Issue 40 - Article 15

The World Bank's experience with cash support in some recent natural disasters

October 10, 2008
Rasmus Heltberg, World Bank

Compared to the famine-inducing disasters of earlier periods, South Asia has progressed in its ability to cope with disaster. Although several major natural disasters have hit the region in recent years – the tsunami, the Kashmir earthquake, floods and cyclones in Bangladesh – it has not experienced a famine for decades. Thanks to better disaster responses in the region, large natural disasters no longer trigger massive food insecurity and high post-disaster mortality.

This article looks at the role of cash support in the wake of natural disasters, drawing out lessons from the World Bank’s use of large-scale cash support in South Asia and Turkey. In the past, the World Bank often focused on reconstruction and long-term development, leaving humanitarian relief to others. But starting with the earthquake in Turkey in 1999 and continuing through the South Asian disasters of 2004 and 2005, cash support to affected households has played a growing role in the Bank’s disaster response. This article reviews some of the lessons from Bank-supported large-scale cash support.

Cash support has performed well. Large-scale cash income support directly to affected households complements other relief and reconstruction efforts and seems to have had positive impacts on short-term food security and long-term recovery. To perform even better in future disasters, donors should consider building up their own capacity and that of implementing agencies to deliver timely and high-quality cash support.

Why income support?

The purpose of cash transfers and other social protection instruments, when used in disasters, is to protect the basic consumption of the affected population and to help preserve and recover assets and human capital. They are a basic safety net in times of critical need, when community support mechanisms are exhausted. The support aims to prevent famines and food insecurity, speed up the recovery of assets and livelihoods and protect a population’s schooling and health status after a disaster. All of this will prevent poverty from deepening.

Income support complements other types of post-disaster efforts, such as housing and infrastructure reconstruction. This complementarity became clear to the Bank in Gujarat: ‘earthquake victims’ need for cash assistance became apparent to the Bank in an indirect way following the 2001 earthquake. Families were using the first installment of house construction funding to purchase food and other necessities to survive, rather than using it toward the construction. As a result, when it became time to issue the second installment, many of the families did not have the first phase of the house to show in order to receive the second installment’.

Cash or food?

Cash remains underutilised in humanitarian relief, although there is growing evidence of its efficacy. The World Food Programme, for example, has embraced cash and voucher-based safety nets in its Strategic Plan for 2008–2011. Cash confers dignity and choice and tends to have lower transaction costs and higher value to beneficiaries than in-kind support. Cash is best delivered through cash transfers or workfare. Cash support must be timely, temporary, not too large and distributed in communities of regular residence, and it must end as livelihoods bounce back. Cash support is inappropriate where food markets are weak or non-existent.

Success requires careful attention to the details of design and implementation: how will support be targeted and distributed, to whom, how much and for how long? Success also requires speedy deployment. Therefore, there are clear advantages on speed and efficiency grounds of drawing up plans for disaster response and helping to build implementation capacity before disaster strikes.

Income support in South Asia and Turkey

Many agencies help countries respond to large disasters. The role of the World Bank in disasters has often been as lead donor, trusted advisor and coordinator of reconstruction and long-term development, but not relief. Since 1984, the Bank has financed natural disaster activities in 528 projects, worth $26 billion, constituting some 10% of total global Bank lending in the last two decades. India and Bangladesh alone had 71 disaster projects from 1984 to 2005. Government agencies implement most of these projects and fund them through a combination of new loans and grants and reallocations from existing loans.

Bank support in major disasters generally comprises a multi-sectoral mix of interventions, such as budget support, emergency preparedness and rehabilitation and reconstruction of schools, housing, infrastructure, agriculture, forestry and health facilities. The World Bank assisted in the deployment of income support as part of its response to disasters in Turkey (1999), Sri Lanka (2004), the Maldives (2004) and Pakistan (2005).

While many different agencies have experience with post-disaster cash transfers, the uniqueness of the World Bank’s recent experience is the scale of the transfers. The marriage of government implementation channels and Bank support (mostly financial, but also some technical assistance) allowed cash grants to be offered on an unprecedented scale. In the Maldives, all affected households (one-fifth of the population) received tsunami cash assistance, most of them within one month. In Sri Lanka, cash grants were given to some 250,000 households (in the first round), covering all of the affected households and even some that were not directly affected (later tranches were therefore targeted more narrowly). In Pakistan, Bank-supported cash grants were given to 250,000 households, approximately 30% of all those affected. Key design details are summarised in Table 1.

The variation in the basic design of post-disaster cash support across countries shown in Table 1 is remarkable. It is particularly surprising that the amounts offered to affected households varied widely. Pakistan, after the Kashmir earthquake, provided the most generous support, especially when compared to GDP. In Pakistan, affected households received between $300 and $2,000, the equivalent of 7–48% of per capita GDP per member. It seems unlikely that Pakistan would be able or willing to finance transfers of this magnitude in future disasters. The comparable amounts per person are 3–9% of per capita GDP in Turkey, 3% in Sri Lanka and 1.5–4.5% in the Maldives. These amounts are in all cases substantially higher than the benefits provided to regular social assistance beneficiaries in these countries (typically the chronically poor).

