash in rehabilitation. The programme aimed toassist in the agricultural recovery of small farmersin El Salvador, Guatemala, Honduras and Nicaraguafollowing the damage caused by Hurricane Mitchin October 1998. Cash payments were included inGuatemala and Nicaragua as one element of arehabilitation package comprising seed, fertiliser,grain and spray pumps. The aim of the cashcomponent was to allow beneficiaries to buycomplementary inputs not included in the package,and/or to purchase food in areas where it was inshort supply.

Cash grants of $30 were delivered to around 17,000families in Guatemala and Nicaragua. Some familiesalso received food aid. In all, cash payments worth

8


a total of around $490,000 were distributed over afive-week period. These payments were alwaysmade to women, while the agricultural inputs weregiven to men. (Cash was not distr ibuted inHonduras or El Salvador for fear that it would createmanagement and security problems; beneficiarieswould not use cash for the intended purpose; andit would encourage dependency.)

According to the evaluation of the programme(British Red Cross, 1999), fears that beneficiariesin Nicaragua and Guatemala would misuse the cashproved groundless. The evaluation team attributedthis to the overriding importance of agriculturalproduction as a source of income. The team believedthat, since cash was part of a package clearly directedat agricultural investment, beneficiaries felt ‘morallyand practically obliged’ to spend cash in the wayintended by those giving it. It was also felt thatdistributing cash to women rather than men helpedto prevent its diversion. The fact that cash wasdelivered just prior to land preparation, when theneed to pay for labour and inputs was at its height,encouraged expenditure on those items.

The evaluation report states that the package waspopular with beneficiaries, most of whom stated apreference for the mixed package over a cash-onlydistribution. The extent to which a cash-onlyintervention (or even a larger injection of cash)would have led to immediate increases in the prices

of inputs, as suggested by the evaluation report,remains unknown. The report also points out thatseed did not seem to be widely available on themarket – or at least that seed was bought or acquiredfrom other sources. The only problem reportedwith the cash distributions was administrative, inthat the beneficiary name on the cheque issued atthe bank did not always coincide with that on theidentity card.

In areas where food aid was not distributed, thecash grant was spent mainly on food. In other areas,cash tended to be spent on inputs not included inthe package, such as labour for preparing land; rentfor land, oxen or tractors; and the purchase of moreseed and agrochemical products, and chickens, pigsor tools. In other words, where food aid wasdistributed, it was an important factor in enablingbeneficiaries to invest the cash they received ininputs and productive assets.

Cash distribution and conflictThe Kosovo crisisIn quite different circumstances, a cash-grant projectof nearly $8m was initiated by the UN HighCommissioner for Refugees (UNHCR) in Albaniaduring the Kosovo crisis in 1999. Grants wereaimed at Albanian families who were hostingKosovar refugees. Some two-thirds of this refugeepopulation, estimated at 285,000 people, werebeing accommodated in Albanian homes.

Box 4: Cash distribution in Bangladesh, 1998

A cash-grant project was included as part of Save the Children Fund (UK)’s emergency response to thefloods in Bangladesh in 1998. One-off payments ranging from 500 to 2,000 taka were made to 6,800families in six different thanas or sub-districts (SCF(UK), 1998). As an indication of the relative value ofthese cash grants, 70 per cent of households in an SCF(UK) survey of the affected area had a monthlyincome of less than 2,000 taka (Hossain and Shuaib, 1998). In addition, around 300 working children,whose income had been affected by the floods, received 450 taka each. As with SCF(UK)’s other interventions,the cash grant project was run by the agency’s partners, and operated in areas where these partners werealready working.

The aims of the cash distributions varied. In one sub-district, cash directly replaced food distributions onthe grounds that local markets were working sufficiently well to enable families to buy food themselves.Here, the value of the cash grant seems to have been calculated in terms of its food value, equating to a1,900 kcal ration (Hossain and Shuaib, 1998a). These grants were targeted at the most vulnerable households:those headed by women, or with a disabled person, or with malnourished children. In other areas, cashgrants were given to assist rehabilitation by helping beneficiaries to repay loans, or to replenish assets soldin the weeks following the floods. Another stated aim in giving cash was to let the beneficiaries themselvesdecide what they needed (SCF(UK), 1998).

On the whole, the programme seems to have been a success, despite problems with the identification ofbeneficiaries in some sub-districts, which meant a slightly smaller number of families took part in theproject than was originally planned. However, it is difficult to judge the effectiveness of this (relativelysmall) project independently of the many other strands of the flood-relief programme.

9


UNHCR considered cash distribution to be themost effective way of supporting the widely-dispersed families, particularly given the logisticaldifficulties involved in distributing food or non-food items (UNHCR, 2000). Host families received$10 per refugee per month, backdated to thebeginning of the refugee influx. The main problemwith the project seems to have been delayedpayments (UNHCR, 2000), which meant thatmoney was not received until after the refugeeshad been repatriated. However, distributing foodand non-food items proved so difficult in terms oflogistics and registration that many host familiesand refugees received no assistance at all.

Logistical challenges aside, food or non-food itemsmay have been more appropriate than cash,especially for poorer host families living in a largelynon-cash economy. However, most refugees werehosted on a commercial basis: they paid their hostfamilies for accommodation and food using theirown money. These economic ar rangements,combined with the generosity of the Albanians,were cited in the UNHCR evaluation report as‘arguably the most significant factors in avoiding apotential crisis’ (UNHCR, 2000).

In Kosovo itself, WFP began providing food aid inmid-1999. Aid covered the needs of the ruralmajor ity Albanian population, urban foodrequirements and the needs of Kosovo’s minorities.In its first three months of operation, blanket foodaid covered some 80 per cent of the 2m-strongpopulation, though by the following Octoberbeneficiary numbers had fallen to 600,000. Thenumber of beneficiaries has continued to declinein line with the gradual scaling back of relief, andthe emergency programme was expected to bephased out completely by early 2002.

This scaling back of aid has coincided with therecovery of the harvest sufficient to meet the majorityof food needs in rural areas, growth in the urbaneconomy and a rise in wage levels. In addition,the UN Interim Mission in Kosovo (UNMIK),effectively the province’s government, has begun asocial-assistance scheme to provide cash to some350,000 especially-vulnerable people. Targetgroups fall into two categories: the unemployableand their dependants; and the very poorest of theprovince’s population. Those eligible in categoryone include the permanently disabled, householdswith no able-bodied adult, the elderly and single-parent families. To be eligible for assistance as partof the second category, beneficiaries must own noproperty, and have no access to any other form ofassistance, either within or outside Kosovo.

