Contents - Previous - Next
This is the old United Nations University website. Visit the new site at http://unu.edu
Issues in evaluating food crisis warning systems
Bruce Currey, Food Crisis Management Group, School of Social Sciences, The Flinders University of South Australia, Bedford Park, South Australia
This conceptual article provides a framework for judging whether a food crisis warning system will be effective. The article is based primarily upon experience in rice-farming regions of Asia with highly seasonal employment calendars, but it formulates concepts that are sufficiently general to be applied to the evaluation of warning systems in other regions. Although conceptual, this paper does not discuss the theory of whether or not warning systems are desirable, nor does it consider budgetary factors such as the unit costs and who pays for them. No attempt is made to detail the operation of any warning system in a particular geographical context according to this framework.
The warning system shown in figure 1 is the flow of information that ensures that effective and efficient interventions are made to prevent food crises. Within the warning system, those making decisions transform knowledge about an imminent food crisis into action to prevent future food crises and minimize immediate suffering. The decision-makers have to decide:
- whether or not to intervene,
- which interventions to use,
- whether or not the interventions are being effective,
- whether changes in interventions are necessary,
- whether to stop the interventions.
During the 20-year period 1963-1983 there were over 400 crises or emergencies to which the international community responded with food aid (1). During the 1970s, three to four Asian countries each year were reported to be coping with a food crisis. These crises were actual or potential food stress situations that drew off funds, energy, and time from long-term development programmes aimed at sustained food security. Now the World Development
FIG. 1. Flow of Information for the Food Crisis Warning System
Report (2) is emphasizing early warning systems for food security. Some Asian countries are now incorporating early warning systems in their development plans.
The Goverment of Japan has recently awarded one million dollars for the establishment of early warning systems for Asian countries in collaboration with the Food and Agriculture Organization and the International Food Policy Research Institute. The World Bank has funded a project for "the early identification of factors suggesting existing or imminent food shortages" in areas where new settlements a replanned. The United States Agency for International Development has funded extensive studies of "managing food security in extreme conditions" and "early warning information and intervention systems" as part of national nutritional surveillance programmes. However, a crisis warning system for food security is not just a part of nutritional surveilance; it is part of the basic socio-economic surveillance necessary for national security.
Previous discussion of early warning systems to achieve food security has been remarkably incomplete. FAO and others continue to emphasize the forecasting of production shortfalls (3-6). The Office of Special Relief Operations sends missions to potential food crisis areas for careful calcuation of the food availability per capita, but they also admit that the control of the consumption of much of that food sometimes falls into the hands of the military sector (7).
On the other hand, nutritionists and epidemiologists, often associated with WHO and UNICEF, have correctly emphasized the problem of detecting changes in consumption rather than only in production. There is a tendency, however, to confuse warning systems with forecasting systems. Their objective has been determining whether or not to sound the bell or flash the red light in order to initiate interventions (8-11). Few have evaluated the interventions as an integral part of the food crisis warning system (12).
Recently, some scholars have also alluded to the possibility that warning systems, such as those outlined in the early South Asian famine codes, may actually reinforce relief dependency (13, 14). They may only be Band-Aids that potentially prevent more equitable development.
FIG. 2. A New Paradigm for Evaluating Food Crisis Warning Systems
Finally, the political nuances of effective food crisis warning systems have been under-emphasized either intentionally or unintentionally (15, 16). The mechanisms and statistics of the food crisis warning system have been studied, but the people those who operate the warning system, and those who suffer if it fails - have been forgotten.
A PARADIGM FOR EVALUATING FOOD CRISIS WARNING SYSTEMS
To remedy the above limitations, there is need for a paradigm, or way of looking at food crisis warning systems. The following section, read in conjunction with figure 2, outlines one possible perspective.
The new perspective must ensure that detection is considered in parallel with intervention. There are always on going interventions in response to early detection, for example, in pest control. Effective decision-making requires information not only about the threat of a crisis, but also about
the state of readiness of various interventions: rice required for a feeding programme may continue to be exported from the area, or relief works programmes may not be adequately prepared to absorb the surplus labour of groups whose consumption is likely to deteriorate.
Food crisis managers must be viewed as central to any food crisis warning system. They are personnel in the normal goverment administration, perhaps at the level of bupatis in Indonesia or of district officers in Bangladesh. The managers have to juggle, usually amidst reduced administrative capacity, information relating to the state of detection of the food crisis and information about the interventions required to combat the crisis.
