Operational definition of a famine

Summary of workshop¹

A one day workshop was held on March 14, 2003 by the Institute of Development Studies (IDS) Sussex, with the aim of developing an operational definition of famine. Convened as a follow-up to an IDS 2002 conference on famine, it was attended by a group of academics, donors, agencies and the ENN.

Dried out sorghum harvest with barely any crops on it , Wollo.

Four cases were highlighted to focus discussion. Ethiopia (1984) was unanimously declared a famine while Iraq during the 1990s, Ethiopia (2000) and Malawi (2002) were more equivocal. The ensuing discussion highlighted factors to consider in defining a famine, including:

• Labelling a crisis 'a famine' tends to elicit donor response, but conversely may result in donor fatigue if 'over-used'. Thus, operationally, the word may need to be used selectively.
• Response to famine involves political calculations, i.e. donors tend to respond through media and international pressures as well as geo-political considerations, rather than solely to objective information on need.
• Should a definition account for what victims themselves believe famine to be?
• How many deaths have to occur before the term famine is used? Should the term be used broadly, or 'saved up' or for extreme circumstances?
• An operational definition should be about 'early warning, response and accountability' and establishing criteria to determine whether a current situation is a potential future famine. Current early warning systems appear to be 'too late'.

Need for an operational definition of famine An operational definition of famine was deemed necessary in order to strengthen the following:

• Accountability and transparency to the beneficiaries of donors, local, national and international community, and all levels of government
• Advocacy - to present a coherent message, set and order priorities for action and to broaden understanding of food crises at all levels
• Response - to promote understanding/common framework for better, faster, co-ordinated response
• Early warning mechanisms - by including indicators that could identify different levels of food insecurity, using a more standardised and comparable approach than current early warning systems.

Identified limitations of developing an operational definition included:

A static definition may restrict responses until a situation is finally defined as a famine, when intervention may be too late. Also, it may focus responses on food at the expense of other sectors, or even at the expense of other foodrelated, non-famine disasters.

Politics and accountability.

An operational definition may provide excuses for donors and agencies not to respond to situations unless a situation is labelled as famine, and could be subject to political manipulation and misuse.

Difficulties in establishing indicators.

It may prove difficult to reach consensus on one operational definition, e.g. how to take into account cultural differences, different target audiences (donors/NGOs), etc.

Following a review of existing famine definitions and group discussions, consensus was achieved on the following issues:

• An operational definition required a verbal description of famine which revolved around access to food, with tools to generate appropriate responses to certain situations. It was felt, however, that this would be extremely difficult to achieve.
• All famines can be measured by outcome indicators, such as nutrition, malnutrition and mortality, but numbers and severity will differ from famine to famine. Similarly, all famines can be measured by process indicators, but the endpoint (mass death) does not have to occur before the situation is labelled a famine.
• Agreeing a timeframe in which a famine occurs is essential for operational usefulness - an acute event will require an emergency response, whereas a developing situation (famine process) will require longer term strategies.
• Responses to famines inevitably tend to be guided by the politicisation of events and media publicity.

Outstanding disagreements/ambiguities included whether a famine was a process or event, what time-frame should be included in an operational definition, for whom the definition was designed, and whether economic and social factors should be taken into account in determining causality.

Strengths and weaknesses of definitions

The strengths and weaknesses of recent definitions/frameworks of famine (Howe and Devereux², Banik³), which involved gradation or scales, were discussed. It was suggested that scaling risks implies different, mutually exclusive, levels of famine. Also, the type of information included (malnutrition, mortality) may be difficult to gather, tended to be outcome focused or may only become available when the famine is well advanced. It was countered, however, that early warning mechanisms, as well as some aid agencies, have the capacities to collect timely key data. Alternative (qualitative) forms of criteria/indicators exist or could be developed for earlier stages. Even if certain data could only be obtained if/after a famine has occurred, it may still be used as a determination of accountability, for future advocacy and improved response.

Despite these issues, the concept of scale or gradation was deemed a useful one, since it includes the idea of a "threshold" which is crossed when a famine crisis occurs, and gives some focus to the process as well as the "event" of famine. Indicators that point to the likelihood of a famine were considered necessary, so that preventative steps may be taken. However in reality, it was considered difficult to mobilise resources for "prior stages of famine" and donors will need to be convinced that "something special is happening". Critical to a definition is how, and who, will determine that the "threshold" has been crossed?

Workshop conclusion

An operational definition was not agreed at this meeting, but key attributes of a definition were, with suggestions for taking the process forward.

Given the high levels of emergency situations and crises, an operational definition was considered a matter of extreme importance. Consensus building was necessary to approach a common position amongst stakeholders and with significant input from donors early on in the process. Gaps in the process as it stands need to be identified, with action points, eg conducting regional case studies on famine threshold issues.

The workshop was deemed successful in gaining operational insights into academic frameworks that are being generated, and giving momentum to taking the process forward through a wider consultation with stakeholders and decision-makers.

For further information, contact Stephen Devereux, IDS Sussex, University of Sussex, Brighton, UK.
Email: S.G.Devereux@ids.ac.uk

¹Report and minutes of "Operational Definition of Famine Workshop" IDS Sussex, Friday 14 March 2003

²Howe, Paul and Devereux, Stephen, 2002. Notes towards an operational definition of famine. Mimeo: Institute of Development Studies, Sussex.

³Banik, Dan, 2002. Democracy, drought and starvation in India: Testing Sen in Theory and Practice, PhD thesis submitted to The Faculty of Social Sciences, University of Oslo. (Chapter 3: Refining Sen (1):Operational definitions of famine and related terms)