Cash payments under the scheme began in August2000. The categories broadly covered those peopletargeted by food aid. The scheme has thus provideda natural link for the final phasing out of food aid,although assistance was still deemed necessary inthe short term to provide some form of food safety-net for beneficiaries who had not been registeredfor the cash scheme, and for marginal cases who,while strictly ineligible, were nonetheless in need.These included people whose land had beenmined, and was hence unusable, as well asminorities and returnees, totalling an estimated100,000 people. In effect, short-term emergencyassistance was giving way to a long-term, cash-based social-welfare system.

Cash distribution and socialinterventionsMozambique, Namibia and ZambiaIn the 1990s, a number of ‘safety-net’ programmeswere established in Africa. Three – in Mozambique,Namibia and Zambia – were assessed as part of aresearch project funded by ESCOR, a componentof the UK’s Department for InternationalDevelopment (DFID). The programmes were:

• a cash-transfer programme called GAPVU(Gabinete de Apoio à População Vulneravel) in14 towns in Mozambique targeted at war-affected and disabled people, who received theequivalent of £16 per person per year (around$22);

• a social pension in Namibia, representing theequivalent of £192 per person per year (about$270), provided to all citizens over 60 years ofage; and

• a CFW project in the drought-affected Westernprovince in Zambia, in which earnings variedby district, according in part to the number ofparticipants.

Devereux (2000b and 2000c) suggests that socialsafety nets can be viewed in terms of Sen’sentitlement theory, and as a means of protectingentitlement following a livelihood shock such asdrought (as in the Zambian case), marketliberalisation or war (Mozambique), or retirement(Namibia).

The initial aim of the GAPVU project was to protectthe urban destitute against rapid rises in food pricesresulting from the war, drought and liberalisationof agricultural markets. As a result, it has alwaysbeen described officially as a ‘food subsidy’. In itsearly years, it faced a number of problems, including

10


serious corruption and fraud which caused thetemporary closure of the programme in 1996. (Theproject was relaunched in 1997 under a new nameand with new management and monitoring systems.)

Another problem arose from the project’s statedobjective, which was to improve the nutritionalstatus of the market-dependent urban poor; as aresult, success had to be measured in terms ofnutritional improvement, which was consistentlynot found. An evaluation of expenditure amongGAPVU beneficiaries supports the evidence fromother transfer programmes that even among thedestitute, transfer income is surprisingly fungible,with food being only one of several priorities.Devereux (2000b) notes that ‘the ingenuity withwhich the poor made this tiny amount of cash gofurther is remarkable’. Over half of respondents inthe expenditure survey said they had purchasedone or more assets using GAPVU income. Thatmost of these were consumption goods – such asclothing or kitchen utensils – rather than productiveassets such as farm implements or education waslargely due to the very low value of the cash payments.

This was exacerbated by the failure to take intoaccount rises in living costs, despite an undertakingat the start of the project to keep levels of paymentin line with increases in the national minimumwage. Despite GAPVU’s low value, the ESCOR-funded survey found that it contributed as muchas 25 per cent of cash income in beneficiaryhouseholds in 1998, and more in the smaller, poorertowns. Devereux (2000b) concludes that ‘GAPVUwas a small transfer that made a small but significantdifference to the livelihoods of its poor recipients’.

In contrast, the relatively high level of wages inZambia’s CFW programme allowed for more

investment in assets, particularly for the richerhouseholds. After the first year of the project –during which people were recovering from asevere drought – the proportion of cash spent onfood declined dramatically, with more going on‘house-related’ items and clothing. ‘Consumptionspending’ was found to be high relative to‘investment spending’, which Devereux surmisesmay have been due to the greater accessibility ofclothes and kitchen utensils, as against fertiliser andseeds, alongside the work sites where traders foundbusiness to be good.

By way of a ‘trickle-down effect’, certain groupsof non-participants also benefited from theparticipants’ wages: these Devereux calls ‘secondarybeneficiaries’, including sub-contracted replacementworkers, ox-cart owners, traders and agriculturallabourers. In addition, CFW income stimulatedlocal trade, attracting new shops and stalls incommunities adjacent to the work sites. Some ofthese were set up by beneficiaries themselves, usingsavings from their wages.

Summarising the expenditure patterns in the threesafety-net programmes, Devereux notes that low-value payments tended to be spent on food andclothes, while higher-value payments allowed forsome investment – firstly in ‘human capital’(education and health) and ‘social capital’ (increasedassistance to relatives), and then in income-generating activities and assets (agriculture andlivestock). Devereux (2000c) gives three preconditionsfor successful safety-net interventions: adequatelyfunctioning markets; fiscal and political sustain-ability; and investment in monitoring systems. Buthe concludes that ‘cash transfers have an enormousimpact on the livelihoods of the poor, and on thelocal economies in which they are situated’.

11


The key question to ask of cash distributions inemergencies must be whether they achieve thesame objectives as food distr ibutions, namelyincreasing the access to food of those in need. Aslong as distr ibution does not involveinsurmountable administrative challenges, this isprimarily a matter for the market, as the success ofcash support is ‘dependent on an adequateresponse from the market in the form of meetingthe demands generated’ (Drèze and Sen, 1989).Does increasing a population’s purchasing powergenerate a sufficient inflow of food at a reasonableprice so that beneficiaries are able to purchase theirrequirements? In short, do cash interventions work?

A mixed pictureThe largest and best-documented programme, theCFF project in Ethiopia, presents a mixed answerto this question. At almost all the sites, beneficiarieswere reported to have been able to purchase theirfood requirements, and often non-food items aswell. Migration was halted and even reversed, andevacuated children returned to their families. Yetat many of the same sites, notably during the worstof the famine in 1984–85, the food value of thepayments declined substantially due to widespreadincreases in the price of grain (not, it should benoted, caused by the cash payments themselves).Beneficiaries at some sites also reported a shortageof grain on the market.

Overall, the Ethiopia example suggests that, for mostof the project areas, the cash payments boosted theinflow of grain from the neighbouring surplusareas. It also indicates that the food value of thesepayments, particularly in the critical period of 1984–85, fell far short of the intended full ration. Thatthe project was nevertheless such a success perhapsowes something to the careful selection of the sites,which did not necessarily include the most severely

affected areas, but rather areas that met the criteriafor selection. One important criterion was, bynecessity, the ‘availability of surplus productionwithin a reasonable distance of the affectedpopulation’ (UNICEF, 1984).

