Protecting livelihoods during drought: some market-related approaches
by Christopher Eldridge May 2003

The history of food crises includes a history of famines that did not occur. Drawing on the experiences of the 1992 drought in Southern Africa, Christopher Eldridge argues that an understanding of the methods people used to avert famine then – notably food purchase – could usefully inform today’s response.

Ten years ago, in 1992, the worst drought in around half a century struck Southern Africa. But there was no famine, largely because of the activities of those people who were most affected, namely the rural poor. Villagers obtained most of their staple food from non-aid sources, particularly through increased purchases. To finance these, most poor villagers intensified non-water dependent income-generating activities, sold animals and/or cut expenditure on education and agricultural inputs. During the planting season, they worked off their own land for longer than usual.

Many of these attempts to finance increased food purchases had adverse effects on livelihoods. Relief programmes during drought should aim to complement and support those responses that do not damage livelihoods. In particular, relief interventions should aim to limit the decreases in poor people’s purchasing power and incomes which normally occur in food crises, especially in rural areas.

Four approaches

This article outlines four possible approaches that address demand issues during drought. The first two address expenditure; they aim to maximise the amount of staple food which poor households buy by temporarily reducing the price and by suspending user fees for essential services. The second two address income; the aim is to minimise the reductions in earnings which most poor households experience during drought.

1. Subsidising staple food prices

During drought, the purchasing power of most villagers decreases because the price of cereals, the item they spend most on, generally rises. This is compounded by reductions in net earnings, especially for the poorest households.

Subsidising cereal prices for the period of the drought would enable people to buy more for the same outlay. As a result, people should have less need to resort to measures that damage their livelihoods, such as working off their own land for excessive periods during the planting season, or selling draught cattle.

Subsidising staple food prices is, in theory, a relatively easy way, both logistically and administratively, of maintaining the purchasing power of the poorest households during drought. However, this has rarely been tried during food crises, and difficulties have arisen where it has been attempted. While subsidies may not be appropriate in every food crisis, these difficulties should not automatically prevent further attempts to apply them. Factors necessary for their successful implementation include adequate preparation, complementary monitoring systems at several levels and comprehensive public information. Where subsidies have not worked, it is likely that these and other prerequisites have been deficient.

2. Waiving user fees and other charges

User fees during drought reduce the amount of money villagers can spend on increased purchases of staple food. Where they exist, fees could be waived for schooling, health services and animal dipping. A clear announcement should be made that the fee waivers would end on a certain date.

Waiving fees on essential services would meet two objectives: it would increase villagers’ disposable cash, so helping to finance increased purchases; and it would lessen the need for villagers to reduce or stop their use of these services so as to save money for increased cereal expenditure. This would benefit poor villagers proportionately more than rich ones, because the former spend proportionately more of their earnings on cereals. It would also be administratively easier to implement than other measures, and would facilitate targeting between districts and between households. Fees could be waived only for drought-affected districts, or a sliding scale could be introduced, in which the proportion of the fee that was waived was higher for more severely affected areas. Fees could also be waived only for poor households.

Other regressive charges on the poor could also be waived, where they exist. These could include land and livestock taxes, and contributions to local government bodies.

3. Policies and programmes to support rural incomes during drought

Cash for work projects

Cash for work projects are one way of supporting rural incomes during drought. They should have two main objectives: to minimise reductions in the incomes of the poorest households during drought; and to produce outputs which will provide lasting benefits for communities, especially for the poorest households.

To achieve these objectives, certain criteria must be met. For example, projects should be planned in advance; villagers should participate in their planning and monitoring; projects should not draw villagers away from working on their own fields during the land preparation and planting months; and they should be limited to outputs which provide lasting benefits, and which villagers genuinely want and are prepared to maintain after the drought has passed.

These and other criteria place limitations on the number of public works projects that can be successfully implemented during drought.

Supporting income-generating activities

Maintaining incomes by supporting existing enterprises and income-generating activities is difficult, since most yield smaller earnings during drought. Furthermore, livelihood shocks, such as drought, often precipitate a shift to illicit or environmentally damaging income-generating activities. These include gold panning, tree-felling, prostitution and theft. It would be inappropriate to support activities such as these.

The number, type and mix of income sources vary greatly between districts; there can accordingly be no standard prescriptions for improving rural incomes. Consequently, interventions should be founded on a thorough knowledge of local livelihoods. Implementing organisations must have a good working relationship with drought-affected communities. Time and personnel constraints, in addition to the various problems caused by drought, mean that it is generally not feasible to support new enterprises and income-generating activities during drought.

A developmental approach: increasing rural incomes during non-drought years

Increasing the incomes of poor villagers during a drought is difficult. A more appropriate approach would be to ensure that poor smallholders enter a period of drought with as high an income base as possible. This means building on smallholders’ strategies during non-drought years.