There must be measures for judging whether the warning system is functioning effectively. These are the performance criteria. Some people may state the latter in terms of never having to give out food doles. Other may see them in terms of never having anyone die of starvation in the district. The performance criteria may eventually be improved to include any crisis of severe malnutrition, but whatever the criteria are they must be clearly stated before the warning system can be evaluated.
There must be an assessment of the likelihood of a food crisis occurring in a particular food system. This decribes the vulnerability of the area. There is little point in praising a local administrator's ability in preventing food crises if his area is not, and has never been, prone to food crisis. Vulnerabililty, or the probabiliy of a food crisis, is likely to be high in areas where there are population groups with limited access to consumption because of an unbalanced distribution of income, severance of traditional patronage for the poor, and a high incidence of natural social hazards or disruptions.
The existing warning system must be described. In any area vulnerable to food crises there will be an existing warning system. Both the government and the population are likely to respond to a perceived food crisis. The existing warning system may be ineffective or inefficient and so it must be improved. It is important, however, for local people to realize that there is a warning system, and therefore information providing the data it requires can help to improve the decision-making process.
The new perspective must include consideration of how the existing warning system is to seek continuous improvement. This learning system is a feedback loop that corrects faults or problems. It must be able to take into account the changing context of the warning system. Vulnerability may increase or decrease as new cropping systems are introduced, or as absentee landlords buy up land in an area, or as landless labourers are encouraged to migrate to another area. New detection procedures may evolve as information on new indicators (for example, the number of people queueing up for work) becomes available or is considered necessary. The possible list of responses or interventions may change when there is a new policy directive to remit revenue, or when a local politician demands that relief resources be used to develop a reservoir or fish pond. Finally, the performance criteria for the warning system may change: if there is an oil boom in the area, or if the area suddenly becomes strategically important, political leaders may become unwilling to accept even a small number of cases of "hunger oedema" (never mind "starvation deaths") in that area.
The people's perception of the problem must be incorporated into a food crisis warning system, or else political pressure through the media and opposition parties may force more effective alternatives. Local knowledge may improve detection by pointing out that, although the officially collated rice prices indicate that all is well, half the population is eating famine foods such as bonn kochu (Araceae spp.) in parts of Bengal, or duan turi (Sesbania grandiflora) in parts of Indonesia. Spontaneous responses by individuals or groups may be much more effective than government interventions. For example, families may move out temporarily to live with relatives elsewhere. People's linguistic perceptions of famine, as reflected for instance in nuances in local vocabularies, may suggest alternative indicators for detecting food crisis: e.g., in Bengali akla, meaning "scarcity," may suggest short-term programmes, while durvickha, meaning a "shortage of alms," may suggest that longer-term programmes are required.
Decisions in a warning system are made to prevent present and future crises. The aim of these decisions is to strengthen the ability of those communities served by the warning system to withstand the disruptions associated with food crises. Any response or intervention made as a result of a decision in the warning system ought to increase such community resilience. The interventions must therefore ensure sustained access to food consumption for the most vulnerable groups. They certainly must not lead to discrepancies in resource allocation and thus increase the likelihood of future food crises and chronic structural hunger.
If there is an integrated rural development programme in the area, its aim should also be to ensure sustained access to food consumption for the most vulnerable groups. The food crisis warning system should be able to detect the impact of integrated rural development strategies. The performance criteria for the warning system would then be able to monitor the impacts of new cropping patterns, village dams, or planned migration projects.
The warning system must be alert even when there are no food crises. There has to be constant revision of all the components that affect decisions made in the warning system. Training drills and simulations must instil recognition that any warning system must change even in the absence of food crises. The readiness of the warning system (fig. 1) depends on the food crisis managers' being fully cognizant of changes in the regional food system (fig. 3) over which they have jurisdiction for food security.
NINE CRITERIA FOR JUDGING ANY FOOD CRISIS WARNING SYSTEM
With this new perspective it is possible to discuss any food crisis warning system. The criteria for judging whether a warning system will be effective include:
(1) understanding the regional food system,
(2) codification of the warning system,
(3) understanding the limits of the information,
(4) the performance criteria for the warning system,
(5) the integration of detection and response,
(6) the administrative capacity of the food crisis managers,
(7) the learning mechanism in the warning system,
(8) the integration of the warning system with rural development,
(9) methods for keeping the warning system alert.