The Ethiopia experience highlights how cashdistribution can work in a country of poorly-integrated markets, and experiencing severe foodshortage. In a country such as Botswana, by contrast,there is a ‘widespread and competitive retailnetwork operating in all but the remote areas’, andthe free exchange of food with neighbouringcountries tends to smooth out in-country pricedifferences (Drèze and Sen, 1989). Here, the marketwould be quick to respond to an increase inpurchasing power within a particular section of thepopulation. This applies equally to other countriesin southern Africa, although the cash-for-workproject in Zambia provides evidence of inflation(particularly of food pr ices) caused by the‘substantial injection’ of cash into three very poordistricts (Devereux, 2000c). Project participantsargued that traders were exploiting the increase ineffective demand, while non-participants foundthat, with no increased income, inflation put themat a particular disadvantage.

On the other hand, the acute food insecurity thatZimbabweans suffered in the drought of 1991–92was contained largely through their purchase offood on the market, made possible through large-scale commercial imports and an effective marketsystem, rather than through food-aid distributions,which were inadequate and arrived too late. Butthe overriding need for cash for food led to distresslivestock sales and the use of cash reserves, andmeant that, by the end of the crisis, an already-poor population had become almost criticallyimpoverished.

3

Assessing the effectivenessof cash-based responses

12


Drèze and Sen also descr ibe the very largeprogramme of public works set up in the Indianstate of Maharashtra during the famine of 1972–73to provide cash income to the drought-affectedpopulation. The income generated on these reliefworks attracted food from neighbouring states viaprivate trade, despite severe restrictions imposedon the movement of food between states. Thisprivately-traded food proved a crucial complementto government sales which, while playing animportant role in relief provision, ‘fell far short ofexpectations’ (Drèze and Sen, 1989).

Certainly, the case studies of food insecurityoutlined above – in Sudan, Ghana and Bangladesh– suggest that beneficiaries had no difficulty inbuying food in the market at reasonable prices. Itis often assumed that local markets lack the capacityto respond to a sudden increase in demand.However, the evidence suggests that, in most cases,deregulated markets respond to demand as longas food is available: ‘Even in regions endangeredby drought, there are private enterprise systemswhich show an astonishing flexibility in providinga supply of food wherever there is sufficientpurchasing power’ (Schubert, 1987). Moreover,there can be positive benefits to stimulating localtrade; the income generated can be used to pay offdebts incurred during the period of shortage, forinstance.

All the sites of the case studies were carefullyselected on the basis that marketable grain wasavailable in the vicinity. Despite Sen’s insistencethat aggregate food supply in Ethiopia in 1983–85was more-or-less adequate, it is hard to imaginethat the injection of large amounts of cash into thefamine epicentre in Wollo would have induced asufficient food inflow at sufficiently low prices toavert starvation.

Keen (1992) gives two examples where cashdistributions among refugees would not havegenerated the required market response: whererefugees had settled in an isolated area that wasalready experiencing a food shortage (Chadianrefugees in western Sudan); and where a largeinflux of refugees had settled among a populationthat was already finding it difficult to get enoughfood (Mozambican refugees in Malawi). Similarly,the inclusion of cash in the agricultural-supportpackage that followed Hurricane Mitch, and itspopularity among beneficiaries, depended on theexistence of an adequately-functioning system ofmarkets through which cash could be exchangedfor the required commodities – seeds, tools andlabour.

The benefits of a cash-based responseThe case studies highlight a number of benefitsthat cash aid has over traditional food distribution.

SpeedPerhaps the chief benefit of cash over food aid isthe fact that delivering cash is much quicker thandelivering food. Cash is logistically simpler, andcan be transported and disbursed rapidly, even toremote locations. While food distribution tends tobe centralised, raising the r isk of populationdisplacement or settlement around distributionsites, cash distribution is typically decentralised.While not without problems of its own, thispresents potential opportunities for better

Box 5: The role and capacity of local traders

The success of cash distributions dependssignificantly on the quantity and quality of servicesupplied by traders. There is a risk that traderswill not have sufficient capacity to meet thesudden increase in demand, or governmentregulations might make it more difficult for themto operate. During the drought in Kenya in 1984,for example, private traders were unable to supplydrought-affected areas because of governmentrestrictions imposed on inter-regional trade (Drèzeand Sen, 1989). Similarly, government regulationsrestricted market integration in Ethiopia’ssouthern regions. In such situations, cashdistribution would not necessarily have increasedfood supply.

Traders are likely to respond with caution to thecreation of new markets. There are a number ofcosts involved in reorienting distributionchannels, and these may deter some traders fromsupplying famine markets. Firstly, traders mustswitch from a sector where the demand is alreadyknown and regularised to one where it isunknown, and where the market may be new ordormant. These uncertainties increase the risk,and mean that high profits must be guaranteed.Secondly, traders will fear the artificial andtemporary nature of this new market, and questionwhether the demand will be sustained when‘normality’ returns. The opportunity cost of losingregular customers may simply be too great.

There is also the risk that, in marketscharacterised by monopolistic control exercisedby a very few traders, artificially high food pricescan be set, and the system exploited to the benefitof traders’ profit margins.

13


household targeting. One of the advantages of theCFF programme was that payments weredistributed near to beneficiaries’ homes, rather thanseveral miles away at distribution sites (see, forexample, UNICEF (1988a)).

CostThe transaction costs of cash distr ibutions aresubstantially lower than those of food. Overheadcosts associated with food aid (handling, transport,storage and administration) on average range from30 per cent to 50 per cent of the total aid provided.A Food and Agriculture Organisation (FAO) studyon the impact of WFP food aid in Ethiopiaestimated that 34 per cent of the cost of food aidwas needed for transport (UNICEF, 1988a);similarly, approximately 40 per cent of UNHCR’soperational budget in Africa is spent on transportingrelief commodities (Keen, 1992). The comparativecost analysis of the Ethiopia CFF project showedthat the cost of cash disbursement wasapproximately half that of the equivalent food-aiddelivery. It was also noted that inland distributionof food aid to famine pockets – which willprobably always be the main target of cashdistributions, as they were in the CFF programme– is relatively more expensive than bulkdistr ibution to major drought-affected areas(UNICEF, 1988a).

The informal monetisation of food aid bybeneficiaries themselves is also expensive due totransaction costs; it is estimated that the value offood rations sold is about three times less than theoriginal cost price (Bryson and Hansch, 1993). Infact, it is often the case that the local value of foodrations is very low in relation to the cost of living,or the value of assets. Keen (1992) identifies thisproblem in the case of Mozambican refugees inMalawi, where the cost of firewood for cookingbeans was five times the value of the beans. Thepoor returns from selling food suggest thatmonetisation does not always constitute the sametransfer value as direct cash distribution.

