What affects the cost of delivering cash transfers in humanitarian settings?

By Clare O'Brien and Fidelis Hove

Clare O'Brien is a senior consultant in the Poverty and Social Protection team at Oxford Policy Management, a development consultancy. She specialises in the design and analysis of social protection policies and programmes.

Fidelis Hove is a consultant in the Poverty and Social Protection team at Oxford Policy Management. He specialises in analytical studies of cash transfer programmes and social security systems, with a particular focus on sub-Saharan Africa.

The study was commissioned by CaLP with funding from Visa. The report from which this article is derived was co-authored by Clare O'Brien, Fidelis Hove and Gabrielle Smith. We warmly appreciate the valuable inputs and time devoted by Oxfam, Concern Worldwide and SOS Children's Villages in Kenya and Somalia, and by other contributors worldwide. 

Cash transfers have been a recognised alternative to in-kind aid in humanitarian disasters for some years¹. Under certain conditions, when local markets can accommodate increased demand and prices will remain stable, they may offer benefits to recipients and donors alike. Households often have great freedom as to what they can buy, where and when. Meanwhile, agencies may experience a lighter logistical burden as cash transfers do not require the procurement, transportation and storage of in-kind goods. 

One consideration among several when determining the appropriateness of a cash transfer is its cost: the 'value for money' perspective means that what matters is not just a programme's effectiveness but whether the money might have been better spent elsewhere. 

The Cash Learning Partnership, CaLP, has funded a set of case studies of emergency cash transfer programmes, to understand their cost and its determining factors. The research was carried out by Oxford Policy Management (OPM)². OPM has also published a guide on the method used, so other agencies can replicate the analysis on their own programmes. This article summarises the findings and provides brief pointers on the method. References to the full papers are provided at the end.

The case studies

We analysed three cash transfer programmes in urban and rural Kenya, and four in Mogadishu, Somalia. The programmes operated between 2009 and 2013 and were run by Oxfam, Concern Worldwide and SOS Children's Villages Kenya (Table 1).

Cost-efficiency or cost-effectiveness?

We conducted cost-efficiency, rather than cost-effectiveness, analysis. Cost-efficiency analysis calculates the administrative costs of delivering a transfer. Cost-effectiveness analysis would have compared the cost with the size of the outcome, e.g. improved food consumption. The latter was not possible because our retrospective analysis had access only to post-distribution monitoring data which could track changes in beneficiary well-being but could not attribute the change to the intervention. 

What costs count?

'Administrative costs' cover everything spent by implementing partners other than the transfer received by the beneficiary. They include not only direct purchases e. g. for transport, printing or buying bank cards, but also the estimated value of time spent by staff for everything from proposal-writing to monitoring and evaluation. NGOs might refer to this expenditure as direct and indirect programme costs, operating costs, management costs or support costs. They should not be confused with the narrower concept of 'overheads'. We used this broad definition because changes in the cost of any of these can lead to improvements in programme efficiency. 

The findings

We see in Table 1 that the costs of administering the seven case study programmes, proportional to the amount of cash disbursed, varied hugely. The cost–transfer ratio, that measures how much it costs to deliver every $1 received by the beneficiary, is up to six times higher in some programmes than in others. 

Note: The 'cost–transfer ratio' is the total administrative cost divided by the total amount disbursed to beneficiaries. A cost of 0.20 means that for every $100 received by beneficiaries, it costs $20 in operations for it to reach them. 

However, a very high cost–transfer ratio does not automatically signal an inefficient programme; nor is a low cost–transfer ratio always good. The context of where the programme operates, what infrastructure it has access to, who its beneficiaries are, and the size and duration of the transfer, drives a large part of the cost. 

So what are these contextual factors, and can anything be done to improve cost-efficiency? We present here some findings from Kenya and Somalia, and consider the implications. 

Kenya

Kenya is vulnerable to regular climate shocks. In rural areas such disasters directly affect livelihoods, while for urban citizens food shortages typically result in price rises. Droughts and accompanying food price hikes have occurred twice recently, in 2009-10 and 2011-12. 

Many agencies delivered emergency cash transfers during this time. They were able to use electronic as well as manual payment mechanisms because Kenya has a rapidly expanding banking sector, very widespread mobile network coverage and enormously popular mobile-money services (primarily Safaricom's M-Pesa). 

• Oxfam's support to residents of Nairobi's informal settlements (case study 1) incurred a high ratio of administrative costs to transfer value (0.64) mainly because its objective was to give a relatively small payment of less than $20 per month to nearly 2,800 beneficiaries. This resulted in high registration costs. As the lead partner in a consortium, it also spent a lot of time securing funding and conducting advocacy campaigns on the crisis. But the programme ran for a long time, 18 months, so its average administrative costs gradually declined as the fixed costs were offset against more transfers; and the costs of regular disbursement using M-Pesa were low. 
• In contrast, SOS Children's Villages Kenya (case study 2) gave much larger transfers of $87 a month to 2,000 beneficiaries in Marsabit, a remote area of northern Kenya. The relative cost of administering the programme was further reduced because the agency received substantial discounts from the payment provider, sQuid and Paystream, keen to trial their smart cards for emergency response. On the other hand, the agency had to set up an office and register beneficiaries and its expenses were increased because poor network connectivity meant that staff travelled regularly to Nairobi to upload data.
• Concern Worldwide's programme (case study 3) fell between these two in terms of its cost–transfer ratio. It had very low set-up costs because it was a continuation of a previous intervention, so it incurred no costs of establishing an office or registering beneficiaries. However, day-to-day costs of disbursement were increased because payments were distributed manually by traders, who charged quite a high commission. 