FIG. 3. The Regional Food System (adapted and translated from ref. 17)
1. Understanding the Regional Food System
Unless food crisis managers or their advisers understand the regional food system, both detection and response will be ineffective. For example, low flood levels may be discounted unless it is recognized that even low levels at particular times in the agricultural calendar may be destructive. A survey of wage rates may show deceptively high or low figures because of the timing of the survey. Crowds of people Iying on a railway platform may be migrant workers, not famine victims. Falling cattle prices in a market may be interpreted as suggesting asset liquidation nearby, unless it is understood that the cattle have been transported to that particular market for sale from a distant area. Food-for-work programmes may increase vulnerability both by diverting scarce food to the management team and draining off local labour through the inducement of excessive wages in kind.
Migration, the role of remittances, stocks, and storage problems are given too little prominence in the diagram, but such a schema as figure 3 provides a useful initial check list for evaluating the food crisis managers' understanding of the food system on which the statistics, rumours, and reports are collected in order to operate the warning system.
Warning system operators may gain this kind of under" standing of their regional food system in various ways. Lifelong experience augmented by the reading of history and folklore are probably the best. The "black books" of the district officers in South Asia transmit experience from one officer to his successor. Younger management personnel may be helped by anthropological and ethnological reports or theses, by journalists' and historians' accounts of past food crises, and by direct local observation. Long-term experience in understanding the food system is highly desirable. Any literature on the cultural ecology and the political economy (18) of the area will be useful. Some might question the cost-effectiveness of such knowledge of the food system, but if the manager has this in-depth understanding, the cheapest warning system would be to pay him extra travel per diem so that he and his principal deputies are encouraged to go and see what is happening within an area.
The major components or boxes in the diagram of the food system (fig. 3) suggest the main areas for which a limited number of indicators ought to operate. The managers of the warning system must be sensitive to critical points in the food system. There may be certain times of year when it is difficult to get relief into an area, or to implement a food-for-work programme. There may also be a seasonality in labour demand and famine-food consumption. Certain areas may have a large proportion of debt-bound, landless households.
The managers must know the dynamics of their food system. The introduction of a "catch" crop in one area may affect migrant labour demand in another area. The mortgaging of labour by a tribal group may make them susceptible to food stress at a later date. Changes in goverment policy on relief food allocation may have consequences on leakages from the food distribution system. Food crisis managers need not know what these impacts will be, but they must have the commitment and ability to investigate such possibilities.
2. Codification of the Warning System
There are informal and formal warning systems in all areas vulnerable to food crises. Informal ones may range from community prayers offered to weather and crop deities or signs of anxiety if no rain has occurred by the second week of the planting month to a tradition of selective migration as increasing dearth is detected.
Formal public warning systems began at least as far back as the "Book of Orders" in sixteenth-century Britain (19). They probably reached their greatest degree of sophistication with the Famine Codes of South Asia in the late nineteenth and early twentieth centuries after a series of severe famines. The managers or administrators during that period gained a great deal of experience in preventing food crises. In other areas the rural administration may have "standing orders" for what to do if there is a food crisis.
My reason for stressing the careful description of the existing warning system, however poorly it may perform, is to emphasize that improvements may be made in warning systems, even though there are no regularly collected statistical series of the kind that nutritional surveillance has traditionally included (20). Codification identifies "gaps in the armour" and hence opportunities for improvement.
Other reasons for carefully describing the existing system include the need to understand the geographical and functional limits of the warning system. The system may be designed to cope with a world-wide food crisis situation. It may be to alleviate food crises resulting from an absolute shortage of food, from difficulties in distributing food from one area to another, or from a drop in purchasing capacity. The food crises being alleviated at any particular time may be completely different, such as when people's ability to benefit from food intake is reduced because of an epidemic of diarrhoeal disease.
The final reason for describing the warning system is for training other managers in the area as well as managers from other areas. The description should emphasize which parts of the existing decision-making process are transferable. The description of the system will be particularly useful in judging how effective the system is, as opposed to how forceful and effective the personality of the present food system manager is.
There is no one way of describing a food crisis warning system. It may be compiled as a manual like the Bengal Famine Code (see Annex at the end of this article) - preferably loose-leaf. Those with a penchant for systems engineering may produce a blueprint of a systems model. These or other methods must clearly detail the attributes of the information used (both the formal information networks and the informal ones, however casual) at particular decision points, such as meetings. They must state who uses which decision rules - e.g., what proportion of the 10 most vulnerable families have to be eating famine foods before a particular intervention is implemented. Any contingency plans, in the event that there is an error in decision-making, should also be described. This description of the warning system, along with details of any changes and the rationale for such changes should all have the common aim of improving decision-making for preventing food crises.