Cash, on the other hand, has minimal transactioncosts. As a result, the unit cost price per beneficiaryis significantly lower than it is for food. Thetransaction costs saved could be used to make cash-disbursement programmes more effective, forinstance by allocating greater resources to targeting,monitoring and supervision, which are identifiedas areas of weakness (see, for example, UNICEF(1988a); Devereux (2000b); SCF(UK) (1998)).Moreover, less overhead could mean beneficiariesreceive a greater overall proportion of the moneydonated. An analysis of costs in Geramider in

northern Shoa, Ethiopia, estimated thatbeneficiaries received nearly 90 per cent of the costsof the cash distribution; in an equivalent fooddistr ibution, the value of food received bybeneficiaries would have been around 35 per centof the total costs (UNICEF, 1988a).

The impact on local markets and tradeIt is also important to consider how a cash-basedresponse benefits local markets and trade. Theincrease in demand attracts food sellers, and linksfood-deficit areas with food-surplus ones, sincetraders tend to shift food from low- to high-priceareas. In this respect, cash disbursement can activateor regenerate local markets and encourage tradingto more remote locations. Market integration islikely to lead to a moderation of food prices indeficit areas, and an increase in food prices insurplus areas, an effect that Drèze and Sen (1989)describe as ‘the sharing of distress over a widerarea’. It is possible, therefore, that cash transferscan help to reduce price disparities and encouragethe restoration of market equilibrium. Furthermore,the potential disincentive effects of food aid onagricultural production are avoided, and theincrease in market demand is likely to discouragehoarding.

Local markets can also profit from the ‘multipliereffect’ of a cash injection, as cash circulates morequickly and for a longer time than food transfers.In the case of the safety-net projects in southernAfrica, ‘secondary beneficiaries of income transfers… included local traders and others who benefitedfrom income multipliers generated by spendingand redistribution of transfer income’ (Devereux,2000c). In other words, cash benefits both the buyerand the seller. In the case of a food transfer,however, the recipient or household consumes, andthe benefit ends there. In addition, cash does nothave a limited shelf-life, as does food, and can bebanked or invested in numerous ways. As Wilson(1991) observes: ‘Given the highly dynamic natureof most Afr ican rural economies, it is notunreasonable to expect that financial injections intothe economy might have marked payoffs to anexisting process of economic growth.’

Identifying requirements and consumptionpatternsThe cash approach also addresses the problem ofascertaining the appropr iate and necessaryrequirements of the population, particularly innon-famine situations. Beneficiaries are in a betterposition to determine their individual needs thanare relief agencies or donors, and the ability to dothis represents a fundamental step towards

14


empowerment. Cash not only gives the recipientthe right to choose, but also enables immediateaccess to markets and economic activity, which isparticularly important for those who do notnormally enjoy any sort of economic status: tarimsin Ghana, for example, or pensioners in Namibia.

In terms of food, cash payments enable bene-ficiaries to avoid the particular kinds of consump-tion patterns determined by donors. General foodrations constitute an unpalatable diet of low micro-nutrient content; as a result, many householdsobtain supplementary food through the sale ofsome of their ration (leading to the transaction costsdiscussed above), or through economic activity. Incontrast, recipients of cash can purchase what theyneed, when they need it. Such flexibility is alsolikely to result in improved nutritional balance sinceit enables the purchase of more varied foods.

The impact on livelihood securityFinally, a cash-based response can be directly linkedto livelihood security, whereas food is essentiallylinked to food security. As Drèze and Sen pointout, whether beneficiaries receive food or cashmatters less in the case of famine victims, for whomfood intake is an overriding concern. But thechoice of food or cash as the medium ofentitlement becomes more relevant in non-faminesituations, in which the root problem is often notso much one of food insecurity as one of poverty.Food aid is nevertheless provided in many cases,in the face of a need that it is not designed to satisfy.As a result, it is frequently monetised to meet cashdemands for livelihood security, such as clothing,soap and transport. How beneficiaries spent theircash in the case studies described here confirmsthat food is not the only concern of households,even when faced with acute food insecurity. Theflexible and fungible nature of cash means that itcan be used directly to meet other livelihooddemands at the early stages of famine, or in post-disaster situations.

The risks of a cash-based responseThe evidence also points to a number of potentialdisadvantages in a cash-based response, as againsta traditional food-aid intervention.

Misuse and misappropriationThe most apparent r isk in cash distr ibutionconcerns the very flexibility and fungibility ofmoney. This is both a strength and a weakness. Onthe one hand, fungibility can be seen as anopportunity since it allows beneficiar ies todetermine their needs, and in a ‘cash-for-food’programme, for example, payment of debts or asset

reconstruction may be as important as buying food.On the other hand, fungibility means cash canalways be put to a use for which it was not intended:non-food consumption, for example, in aprogramme that aims to address nutr itionaldeficiencies. It is often assumed that cash is likelyto be diverted into activities that do not benefitthe whole household and, in particular, squanderedby men on alcohol or gambling, although thisassumption may have more to do with thepreconceptions of donors than with the actualpriorities of those in need.

The case-studies confirm that cash is certainly usedfor non-food consumption in situations of acutefood insecur ity (see, most notably, the CFFexperience in Ethiopia). But they give no empiricalevidence of widespread squandering of cash infamines, or in post-disaster situations. Thewidespread purchase of non-food items in the CFFprogramme in Ethiopia may have been surprising,but at many sites it meant that beneficiaries were‘undoubtedly in a far better situation to face thefuture than they were at the commencement ofthe project’ (UNICEF, 1984). In the southern Africasafety-net projects, Devereux identifies three waysin which beneficiar ies used cash to enhancehousehold food security: food purchase, investmentin foodcrop farming, and as working capital ininformal activities such as petty trading. Devereuxnotes (again in the context of safety nets rather thanemergency interventions) that ‘income transferswill impact on productive investment only if theyare large enough to cover consumption needs’. Healso highlights the dilemma for policy-makers:whether to divide resources among as many of thepoor as possible, with minimal impact on liveli-hoods, or to concentrate on fewer beneficiaries, togreater medium- or long-term effect (Devereux,2000c).

That beneficiaries use cash to satisfy consumptionrequirements first is also reflected in the HurricaneMitch example, in which productive investmenttended only to be possible where cash did not haveto be spent on food – that is, where food aid wasprovided. In the Hurricane Mitch case, the use ofcash for food and agricultural investment may havebeen encouraged by certain ‘signals’ built into theprogramme: the inclusion of cash within anagricultural package, the timing of the distribution,and the distribution of cash to women (British RedCross, 1999).

