Swaziland Cash and Food Transfer Programme

By Rosie Jackson

Rosie Jackson currently works for Save the Children UK as an Emergency Food Security & Livelihoods Advisor. Based in London, she provides technical support to global programmes. Prior to joining the London team she worked in Peru, India, Pakistan, Zimbabwe and Swaziland.

Parts of this article summarise the findings of an independent evaluation carried out by Stephen Devereux, Institute of Development Studies, and Paul Jere, Independent Consultant, in June 2008¹.

This article outlines a short term emergency safety net programme by Save the Children UK, using a combination of food and cash transfers to households during a national food crisis in Swaziland.

Severe drought during the 2007/2008 agricultural season caused a 60% drop in national maize production, resulting in the lowest harvest on record. This was exacerbated by forest fires in the highveld region of the country, which damaged the forestry industry and resulted in many lost labour opportunities. The combined shocks of drought and forest fires impacted heavily on an already vulnerable food security context. The majority of rural households in Swaziland depend on smallholder agriculture but yields are low and have been declining since 1996. The national poverty rate in Swaziland is currently 43%, while HIV prevalence is 26% (2007) and is believed to be the highest in the world. This has been significant in the reversal in human development indicators that had been rising until the mid 1990s. The current legacy from AIDS deaths is 69,000 orphans, which weighs heavily on the economy and society.

Cash transfer beneficiaries receiving cash at the Siphofaneni Post Office in Swaziland

Although triggered largely by natural shocks, the clear trend of diminishing national food production indicates that chronic food insecurity is rising and that vulnerability to such shocks is likely to increase in future years. It also raises questions about existing strategies to combat or reverse these trends. Traditionally, safety net programmes in Swaziland have been in the form of monthly food distributions, causing significant dependency in some regions. In recent years, the Government has started to implement a number of cash based safety net programmes, including an old age grant, an orphans and vulnerable children (OVC) grant and a disability grant. Although this signals positive change, these have had marginal impact due to poor coverage and insufficient grant size.

Design of programme

Agencies in Swaziland have routinely responded to food and income shortage with a standard food parcel of cereal (usually maize), pulses and vegetable oil. Cash transfers, in response to an emergency, are an entirely new type of intervention in the country. A detailed market feasibility assessment for using cash provided the foundation for action, as it confirmed that local and national markets would support an increased demand for (largely food) products and suggested the inflationary effect would be insignificant. A cash transfer safety net programme was therefore designed by Save the Children (SC), to support beneficiary households with access to food and other basic items to meet immediate humanitarian food and non-food needs, while also protecting and promoting livelihoods. The Save the Children Emergency Drought Response project (EDR), funded by the UK Department for International Development (DFID), distributed a combination of food aid (50% ration) and cash (market value of a half food ration) to 6,200 households each month for 6 months. Beneficiary households were identified using nationally agreed vulnerability criteria from SC operational areas. Cash was transferred to private bank accounts opened with the support of SC at Standard Bank (SB), or through Post Office accounts. The exception was child-headed households who received their cash transfers directly from SC. Lump sum grants were transferred to each household during the first and last payments to strengthen livelihoods.

The cash transfers were accompanied by financial literacy training for all beneficiaries and a multimedia approach to communicating the function of bank accounts and use of ATMs (automated teller machines) to access the cash. Interactive community training sessions were organised to draw out issues for discussion arising with the cash transfers. Mock-up ATMs were used to demonstrate the cash withdrawal process. Posters, picture boards and information leaflets reinforced project messaging around savings, expected/appropriate spend of the transfers and the complaints procedure for the programme.

Considerable investment in a comprehensive monitoring and evaluation system generated useful data before and during the intervention. The investment included a market feasibility study and baseline survey (preimplementation), and monthly monitoring of disbursements (cash and food), markets (prices and availability), and households (income, expenditure, assets and diets). A final evaluation survey was implemented in May 2008 (post-implementation). The sample of 1,784 households included 1,225 'cash plus food' recipients, 491 'food only' recipients and 68 child-headed households. Households who received food only were sampled from neighbouring districts, where agencies were implementing the national food distribution response. Child headed households received cash transfers and food rations but were treated as a separate category given their unique characteristic as minors.