3. Understanding the Limits of the Information
Almost all managers operate with incomplete information. It is important that food crisis managers know the limits of their data. The reasons data for detecting food crises are often incomplete and inaccurate have been analysed elsewhere (21). Jodha has presented data clearly illustrating the errors sometimes as much as 300 per cent - in information on interventions during a food crisis (22).
Whatever information is being used in making the decisions to prevent food crises, the limits of that information should be known by the manager. A subtle but crucial example of the limits of information is illustrated by the Indonesian saying that summarizes the attitude to information relating to food crises in many other countries, "ABS," which stands for Asal bapak senang ("As long as the boss is satisfied"). If the boss wants to hear about imminent food crises, it is highly probable that the warning system will be very sensitive to potential food crises. All of these data streams involve judgements that may be made by underpaid clerks. In almost all cases the person making the judgement is accountable to someone. The judgement may be affected by the person's accountability.
The manager must know how limited his information is.
- To which areas do the data refer?
- For which units were the data acquired (individual, household, or village)?
- With what frequency are the data collected?
- At what time are the data collected? For example, the hour of the day or the day of the week.
- Are the data consistent? For example, is the same type of rice used each time the price is recorded?
It is necessary to know these limits to the information, but this does not necessarily mean that more precise information must be gathered. As an example, when one hears that an intervention was triggered on the basis of a quick aerial reconnaissance of a war-torn country, a more detailed and costly survey to detect the magnitude of the problem may not be necessary. The first question ought to be about the costs, lead times, and manpower required to implement the response recommended after the aerial reconnaissance. Only then should these constraints on the response be compared with the limitations of the data that suggested the need to respond. Some examples of factors that affect the data are:
- the experience of the reconnaissance team or person,
- the agricultural season of the flight,
- the representatives of the area,
- the likelihood of a food crisis occurring amidst scenes of apparently plentiful production,
- the likelihood of only certain sections of society having access to the production or being unable to consume it because of illness,
- the likelihood of certain types of food production, such as house-gardens, being invisible to any aerial survey team.
Most of all, the decision to spend more resources on detection should be made only if the results of the air reconnaissance have been inconsistent with other available information. In medical terms, what is suggested is a "therapeutic trial" based on the knowledge that the available information is limited, rather than an epidemiological survey that will document symptoms rather than the processes causing those symptoms. The accuracy and precision of the warning system's ability to detect a food crisis should be no greater than the accuracy and precision of its ability to respond.
4. The Performance Criteria for the Warning System
These criteria indicate what the warning system is meant to do. They are the criteria for judging its success or effectiveness. Different groups will have varying ideas about what a food crisis is. There should be an operational definition agreed upon by all the people involved with the warning system. Those involved in improving the warning system must decide whether it is meant to prevent communities fragmenting because of the crisis, cases of hunger oedema, starvation deaths, any delay in the distribution of food relief, structural hunger, or seasonal scarcity. They must also decide whether the system is to prevent two out of three food crises in which over 1,000 starvation deaths are recorded, or whether it aims to prevent food crises in which there are 10 or fewer cases of hunger oedema twice in ten years.
Performance criteria should also contain a cost specification in determining what constitutes a successful system. It might be suggested that the costs of success are potentially high because the political costs of an unreliable system are high. In other words, regimes could fall if a food crisis occurs. Nevertheless, there are real political limits to costs because no government wants to have to provide food relief to an area year after year with no sustained impact.
Performance criteria are necessary politically in order to maintain the credibility of the system. Reporters and opposition spokesmen may bandy about unique photographs of a single child suffering from marasmus precipitated by post-measles diarrhoea. The managers of the warning system can refute such evidence by citing the performance criteria if they are written to exclude such conditions. The system itself has not failed in this case.
FIG. 4. Hypothetical Detection-Response Calendar
5. The Integration of Detection and Response
The response and detection system must be integrated. Traditionally, only the detection section of the agricultural calendar (fig. 4) has been emphasized as part of the warning system (23, 24). Few would argue, however, that work programmes should be maximized when wages peak in agricultural production activities. Understanding the response calender is as important as understanding the detection calendar.
The management of a food crisis warning system must have an understanding of the trade-offs between lengthening the lead time for detection and shortening the response time for interventions. The manager must also understand the trade-offs between the additional costs of increasing the validity of indicators and lowering the costs of intervening.