A sudden increase in income through cash transferdoes not, however, necessarily imply a boost inpurchasing power. In a Cash-For-Work inter-

15


vention in southern Ethiopia, SOS Sahel found thatthe majority of households preferred food to cashsince they feared that they would be forced to repaytheir debts if given money (Jenden, 1995). In theworst case, cash can be diverted and used for anti-social activities, such as buying weapons.Undoubtedly, in situations of war-induced faminethere is a clear danger that cash may bemisappropriated by warring parties, contributingto the political economy of war.

It can also be argued that misallocations of cash aremore easily hidden than with food or non-foodcommodities since cash is less visible; and that cashis therefore more vulnerable to diversion andcorruption. This makes good monitor ing andaccounting systems even more important than withfood distributions, and is perhaps why these twoareas were cited as the main focus of problems inmore than one case study (see, for example,UNICEF (1988); Devereux (2000c)). In Mozam-bique, ironically, inadequate monitor ing andsupervision systems were a result of donor concernsto maximise the funds transferred to beneficiaries,and resulted in massive leakage and diversion ofcash (Devereux, 2000b). But both Keen (1992) andWilson (1991) argue, albeit with regard to cashdistributions to refugees, that accountable systemscan be created if the will to do so exists. Certainly,the savings on transport and storage relative to fooddistributions suggest that more resources couldpotentially be allocated to monitor ing andaccounting.

TargetingTargeting also becomes a more contentious issuein the case of cash distr ibutions, since cash isinherently of value to everyone. Cash distributioninevitably invites more stakeholders, and has thepotential to further marginalise politically weakerand vulnerable groups. Direct cash transfers do not,moreover, lend themselves to targeting by self-selection, as do CFW programmes in which a levelof payment can be set that is sufficiently low toattract only the poorest (see Drèze and Sen, 1989).However, while targeting can undoubtedly be moredifficult for cash distributions, it does not appearto have arisen as a fundamental problem in thecase studies. This may have been due, in theHurricane Mitch case, to its inclusion in anagricultural package; or, in Ethiopia, to the emphasison geographical selection; or, in the Ghana andBangladesh examples, to implementation at arelatively local level, on a relatively small scale.

Perhaps more seriously, the very fact of targetinghas implications for cash distributions that do not

apply to distributions of food. Food distributioncan bring down food prices, thus benefiting peoplewho have not been targeted, but who have topurchase their food. Conversely, if not all thevulnerable population is included in the cashdistribution, those that have been excluded willstill have to deal with the consequences of any priceincreases arising from the injection of cash. Thus,the overall vulnerability of the population couldbe exacerbated, rather than diminished (Drèze andSen, 1989). This fear of sudden inflation and itseffects can deter agencies from distributing cash.The Zambia cash-for-work project provides someevidence that food prices can rise sharply followinga large injection of cash. The only other case studyto monitor local prices before and after a cashdistribution – the CFF project in Ethiopia – foundthat they increased. But it seems that this was aresult of widespread food shortages, rather thanthe cash transfers themselves. Two factorsdetermining the effect of a cash injection on localpr ices must surely be the overall size of thedistribution, and the size of the population or localeconomy in which it is disbursed. In other words,is there a threshold below which the inflationaryeffect is minimal, but above which prices tend torise rapidly? Apart from the effect on beneficiariesand non-beneficiaries, ‘inflation offsets the incometransferred, so the total intervention effect may beconsiderably less than the hard currency value ofthe cash’ (Bryson and Hansch, 1993).

There is also a risk that women will not gain accessto purchasing power since, in many societies, mentraditionally dominate control of money in thehousehold, and may not spend it in ways thatbenefit the whole family. Even transferring incometo women, as was done in the Hurricane Mitchintervention, may not necessar ily confer theequivalent purchasing power if they are obliged tohand over the cash to their husbands, or if they areanyway restricted in what they can buy or own.Food, on the other hand, is easier to target andmore likely to benefit women, children and theelderly. In Zimbabwe’s drought-relief programmesof the mid-1980s, most women argued for foodassistance, while most men argued for cash (Keen,1992). That the most vulnerable are often unableto participate in purchasing and economic activitycan represent a strong argument for food transfersover cash. Gender is an even more complexconsideration in safety-net cash transfers; asDevereux (2000c) points out: ‘Southern Africanwomen are restricted in the assets they can own,so income transfers are more likely to be consumedthan used to finance independent wealthaccumulation.’

16


Economic impactsInjections of hard currency are also likely to affectthe local exchange rate, particularly if an unofficialrate exists. If an agency decides to disburse USdollars, which is by far the most common hardcurrency used for economic transactions in reliefoperations, it risks establishing a dual economy.Davies (1996) explains the economic implications:

the unintended effect of the massive influx of dollarsis the creation of an environment conducive to moneylaundering. The inevitable result is that theAuthorities lose what little control they might havehad over monetary policy and their economicmanagement becomes, as a consequence, even moredifficult. The net effect is usually an inappropriateexchange rate with subsequent negative effects onthe profitability of the export sector, once it is revived.

The quantity, quality and promptness of supply alsoinfluence the effectiveness of cash distribution. Anydelay between the distribution of cash and themarket supply of food will have negative effects. Asurplus of cash without immediate food supply mayencourage other types of purchasing. Surplus cashin the household may be seen as a security hazard,forcing quick spending for fear of theft.Furthermore, in a famine region cash may devaluerapidly, to the extent that the final income transferat the point of trade may be considerably less thanat the time of distribution.

SecuritySecurity risks also face the agencies distributingthe cash. The idea of transporting and distributingthousands of dollars in cash must be a frighteningprospect for any field officer; it was partly due tosecurity fears that cash was not distributed in theHurricane Mitch agricultural-support packages inEl Salvador and Honduras (British Red Cross,1999). The nature of the political environment canalso increase the security dangers involved in cashdistribution. For example, assisting populationswith cash during a conflict-related emergency maymake them a political or military target of warringparties.

MobilityA risk for governments to consider concernsmobility of people. Since money may be spent ontransport, the authorities may fear that cashdistribution will encourage migration or, at theleast, increase the mobility capacity of ruralcommunities. Increased mobility and access couldbe seen as a threat to state control. While this maysometimes be so, no beneficiary in the case studieswas reported to have used their cash in this way.

On the contrary, the cash distributions in, forexample, Tatch Gaint in Ethiopia enabled migrantsto return home, and allowed the return of childrenwhose parents had sent them away at an earlierstage (UNICEF, 1984). The mobility question isheightened in the case of refugee populationswhere host governments are reluctant to allowrefugees to move freely. Traditional fooddistribution, on the other hand, ensures control ofpopulation movements around designateddistribution centres.