Somalia

The 'south–central' region of Somalia, i.e. the area outside the largely autonomous regions of Somaliland and Puntland, is governed by no single authority and suffers repeated conflicts. In 2011, as in Kenya, it suffered a severe drought. An estimated 1.4 million people were internally displaced, often to camps around Mogadishu. Famine was declared in July 2011, triggering a massive response by non-governmental organisations (NGOs). 

However, there were few options for NGOs wishing to deliver cash. The financial services sector was dominated by money transfer (hawala) agents in the absence of formal banking. There was no central bank until 2012, no registered private banks, no ATMs and no point-of-sale terminals in stores. In 2011 two mobile phone companies launched mobile-money services which quickly became popular. Oxfam used the hawala agents for its major emergency response in 2011–12. It then piloted the use of the mobile-money scheme in mid-2012. Concern Worldwide used mobile money for the two programmes we reviewed from 2012–13. 

• Oxfam's Emergency Cash Transfer Programme (case study 4) was by far the largest intervention we analysed. It benefited from economies of scale as it reached some 12,500 households who received six payments totalling $5.57 million in transfers. The $1.12 million it cost to administer the programme included commission to the hawala agents at 2.5% of the transfer value, and several hundred thousand dollars of staff time to oversee disbursement. Nonetheless, even if a mobile-money system had been available at the time (which it was not), a similar amount might have been spent on purchasing mobile phones and SIM cards for beneficiaries and commission would have been paid to a 'cash facilitator' to transfer money to the mobile network operators. The manual payment system was therefore not necessarily inefficient for a programme of that size and duration.
• Oxfam's 'E-cash' mobile-money pilot (case study 5) tested the feasibility of using the new technology by trialling it for a one-off transfer. This means that the administrative cost looks quite high compared with the amount disbursed, because the purchase of phones and SIM cards was not offset by many transfers. But this brief use was central to the objective of testing the system. 
• Once the system had been tested and found to be effective, Concern Worldwide was able to take up the technology and to distribute cash transfers using mobile money at very low cost (case studies 6–7). They benefited from economies of scale, using the mechanism for numerous transfers; in the case of the IOM-funded intervention, they retargeted existing beneficiaries so had no costs of identifying new recipients or providing them with phones. 

Implications for implementing agencies

Agencies and their funders are, naturally, keen to pinpoint the high-cost items in their interventions, so that they can consider whether the intervention can be delivered in an alternative, more cost-effective way. At the outset of this research, for instance, we aimed to explore whether electronic payment mechanisms are generally more cost-efficient than manual payment mechanisms. 

What we found, though, was that costs are driven not so much by individual line items, but rather by processes. We identify three key lessons:

1. Many costs are negotiated rather than fixed, so savings can be made if aid agencies make their programme attractive to participating companies. In Somalia, the hawala agents tried to negotiate an increase in their commission when they realised the inconvenience of having to pay 12,500 beneficiaries in addition to their regular customers; in contrast, in Kenya the smart card provider offered a discount because it wanted to participate. 
2. The state of infrastructure development has a huge impact on cost. Aid agencies can try to drive innovation in infrastructure development but this risks being complex and expensive. Agencies will incur extra costs, for example, if they attempt to use electronic payment mechanisms in locations where the network coverage is poor, or where recipients and payment agents are unfamiliar with the technology and require training and ongoing support. 
3. More broadly, a key determinant of cost is the amount of new activity required in a programme. This could entail training new partners, registering beneficiaries, or introducing a new payment mechanism. If cost is the driving force in the design of a programme, there is therefore a risk that innovation will be lost: it is cheaper to keep paying the same beneficiaries with the same payment mechanism than to look for new ones, but this may not result in the programme achieving its objectives.

Cost-efficiency is also affected by choices about the size of the transfer. A programme will look more cost-efficient if the payment to beneficiaries is increased, because although the commission on the payment may also rise, other expenses such as registration costs will remain the same. Again, though, this may not be in line with programme objectives: it may be necessary to pay smaller amounts to a larger population. We conclude that it is therefore more appropriate to make decisions about programme design, such as the selection of the targeting or payment mechanism, on factors other than simply the cost.

For more information, contact: Clare O’Brien, email: clare.obrien@opml.co.uk

For more information:

O'Brien, C, Hove, F and Smith, G (2013). 'Factors Affecting the Cost-efficiency of Electronic Transfers in Humanitarian Programmes'. http://www.cashlearning.org/resources/library/416-factors-affecting-the-cost-efficiency-of-electronic-transfers-in-humanitarian-programmes

O'Brien, C. (2014), 'A guide to calculating the cost of delivering cash transfers in humanitarian emergencies: With reference to case studies in Kenya and Somalia'. OPM Working Paper, June. http://www.opml.co.uk/publications/opms-publications (search under publication type 'OPM Working Paper'). 

¹See for example, Harvey and Bailey (2011).

²http://www.opml.co.uk/