Attempts to select the best detection and response procedures should be developed while the warning system is operating in an ongoing food crisis situation. For example, would it be better to import food from farther afield with a longer lead time or try to import another type of food from closer at hand if it is difficult to increase the lead time using available methods of detection? Such decision-making studies will form the basis for training and keeping the warning system active. Any warning system is only as good as its interventions allow. The time and cost of the most appropriate interventions should help decide whether or not there is a need to improve the ability to detect food crises.
This co-ordination of detection and response is crucial at all levels in the administrative hierarchy. Some interventions, such as price control, procurement, and revenue remission at the district level may be pre-ordained by national policies. Other interventions may be constrained by local conditions; thus one cannot dig wells in basalt areas, especially if the water table lies far below the surface. There are probably many interventions that the food crisis managers at the district level can trigger by themselves, such as opening a gruel kitchen or allowing people to graze cattle in a public forest. The manager must know how effective his interventions are. For example:
- Have the brickworks that he has opened attracted the unskilled landless labourer?
- Has the subsidized rice been bought by the poorer groups?
- Are the chronically malnourished working in the stone-chipping programme?
The objective at all levels is to target the interventions to the people who need them at the most appropriate time in the most appropriate place.
6. The Administrative Capacity of the Food Crisis Managers
Like all managers, those managers who ameliorate and prevent food crises juggle personnel, resources, scheduling, costs, and objectives. They may be military or ax-military people with a particular style of management and allegiances to particular political groups. Sometimes they are locally elected people who are sensitive to the local informal channels of communication. In other cases they are rotated from area to area within the civil service and principally concerned with rapid upward mobility. Thus they are motivated to achieve short-term rather than long-term performance criteria. Their training may be both academic and vocational. Some may have been trained using a specific set of standing orders in manuals or codes. They may have experience in dealing with other food crises, which makes it important to know where and what type of warning system was in use. They may be urban-bred or rural-born and sensitive to the rhythm of rural life. Their operation of the warning system may be a vocation or merely an assigned role. As well as their need to understand the regional food system, they must also be constantly able to improve their warning system, to seek out weaknesses in the way it operates and then to take corrective measures.
The key actors must have strong working relationships, not only with their superiors in the national and international warning system, but also with the lower-level managers. They must work well with their colleagues in agriculture, food, health, public works, logistics, and local government. In concert with their teams at the district level, they are probably the only people capable of breaking out of the "administrative trap" (25) and co-ordinating the normally fragmented government approach. They should be forceful personalities with the ability to cut through red tape when necessary. The personality and cultural or ethnic background of the food crisis manager can easily affect the effectiveness of the warning system.
Any evaluation of a warning system must also determine how discriminating the manager is when making decisions based on casual dialogue with friends rather than checking the findings against other sources of information. Perhaps he is under pressure from those above him to meet targets. If so, will he be loath to admit to problems and also less ready to be seen responding to them? On the other hand, there may be pressures for him to exaggerate problems and intervene excessively. The pressures on the food crisis manager are just as great as those on any other crucial managerial positions, but to date the training of this specific type of management has only been through experience.
7. The Learning Mechanisms in the Warning System
Any food crisis warning system should be improved continuously. It must have a mechanism for coping with changes in the performance criteria, changes in crops, and changes in the agricultural calendar. Populations may outgrow planned storage levels. New migrants may settle in the area. Policies may change, managers may change, and new interventions may be tried. Modernization may mean that face-to-face contact is replaced by formal communication channels. Particular detection indicators may become irrelevant. There is little to be gained by monitoring the sale of gold ornaments if all of the poor sold their ornaments in the last food crisis.
The food crisis managers must learn how to learn. As new processes begin to operate in the regional food system, the managers must make modifications in the warning system. Post-crisis or yearly audits should be held, such as the former Famine Commission reports and more recent appraisals of the response to food crises in India (26, 27). There may be a special cell charged with studying and integrating changes, but there must be somewhere a mechanism to identify new alternatives. The reasons for any changes made should be documented in order to avoid repetition of previous weaknesses.
The manager should recognize that other interventions are going to affect the operation of the warning system. A rice field attacked by blight or insects may be detected, precipitating an immediate, effective spraying of the field by the pest control service. If the activities of the pest control service are not being monitored by the warning system, a food relief programme may be inefficiently triggered in the area.