Balancing benefits and risksGiven that there are both benefits and r isksassociated with cash distribution, the decision todistribute cash or food depends on an assessmentof whether the good outweighs the bad. In certainsituations, the risks will be too great to justify thedistribution of cash; for example, if there is a dangerof creating excessive demand, or of not being ableto target women and children effectively.Elsewhere, however, the opportunities mayoutweigh the risks. There can, of course, be nobluepr int policy on using cash; the decisionrequires a case-by-case approach, calculating thetrade-off at the local level through ongoing marketanalysis and food-security monitoring. A starting-point could be to juxtapose the pros and cons offood and cash aid, as outlined in Table 1.

Agencies need to consider two key variables. Thefirst concerns accessibility. The experiences to datein cash distribution suggest that this type of responseis appropriate in famine ‘pockets’ in close proximityto surplus-producing areas. If famine-affected areasare isolated and inaccessible, a sudden boost inpurchasing power is unlikely to increase foodsupply. Furthermore, while markets may beaccessible geographically and logistically, tradersmay not necessarily be assured access, since thiscan be determined as much by market controlsand political imperatives. It is equally important toconsider the size and demand of the famine market,which must be great enough to justify tradersreorienting their distribution channels. Furthermore,the spread of markets will limit the shocks causedby a dramatic cash injection.

The other key variable is regional or national foodavailability. Determining this involves analysingfood production, regionally and nationally; ananalysis of food in storage and on the market; andan analysis of pr ice trends and consumptionpatterns. It also requires a good knowledge of theexisting trade and retail structures and the targetpopulation. A good example of the understandingneeded is reflected in the plans to phase out WFP

17


food aid in Kosovo and replace it with cash aid:cash payments were to start only once food avail-ability had increased after the harvest, and food aidwas to continue for certain groups, including Serbs,who did not have access to the market. In thisrespect, an effective cash intervention may initiallyrequire a more cautious approach than a blanketfood-aid programme, but the time invested may bewell spent in terms of impact and cost-effectiveness.

While an adequate food supply is a prerequisitefor an effective cash-based response, themechanisms by which that food supply can beassured do not necessarily all have to be market-driven. As Drèze and Sen (1989) point out:

To recommend greater use of cash support is not tosuggest that importing food into famine affectedcountries or areas is undesirable or unnecessary. Cashsupport and food supply management are not, byany means, mutually exclusive activities … Ourcontention is not that cash support should replaceefforts to improve food availability, but only that in

Table 1: Food versus cash

Source: Adapted from Bryson and Hansch, 1993

Cash

Advantages

More cost-efficient

Transfer costs minimal

More easily converted (fungible)

Allows beneficiary choice

Encourages productivity and stimulates markets

Disadvantages

Limited donor budget for this kind of intervention

Losses from inflation

Losses from leakage

More difficult to target – usage favours men

Can be used for non-food consumption and anti-social activities

Security risk

Food

Advantages

Greater donor surplus

Immediately increases food availability (althoughthere may be delays in delivery)

Directly addresses nutritional deficiencies

Can be self-targeting

Favours women, children and the elderly

Disadvantages

Transport and storage costs high

Administrative costs high

Losses from spoilage and theft

Less easily exchanged

Disincentive effects on production

Competes with local markets and trade

many circumstances there is no need to makeentitlement protection conditional on the directdelivery of food.

The same principle could be applied to the deliveryof non-food items in rehabilitation programmes.Devereux (2000b) suggests that: ‘Althoughparticipants cannot of course be forced to purchase“investment goods” [such as seeds, fertiliser orlivestock] rather than consumer items, futureinterventions might consider ways of making suchcommodities more readily available or moreattractive to participants.’

Cash-based support does not preclude thesimultaneous implementation of otherinterventions. Cash as a complement to fooddistr ibution, for example, may be the mostappropriate response where agencies or donors fearrapid increases in food prices and/or inadequatesupply, or where they wish to see the cash used forproductive investment and livelihood recovery,rather than food consumption alone. Similarly, the

18


inclusion of cash as one element within arehabilitation package, as in the Hurricane Mitchagricultural programme, may be one way of makingthe most of the benefits of cash delivery – including

the speed and low cost of implementation, itsflexibility for beneficiar ies and the secondarybenefits – while ensuring that its use accords withprogramme objectives.

19


Agencies seem increasingly willing to use cash as acomplement to traditional forms of relief andrehabilitation. While this is due in part to thedeclining availability of food aid, it also reflects anattempt by agencies and donors to take into accountthe increasing importance of cash in even thepoorest rural economies. This has been reinforcedover the past decade by the need to consider formsof relief and rehabilitation in Western, industrialisedeconomies that have long been cash-based. As aresult, the term ‘food insecurity’ has become anincreasingly fragile analytical concept; instead,‘livelihood security’ has been the focus of muchdonor thinking in recent years.

The case studies reviewed in this paper illustratethe implicit use of a livelihood framework in manycash-transfer programmes over the past 20 years.They are limited to cash distributions in naturaldisasters, or in situations of at least relative stability;none covers distribution in an ongoing conflict orother complex political emergency. Nonetheless,these studies offer a number of indications, bothpositive and negative. The ‘cash-for-food’ projectsin Ethiopia and western Sudan, for example, lendqualified support to the idea that during a famine,cash distributions to populations in pockets ofdeficit close to surplus-producing areas can bringabout an inflow of food at reasonable prices. TheEthiopia case in particular also provides solidevidence of the advantages of cash over food aid interms of cost and speed. The Hurricane Mitchagricultural programme, which had a quite differentobjective, shows how cash can be included to goodeffect in a mixed agricultural-recovery package. Thisexample underlines the importance of consideringcash as one element of a package, or as one strandof a programme, rather than as an option that excludesall others. The southern Africa safety-net studiesindicate that beneficiaries tend to use cash for social

and productive investment only after consumptionneeds have been met, and show how cash can actas a stimulant to the local economy. Evidence of‘squandering’ – on alcohol or gambling, forexample – was not found in any of the case studiesthat looked at how grants were spent.

On the other hand, the case studies indicate someof the difficulties that can arise. The only study tosystematically monitor the effect of cash distributionon prices – in Ethiopia – shows what can happenwhen cash is delivered in an economy where foodprices are rising rapidly due to overall shortages.Several districts in the CFF programme, and in theUNHCR project in Albania, found theadministrative burden imposed by a cashdistribution (especially on the banking system)exceeded capacity, causing delays (see UNICEF(1988); UNHCR (2000)). Furthermore, themonitoring and accounting of cash movementsclearly needs to be even more stringent than forfood or non-food items, and was found to beinadequate in at least two of the projects (seeUNICEF (1988) and Devereux (2000b)). Anothermajor issue is the exclusion of certain groups fromeconomic activity or ownership (women insouthern Africa, for example), which can make itdifficult to achieve equitable cash distr ibutionwithin households (see Devereux, 2000b).