Responses to drought should be seen in a longer-term, developmental perspective, particularly as drought can no longer be regarded as an occasional, exogenous shock. The likelihood of drought, and of seasons with below-average rainfall or with poor rainfall distribution, needs to be factored into longer-term development policies and programmes, especially those concerned with poverty reduction.

Attempts to increase the income of poor smallholders during non-drought years should focus – as many longer-term projects do – on diversifying and increasing income sources, especially off-farm sources. This involves a range of measures, including micro-finance, greater access to inputs, improved access to output markets, improved rural roads and better rural transport. Women require particular support, given their varied contributions to household livelihoods.

4. Stabilising livestock prices

Livestock prices fall during drought for two main reasons: increased supply (because animals cannot be kept alive and/or because larger numbers than usual have to be sold to finance increased expenditure on cereals); and decreased quality (as worsening supplies of water and grazing make animals increasingly thin and sick).

During drought, decreases in livestock prices (especially cattle prices) interact with increases in cereal prices, resulting in sometimes large deteriorations in the pastoral terms of trade (the rate at which livestock products can be exchanged for cereals). Rural livelihoods are consequently compromised, both in the short and medium term. It may take many years to rebuild cattle herds to their pre-drought levels. Minimising decreases in livestock prices during drought will therefore reduce the extent to which drought, and villagers’ responses to it, adversely affect livelihoods. Three ways of lessening reductions in livestock prices during drought are described below.

Selling feed at subsidised prices

Commercial stockfeeds (and/or fodder) could be sold at subsidised prices, preferably through rural stores, but also by NGOs and churches, and/or by ministries or local councils.

Relaxing purchasing criteria

Where parastatals and other organisations are involved in the purchase of livestock, consideration should be given to relaxing their purchasing criteria during drought; adhering to standard criteria can reduce the income-earning possibilities open to livestock owners.

Supporting prices

Where no parastatal is involved in livestock purchase, the Ministry of Agriculture, the local council or an NGO could buy cattle (and possibly goats) at reasonable prices, regardless of their condition. The animals could be slaughtered and the meat used in feeding programmes at schools or hospitals; or the meat could be dried and sold as biltong. Alternatively, animals could be purchased and kept alive, on the understanding that their owners could buy them back after the drought at a similar price.


Interventions which address food demand issues during drought are arguably more effective at protecting sustainable livelihoods than those which address food supply problems, at least in areas where markets are functioning reasonably well. Food demand interventions may be categorised as either expenditure or income measures. It is easier to increase villagers’ purchasing power and disposable cash with interventions which address expenditure (such as subsidies and fee waivers) than by interventions which address income (such as livestock price maintenance and support for income-generating activities). The former, while not straightforward, have fewer logistical and local management requirements than the latter. They also have the potential advantage of minimising reductions in expenditure on essential services (and goods) which some villagers, especially the poorest, may make to finance increased expenditure on staple food.

Attempts to increase the incomes of people affected by drought should generally be complemented by other interventions, and should begin before drought occurs, within a developmental framework.

Timing and seasonality are important. Demand issues must be addressed early in crises, not only because they should be, but also because they can be. The relative speed with which polices can be implemented to protect the purchasing power of the poorest households and their use of essential services and farm inputs during drought is a key advantage of this approach. Relief programmes take longer: in 1992, most of the relief food imports arrived in Southern Africa six or more months after orders were placed. The balance between demand- and supply-side measures should vary during the course of a food crisis, with the latter perhaps becoming more dominant in the later phases.

Adequate financing is essential for the success of marketrelated interventions, especially for subsidies, which have significant budgetary implications. The four approaches outlined above may not be appropriate in all food crises. Nevertheless, demand issues should be addressed in some way in most food crises, especially where purchases of staple food are likely to comprise a substantial proportion of the responses of poor people. Interventions which address demand issues during food crises have been given much less attention than relief food programmes. They are not without their problems, but nor are food distribution operations. More policy-oriented research into factors which would lead to their success and more pilot projects are needed.

Christopher Eldridge was Save the Children-UK’s country director in Sudan during the 1985 famine, and in Zimbabwe during the 1992 drought. While head of Save the Children-UK’s regional office for Southern Africa between 1995 and 1998, he carried out participatory research on household responses to the 1992 drought, involving 936 households in 72 villages in Malawi, Zambia and Zimbabwe.

References and further reading

Stephen Devereux, Theories of Famine (Hertfordshire: Harvester Wheatsheaf, 1993).

Stephen Devereux and Simon (eds), Food Security in Sub-Saharan Africa (ITDG, 2001).

Christopher Eldridge, Some Recommendations for Relief and Rehabilitation Programmes During and After Food Crises in Southern Africa (London: Save The Children-UK, 1995/2002).

Christopher Eldridge, ‘Why Was There No Famine Following the 1992 Southern African Drought?’, IDS Bulletin, vol. 33, no. 4, October 2002.

Michael Hubbard, Improving Food Security: A Guide for Rural Development Managers (ITDG: 1995).

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