There is limited evidence of the impact of this income support. The most detailed is from the Maldives, where a survey carried out six months after the tsunami showed that the support was adequate, well-targeted and helped people to cope and recover. Targeting appears to have been good and few eligible households, if any, were missed. The cash grants infused purchasing power into affected islands and helped to revive retail trade. Incidences of food shortage fell rapidly after peaking one week after the tsunami. Six months afterwards, household income was higher than it had been before the disaster.

From Pakistan, there are anecdotal reports that the government’s decision to pay compensation for death may have led to inflated casualty figures: verification of claims for death compensation proved difficult as the dead had often been buried without vital registration.

There is little solid evidence on the extent of corruption. In Sri Lanka, reports found mistargeting (inclusion of unaffected households) but not corruption. Nevertheless, risks are large. Many studies of South Asia’s safety nets – albeit from non-disaster settings – have pointed to leakages. Most reported problems arise with workfare and in-kind transfers, such as India’s subsidised food grain programme. There are fewer reports of mistargeting and leakages of cash transfers, probably because they are simpler to implement.

Implementation arrangements are critical and a particular challenge is to combine speed and quality in the delivery of assistance. It proved difficult to align technical assistance for design and implementation with the desire to roll out cash transfers swiftly. Some governments went ahead with disbursing cash transfers without waiting for technical assistance: Pakistan (death/disability grants), Sri Lanka (cash grants) and the Maldives (cash transfers) all started without external technical assistance, while Pakistan’s livelihood grants benefited from technical assistance but were paid out somewhat later. As a consequence, monitoring arrangements were often simple, with little or no data collection, sometimes hampering ex-post verification and audits.

Lessons learned

The main lesson that emerges from these experiences is that speed is of the essence: if donors are slow, governments will design and implement programmes on their own. Some further tentative lessons are listed below.

Cash preference: Income support should be in cash unless there is evidence that food markets are disrupted. In rapid-onset disasters, food and other in-kind support can be useful if the disaster has disrupted markets. However, in-kind support should subsequently be converted into cash support except where food markets remain interrupted. In slow-onset disasters (with the possible exception of conflicts) food markets are more likely to function and the case for cash is therefore stronger.

Duration and target group: Cash support should be provided only for a limited period for most households, but could be stretched into the medium term for the most vulnerable. In small-to-medium emergencies, the best approach may be to provide an initial fairly large transfer to all those affected, followed by a second, smaller transfer (perhaps after three months). Later transfers should be targeted to vulnerable households. In very large emergencies, household targeting within affected areas may be required for all transfers. For the most vulnerable groups (such as orphans and disabled people), additional social services are also required.

Benefit levels: The amount of support should be adequate for subsistence but not so high as to jeopardise work incentives or cause inflation. Larger amounts should be given only as one-off compensation, for example triggered by the loss of a house.

Eligibility: Eligibility criteria would ideally be predefined, transparent, easy to explain and simple to administer. Criteria could emphasise loss of house or other assets for immediate support (but without discriminating against renters), shifting to poverty criteria and possibly even a proxy means test for medium-term support.

Delivery agency: Income support would ideally be coordinated and delivered by one or more of the regular safety net or social security programmes with experience in handling cash transfers. Given the logistical challenges in large emergencies, support from others such as army staff or local governments will probably be needed.

Delivery channel: For rapid support to people who do not have bank accounts, cash-in-hand may be used to avoid the delays of opening bank accounts. Cash is also often the best way to pay workfare participants. For medium-term support, bank or post office accounts should be used. In the future, transfers via mobile phones may become more common.

Systems are needed: Systems for targeting, implementation and monitoring and evaluation should be prepared in advance.


Large-scale cash support has been an important and effective part of the response in several recent major disasters. As cash transfers will continue to be needed, it makes sense to improve their design and seek to make them swifter, more equitable and more consistent. Affected countries and donors should work this into their disaster preparedness plans. Ex-ante planning should answer the core design questions: in what form, amount and duration should income support be offered? which households should be eligible for support and how can they be targeted? which agency is responsible for delivery? Implementing agencies should start building the necessary response capacity. These steps would help the international humanitarian community to respond better and faster to the next emergency.

Rasmus Heltberg ( is a Sr Technical Specialist at the World Bank. For more on this subject, see Rasmus Heltberg, ‘Helping South Asia Cope Better with Natural Disasters: The Role of Social Protection’, Development Policy Review, vol. 25, no. 6, 2007, pp. 681–98.


Comments are available for logged in members only.