Despite the problems they highlight, the casestudies give cause for cautious optimism, andprovide pointers as to the circumstances in whichcash relief can work to best effect. But they are attheir most useful in highlighting just how little weknow about cash transfers, and about the behaviourof the local economy, of traders and ofbeneficiar ies. Under what circumstances willtraders respond to an increase in demand? Whatlevel of purchasing power is necessary, and at what

Conclusion

20


distance from supply, to ensure an inflow of foodor other items? How do prices behave followingan injection of cash? At what level of cash inflowdoes inflation become inevitable? How dobeneficiaries spend cash, and how do these patternschange in different situations? How can investmentin livelihoods be encouraged, while at the same

time allowing beneficiaries the flexibility to satisfytheir consumption needs? These questions mustbe tackled through carefully planned andmonitored experimentation and pilot projectsbefore we can be confident in cash-basedinterventions, and learn how to minimise the risksand maximise the benefits.

21


British Red Cross (1999) Hurricane Mitch Agricultural Support Programme: Rebuilding Rural Livelihoods. AComparative Analysis of the Effects of Cash Assistance on the Recovery of Rural Livelihoods. London: British RedCross.

Bryson, J. and Hansch, S. (1993) Food/Cash for Work Interventions in Famine Mitigation. Washington DC:Office of Foreign Disaster Assistance, Famine Mitigation Strategy Paper.

Buchanan-Smith, M., S. Jones and B. Abimbilla (1995) Review of the Bawku Emergency Programme (draft).London: Action Aid.

Davies, R. (1996) Humanitarian Aid: Economic Implications and Consequences. Oxford: University of OxfordRefugee Studies Programme.

De Waal, A. (1989) Famine That Kills: Darfur, Sudan, 1984–1985. Oxford: Clarendon Press.

Devereux, S. (1988) ‘Entitlements, Availability and Famine’. Food Policy, August.

Devereux, S. (2000a) Famine in the Twentieth Century. Sussex: Institute of Development Studies.

Devereux, S. (2000b) Social Safety Nets for Poverty Alleviation in Southern Africa. ESCOR Report R7017. Sussex:Institute of Development Studies.

Devereux, S. (2000c) Social Safety Nets for Poverty Alleviation in Southern Africa. Summary Research Report.Sussex: Institute of Development Studies.

Drèze, J. and Sen, A. (1989) Hunger and Public Action. Oxford: Clarendon Press.

FAO (2000) The State of Food and Agriculture. Rome: FAO.

Grosh, M. (1992) ‘The Jamaican Food Stamps Programme’. Food Policy, February.

Hossain and Shuaib (1998a) Flood 1998. Nutrition Assessment, Coping Mechanism and Rehabilitation Activities.Round 2, Nov–Dec 1998. Dhaka: Save the Children.

Hossain and Shuaib (1998b) Rapid Nutrition Assessment During Flood 1998. Coping Mechanisms. CommunityPerception of Suitable Rehabilitation. Volume 1. Dhaka: Save the Children.

Jenden, P. (1995) Cash-for-Work and Food Security. Network Paper 11. London: Relief and RehabilitationNetwork.

Keen, D. (1992) ‘Providing an Income: The Other Side of the Relief Coin’. Rationing the Right to Life.London: Zed Books.

Kumar, B. (1985) The Ethiopian Famine and Relief Measures: An Analysis and Evaluation. Addis Ababa: UNICEF.

Lawrence, M. (1999) Food Economy Assessment of Kosovo Province. July– September 1999. Rome: World FoodProgramme.

Lawrence, M. (2000) Summary of Projected Food Aid Needs for Kosovo. April 2000–March 2001. Kosovo:World Food Programme.

Mitchell, J. (1996) The Monetisation of Food Aid. Brussels: EURONAID.

References

22


SCF(UK) (1998) DEC Appeal: Flood in Bangladesh. Save the Children UK: Flood Recovery Plan 1998–FinalReport. London: Save the Children.

Sen, A. (1981) Poverty and Famines: An Essay on Entitlement and Deprivation. Oxford: Clarendon Press.

Schubert, B. (1987) ‘Cash Transfers to the Poor’. Food Policy, February.

UNICEF (1984) Local Purchase of Food Commodities Project, Ethiopia 1984. Evaluation Report. Addis Ababa:UNICEF.

UNICEF (1985a) Evaluation Report on the Cash-for-Food Project in Siqe Wereda, Kembatana, Shoa. Addis Ababa:UNICEF.

UNICEF (1985b) Evaluation Report on the Cash-for-Food Project (II) in Mama Midir Wereda (Menzna Gishe,Shoa, Ethiopia). RRC/UNICEF, Addis Ababa.

UNICEF (1985c) Evaluation Report on Cash-for-Food Project in Konso Wereda, Gardula Awraja, Gamu Gofa.Addis Ababa: RRC/UNICEF.

UNICEF (1986) Evaluation Report on Cash-for-Food Project in Yifatna Timuga Awraja (Shoa Region, Ethiopia).Addis Ababa: RRC/UNICEF.

UNICEF (1987) Evaluation Report on Cash for Food Project in Haraghe Region, Hara Awraja, Fedis Woreda. AddisAbaba: UNICEF.

UNICEF (1988a) Quick Assessment: Cash For Food Programme, UNICEF/RRC Emergency Intervention. AddisAbaba: UNICEF.

UNICEF (1988b) Cash For Food Project Evaluation of Arsi I, II & III Sites. Addis Ababa: UNICEF.

UNHCR (2000) The Kosovo Refugee Crisis. Geneva: UNHCR Evaluation and Policy Analysis Unit.

Wilson, K. (1991) Enhancing Refugee’s Own Food Acquisition Strategies. Paper for the Symposium: Responding tothe Nutrition and Health Crisis of Refugees: The Need for New Approaches. Oxford: University of Oxford RefugeeStudies Programme.

23


Good Practice ReviewsGood Practice Reviews are commissioned ‘state of the art’ reviews on different sectors or activities within the relief andrehabilitation field. Prepared by recognised specialists, and subject to peer review, they are produced in a format that is

readily accessible to field-based personnel.

Email: <[email protected]>Website: <www.odihpn.org>

Network Papers are contributions on specific experiences or issues prepared either by HPN membersor contributing specialists.

Study III ed. J. Borton (1996)17 Monetisation: Linkages to Food Security? by J. Cekan, A.