Impact

Commuinty training on savings and investment

Although the EDR project was primarily a humanitarian intervention, cash transfers were envisaged as a mechanism that might help break the cycle of reliance on food aid and encourage debate about how to strengthen the capacity of chronically vulnerable rural families. The EDR served as a pilot for introducing cash transfers into the Swaziland context for the first time within a humanitarian context and had to contend with significant challenges as a response to acute (and possibly also chronic) food insecurity in Swaziland. Nine hypotheses were outlined at the outset of the programme as part of the monitoring and evaluation system² in order to test the impact of cash transfers, and compare the impact of a combination of food and cash to food only as an intervention. The key findings from a comprehensive independent evaluation³ with regard to these hypotheses were as follows:

1. Cash improves nutrition and dietary diversity.
   Cash transfers were unconditional, yet all recipients spent some of this cash on purchasing food for consumption. Cash recipients spent almost double the amount on food compared to 'food only' recipients, and purchased a wider variety of food groups, as evidenced by consis- t e n t l y higher dietary diversity scores among children in 'cash plus food' households. Selfreported hunger fell immediately after cash transfers were introduced (from 70% to 22% of 'cash plus food' recipients), but less dramatically among 'food only' recipients (from 79% to 61%). This positive impact on household food security was sustained throughout the project period, with fewer households that received cash transfers reporting hunger than those that received only food aid. Hypothesis #1 was accepted.
2. Cash enables purchases of essential nonfood items.
   Apart from purchasing food, EDR cash transfers were also allocated to a range of non-food needs, from household non-food items (99% of cash recipients, 7% of total spending), to health (58% of recipients, 2% of spending), to education (38% of recipients, 7% of spending), to clothing (39% of recipients, 3% of spending), to debt repayment (31% of recipients), to transport (7% of total spending). Cash was especially useful for meeting seasonal needs, such as school fees that are payable each year in January. 'Food only' recipients spent less on all these categories, and were forced to deplete their assets or sell food items to raise the necessary cash. Hypothesis #2 is accepted.
3. Cash is invested in assets and livelihoods.
   After food, livelihoods were the second largest category of household expenditure (16% of total household spending). Many cash recipients invested in their farm (e.g. buying fertiliser) or business (e.g. retailing or selling cooked food). Monitoring of harvests and business enterprises some months after the EDR intervention would tell us whether the production and income gains attributable to the cash transfersare sustained. Some cash transfers were saved as working capital, or as security against future shocks, and membership of savings clubs doubled. Cash recipients also protected their assets against depletion to meet food and nonfood needs more effectively than did 'food only' recipients. Although both groups raised cash through asset sales, cash recipients also purchased assets using their cash transfers, including livestock, farm implements and household goods. Extrapolation from survey data suggests that more than 650 households acquired chickens using cash transfers, more than 200 households bought goats, about 375 bought hoes, and more than 150 bought radios. Hypothesis #3 is accepted.
4. Local markets are strengthened by cash injections.
   In August 2007, the market feasibility survey predicted that cash transfers would cause food price inflation of 5-7% in local markets. In fact, food prices rose much more than this. The cost of a half ration in monitored markets peaked in January 2008 at 37% higher than when the baseline survey was conducted in October 2007, but the evidence is inconclusive as to what proportion of this was attributable to the cash transfers. Hindsight shows that international factors were influencing prices across the globe. Some traders opportunistically increased their prices on cash paydays. It seems most likely that EDR cash transfers had negligible impact on aggregate food prices at national level, but a (temporary) inflationary effect on prices at the local level. On the other hand, stocks of food and non-food commodities in shops and local markets did increase, confirming that supplies were responsive to increased demand. Purchases of these items by cash recipients certainly contributed to an income multiplier effect, but its magnitude cannot be quantified with the available data. Hypothesis #4 is neither accepted nor rejected.