MacNeil and S. Loegering (1996)18 Beyond Working in Conflict: Understanding Conflict and

Building Peace (The CODEP Workshop Report), by J. Bennettand M. Kayitesi Blewitt (1996)

19 Human Rights and International Legal Standards: what reliefworkers need to know by J. Darcy (1997)

20 People in Aid Code of Best Practice in the Management andSupport of Aid Personnel ed. S. Davidson (1997)

21 Humanitarian Principles: The Southern Sudan Experience by I.Levine (1997)

22 The War Economy in Liberia: A Political Analysis by P.Atkinson (1997)

23 The Coordination of Humanitarian Action: the case of SriLanka by K. Van Brabant (1997)

24 Reproductive Health for Displaced Populations by C. Palmer(1998)

25 Humanitarian Action in Protracted Crises: the new relief‘agenda’ and its limits by D. Hendrickson (1998)

26 The Food Economy Approach: a framework for understandingrural livelihoods by T. Boudreau (1998)

27 Between Relief and Development: targeting food aid fordisaster prevention in Ethiopia by K. Sharp (1998)

28 North Korea: The Politics of Food Aid by J. Bennett (1999)29 Participatory Review in Chronic Instability: The Experience of

the IKAFE Refugee Settlement Programme, Uganda by K.Neefjes (1999)

30 Protection in Practice: Field Level Strategies for ProtectingCivilians from Deliberate Harm by D. Paul (1999)

31 The Impact of Economic Sanctions on Health and Well-beingby R. Garfield (1999)

32 Humanitarian Mine Action: The First Decade of a New Sectorin Humanitarian Aid by C. Horwood (2000)

33 The Political Economy of War: What Relief Agencies Need toKnow by P. Le Billon (2000)

34 NGO Responses to Hurricane Mitch: Evaluations forAccountability and Learning by F. Grunewald, V. de Geoffroy& S. Lister (2000)

Network Papers

1 MSF-CIS (Celula Inter-Secçoes), Mozambique: A DataCollecting System Focused on Food Security and PopulationMovements by T. Dusauchoit (1994)

2 Responding to the 1991/92 Drought in Zambia: TheProgramme to Prevent Malnutrition (PPM) by D. Mukupo(1994)

3 An Account of Relief Operations in Bosnia by M. Duffield(1994)

4 Bad Borders Make Bad Neighbours - The Political Economy ofRelief and Rehabilitation in the Somali Region 5, EasternEthiopia by K. Van Brabant (1994)

5 Advancing Preventive Diplomacy in a Post-Cold War Era:Suggested Roles for Governments and NGOs by K. Rupesinghe(1994)

6 The Rwandan Refugee Crisis in Tanzania: initial successes andfailures in food assistance by S. Jaspars (1994)

7 Code of Conduct for the International Red Cross and RedCrescent Movement and NGOs in Disaster Relief ed. J. Borton(1994)

8 Targeting the Poor in Northern Iraq: The Role of Formal andInformal Research Methods in Relief Operations by P. Wardand M. Rimmer (1995)

9 Development in Conflict: the Experience of ACORD inUganda, Sudan, Mali and Angola by ACORD (1995)

10 Room for Improvement: the Management and Support of ReliefWorkers by R. Macnair (1995)

11 Cash-for-Work and Food Insecurity in Koisha, SouthernEthiopia by P. Jenden (1995)

12 Dilemmas of ‘Post’-Conflict Transition: Lessons from the HealthSector by J. Macrae (1995)

13 Getting On-Line in Emergencies: A Guide and Directory to theInternet for Agencies involved in Relief and Rehabilitation by L.Aris, P. Gee and M. Perkins (1996)

14 The Impact of War and Atrocity on Civilian Populations: BasicPrinciples for NGO Interventions and a Critique ofPsychosocial Trauma Projects by D. Summerfield (1996)

15 Cost-effectiveness Analysis: A Useful Tool for the Assessmentand Evaluation of Relief Operations? by A. Hallam (1996)

16 The Joint Evaluation of Emergency Assistance to Rwanda:

1 Water and Sanitation in Emergencies by A. Chalinder (1994)2 Emergency Supplementary Feeding Programmes by J.

Shoham (1994)3 General Food Distribution in Emergencies: from Nutritional

Needs to Political Priorities by S. Jaspars and H. Young (1996)4 Seed Provision During and After Emergencies by the ODI

Seeds and Biodiversity Programme (1996)5 Counting and Identification of Beneficiary Populations in

Emergency Operations: Registration and its Alternatives by J.Telford (1997)

6 Temporary Human Settlement Planning for DisplacedPopulations in Emergencies by A. Chalinder (1998)

7 The Evaluation of Humanitarian Assistance Programmes inComplex Emergencies by A. Hallam (1998)

8 Operational Security Management in Violent Environmentsby K. Van Brabant (200)

Background

The Humanitarian Practice Network (HPN) was launched in 1994 in response to research that indicatedsubstantial gaps between practitioners and policy makers in the humanitarian field, as well as serious weaknessesin the ability of the sector to learn and become more ‘knowledge-based’.

Purpose

To stimulate critical analysis, advance the professional learning and development of those engaged in andaround humanitarian action, and improve practice.

Objectives

To provide relevant and useable analysis and guidance for humanitarian practice, as well as summaryinformation on relevant policy and institutional developments in the humanitarian sector.

Activities

• Publishing in three formats: Good Practice Reviews (one per year), Network Papers (four to six per year)and Humanitarian Exchange (two per year). All materials are produced in English and French.

• Operating a resource website: this is one of the key reference sites for humanitarian actors.

• Collaborating with international ‘partner’ networks: this increases the reach of the HPN, and bringsmutual benefit to the participating networks.

• Holding occasional seminars on topical issues: these bring together practitioners, policy-makers andanalysts.

HPN Target Audience

Individuals and organisations actively engaged in humanitarian action. Also those involved in the improvementof performance at international, national and local level – in particular mid-level operational managers, staffin policy departments, and trainers.

While a project and Network with its own identity, the HPN exists within the Humanitarian Policy Group atthe ODI. This not only ensures extended networking and dissemination opportunities, but also positions theHPN in a wider ‘centre of excellence’ which enhances the impact of the HPN’s work.

FundingThe Humanitarian Policy Group is supported by the British Red Cross, CARE, DANIDA, DFID, ECHO, MFANetherlands, OCHA, SCF(UK), SIDA, UNDP, USAID and the WFP.

H U M A N I T A R I A N P R A C T I C E N E T W O R K

cash transfers in emergencies: evaluating ... - odi hpn · network coordinator would appreciate...

Documents