Cash transfer beneficiaries receiving cash at the Siphofaneni Post Office in Swaziland

5. Harmful coping strategies are avoided.
   Drought-affected households in rural Swaziland adopted the full range of 'coping strategies' observed in food crises elsewhere in Africa - rationing food, borrowing food or cash, migrating for work, selling livestock or other assets, and withdrawing children from school. There were no statistically significant differences in coping strategy adoption rates between 'cash plus food' and 'food only' households: 68% of cash recipients and 67% of 'food only' recipients rationed consumption, for instance. One explanation is that all EDR households were exposed to breaks that occurred in the food aid pipeline. If cash transfers had been increased, or if pure 'cash only' transfers had been delivered, these households would have been better protected. More severe strategies, such as withdrawing children from school (4%) and selling assets (2-3%), were adopted by small minorities of households. Nonetheless, the similarity across cash and food recipients implies that cash and food transfers were equally (in)effective in protecting families against the need to adopt austerity measures that could undermine their future livelihoods. Hypothesis #5 is rejected.
6. Caring practices for children improve.
   There is insufficient empirical evidence on this topic, which requires a special in-depth study. Hypothesis #6 is neither accepted nor rejected.
7. Women are empowered by receiving cash.
   Women were registered as cash recipients and bank account-holders in 90% of households receiving cash transfers, as a deliberate strategy to empower women and ensure that cash was used responsibly to meet the basic needs of women and vulnerable children. Concerns that disbursing cash to women in male-headed households could result in genderbased violence proved to be unfounded; most men accepted that women spend cash sensibly ("our wives know what to buy"). The main challenge to this decision at the intrahouse-hold level was not gendered but 'generational'. Many chidren, knowing that the transfer was calculated at E30 per person, demanded their 'share' of the money given to their mother or carer, even though the intention was that the cash should be used to benefit the household as a whole. Hypothesis #7 is accepted.
8. Cash delivery systems are appropriate, timely, safe, well targeted and scaleable.
   The delivery of cash transfers raises a number of 'customer care' issues, including targeting, accessibility, timeliness and security at cash collection points. The EDR project performed well on all of these issues. Cash transfers were delivered in full to all recipients on specified dates with no 'pipeline breaks' (unlike food aid), queuing times were long initially (over 4 hours) but fell (to under 2 hours) as efficiency improved, transport costs to pay-points were reimbursed, there were no reports of serious security problems, bank and Post Office staff treated cash recipients courteously and SC staff were always on hand to provide assistance (e.g. in using bank ATMs) and to listen to complaints. Many cash recipients switched from the Post Office to ATMs with little difficulty, and many expressed their intention to continue using the ATM after the project ended. A very small number of unpleasant incidents (e.g. verbal or physical abuse, drunkeness) were reported. On this evidence, there is no reason why cash transfers delivered through Post Offices and/or bank ATMs, could not be scaled up to national level in Swaziland, either in future emergencies or as a predictable social protection measure. Hypothesis #8 is accepted.
9. Beneficiaries are more sympathetic to cash transfers than before.
   The pre-intervention market feasibility and baseline surveys both recorded an overwhelming preference for food aid among rural households. Given a hypothetical choice, 60% of respondents in the market study (July 2007) chose food, 23% chose cash and 17% chose half food, half cash. In the post-intervention final evaluation survey (May 2008), this question produced very different responses: only 6.3% chose food, just 2.4% chose cash, and 91.3% chose half cash, half food. This finding is ambiguous; pure cash transfers were less preferred after the intervention than before due to perceived difficulties in accessing sufficient grain stocks (down from 23% to 6%), but the preference for cash or 'cash plus' increased from less than half to almost all respondents (up from 40% to 97.6%). This represents an endorsement of cash transfers as at least part of the solution to future food crises. Hypothesis #9 is accepted.

In addition to the above hypotheses, and in line with SC's mandate, the evaluation looked at spend by and on children. Child headed households spent their cash very responsibly, prioritising basic needs (food, non-food household items, clothing), followed by health and education. Some also used the cash transfers to pay off debts, and a smaller number shared their windfall income with others. This pattern followed the spend of adult headed households very closely, except that child-headed households did not report investing in farming or business, presumably because most are too young to farm or run a business. Child headed households had a lower overall spending on a monthly basis as evidenced by data on coping strategies, where 90% of child headed households had rationed food during 2008, compared with 68% of adult headed households. This implies higher levels of vulnerability and poverty and suggests greater food insecurity within this group.

Six months on - how sustainable was the programme impact?

Continued monitoring of 920 programme beneficiary households, aimed to answer certain questions following the evaluation, namely the appropriateness of the substantial investment made in the transfer mechanism and the relative sustainability of impact on income generation.

For the households still accessing their private bank accounts beyond the end of the programme, there have been relatively few problems and a small proportion of these account holders have taken advantage of additional financial services provided by the bank, such as receiving the transfer of funds from outside of the country. Deposits and withdrawals have continued up to the date of this report and understanding of the financial aspects of the accounts remained strong following the project period.

However, not surprisingly, lack of disposable income prevented the majority of the EDR beneficiaries from using their accounts following the final transfer, when over 70% of households withdrew all funds and have not since used the accounts. Although activity within the accounts remained high during the 3 months immediately following the project, a combination of shocks such as further forest fires, extended dry spells and a 100% rise in cereal prices contributed to a decline in disposable income and subsequent account activity. By November 2008, only 4.6% of all EDR households were accessing their accounts.

The lump sums transferred during the EDR project designed to support livelihood activities and assets were invested well, with appropriate uses by over half of all beneficiary households (whilst some households prioritised food piling, healthcare and education). Post project monitoring confirmed that this investment continued to contribute to household income for 45% of all beneficiaries, a figure that declined during the months of August and September but stabilised at 30% in December. This figure represented around 60% of all households generating an income.

Withdrawing money from an ATM as part of cash transfer programme

Transfers through bank accounts, whilst time consuming in the set up period, was a suitable approach and could be considered in short and long term cash transfer programmes. Mainstreaming of cash transfer procedures into long term safety net programmes and into disaster preparedness systems would support rapid scale up in emergency contexts.

Conclusions

In the programme evaluation in May 2008, evidence of the appropriateness of the response and the additional value of providing cash as an alternative to food were clearly identified, with especially encouraging patterns in child and household dietary diversity. Chronic malnutrition is a big challenge in Swaziland. The evaluation did, however, conclude that the project started too late to avoid adoption of harmful coping strategies. Considering the rapid decline in food security for the majority of households following the project, this report also concludes that the funding for the project finished too early.

The Government needs to take responsibility for the underlying issues contributing to food insecurity in Swaziland, namely poor investment in agricultural production, poor access to healthcare, insufficient social protection, lack of labour opportunities and rising food prices, and work together with partners towards providing long term policy based solutions that support families in a practical manner. Given the regularity of emergency events in Swaziland, the high level of vulnerability to hunger and chronic poverty, particularly in the Lowveld, predictable long term safety net transfers should be designed by Government in collaboration with relevant United Nations agencies and non-governmental organisations.

A range of long term safety net programmes, including a monthly unconditional cash transfer to the most vulnerable households, which could be scaled up to meet the additional needs during a drought year, would support the existing development efforts by the Government. For this purpose, the efficiency and proven beneficiary adoption rates make private bank accounts a feasible transfer mechanism for long term transfer programmes.

Cash transfers had a greater positive impact on children's diet than food distribution only and in such a fragile and vulnerable context, predictable safety net transfers must be designed as part of a broader Government social protection system.

For further information, contact: Rosie Jackson, email: r.jackson@savethechildren.org.uk

Two related external evaluations of the project are available for download from the online library on the Save the Children website: http://www.savethechildren.org.uk/en/54.htm

¹Devereux, S. & Jere, P. (June 2008) Choice, Dignity and Empowerment - Cash and Food Transfers in Swaziland. An evaluation of Save the Children's Emergency Drought Response 2007/08.

²Part funded by Department for International Development (UK), Regional Hunger and Vulnerability Programme and World Food Programme.

³By Stephen Devereux (Institute for Development Studies) and Paul Jere (independent consultant). See footnote 1.