8. The Integration of the Warning System with Rural Development
Some have argued that a food crisis warning system should successfully generate curative actions to eliminate the adverse effects of food crises and later turn to more preventive, long-term approaches to removing all malnutrition and hunger. Most curative interventions should concurrently serve a long-term, preventive function. Similarly, many preventive interventions should be adaptable to serve a curative function. Relief must be put in a development context. The ultimate goal is a resilient food system in the area-free not only from food crises but also from the hunger associated with structural poverty (fig. 5).
The food crisis warning system can be a vehicle for integrated rural development. Food relief and irrigation programmes must be designed to reach the poorest of the poor by giving them greater access to food and to the means of producing their own food. The manager should be concerned with redesigning relief programmes so that they are not merely palliatives. This will involve collaboration with the political process to encourage those changes most likely to bring greater equity of access to food in the agrarian structure at times of crisis. The manager should also be concerned with preparing work programmes or immunization schemes targeted to particular areas in order to ensure that they mesh with long-term integrated rural development plans.
The co-ordination role for the warning system should be through the Ministry of Rural Development or Provincial Development to the Ministry of Home Affairs, or whichever ministry co-ordinates development administration. Health ministries and traditional nutritional surveillance expertise should preferably not be charged with implementing early warning systems. Their role should be to evaluate the performance of the early warning system. Such surveillance should monitor whether the relief ministries, the public works department, and all the rural development projects are indeed reducing hunger and its causes in the regional food system.
FIG. 5 The Hunger Transition: A Normative Model - Putting Food Crisis Management in a Development Context
9. Methods for Keeping the Warning System Alert
"Crises" are not frequently recurring events. There must be a method for preventing managers from losing their vigilance and the procedures from becoming rigid. The Bengal famine of 1943 appeared almost 30 years after any previous food crisis, and Kampuchea's problems of the 1970s evolved almost 40 years after the last food crisis in what is normally a food-surplus nation.
There are at least three approaches to coping with atrophy in the warning system. The first is internal, through regular training procedures with mock food crisis manoeuvres. Food crisis "games" can be played like war games. They can be played by the food crisis managers in response to the collection of a piece of false information in the response system, such as a national directive to release a certain amount of food to the market at a reduced price. Simulated ideas for training can come from past food crises in the regional food system. They could also come from other food crisis managers throughout the world who have successfully prevented food crises.
The second approach to preventing ennui resulting from the system's lack of priority because of more pressing social or economic concerns is to persistently reform the performance standards so that the warning system is also dealing with those concerns. As the performance criteria are changed in order to prevent structural hunger, the manager of the system will have to explore the causes of hunger in the regional food system if he is to provide development solutions.
A third approach to keep the system alert is to train and educate those outside the formal food crisis warning system. Discussions can be held with journalists, who normally report on the symptoms of crises, in order to increase their sensitivity to signs and signals of the emerging processes that lead to the crisis. Education can raise the consciousness of children in areas vulnerable to food crises. Similarly, government officers in their initial training can be immersed for a short time in a rural food system and given details of the food crisis warning system. Constructive questioning from outside may be one method of increasing vigilance within the warning system and of preventing people from beginning to take advantage of the system.
This conceptual article has defined what food crisis warning systems are, has illustrated their complexity, and has emphasized the research and training needs necessary to create reliable warning systems. The nine criteria are probably more eclectic than those considered by some groups designing warning systems. This article may provide the basis of a first checklist for an evaluation handbook for the many new warning systems that are being set up in different countries. It will provide the springboard for training food crisis managers to learn from past and present warning systems in other areas of the world. The comparative evaluation of warning systems in different areas will improve their efficiency and lead to the more effective management of food crises.
This paper is based primarily on experience gained on four missions to evaluate food crisis warning systems in Asia - one mission to Bangladesh, and three in Indonesia. The missions were financed by the Us Agency for International Development, project nos. 388-0027 INDO-81-146 and 497-0272 and contract USAN-C-0272. In addition to the co-authors on the mission reports, the author is particularly grateful to his early mentors Dean Wilson and Roy Miller of the Community Systems Foundation, Ann Arbor, Michigan, USA; to Fariduddin Ahmed, A. N. M. Eusuf, A. A. Loedin, and Mark Brooks for their knowledge of loci conditions; and to J. P. Habicht, Jon O'Rourke, George Woods, Joseph Stepanek, and Nicholas Studzinski for their constant support. D. P. Chaudhri, David Morton, the late David Penny, Alberto Pradilla, Gordon Wolman, and my colleagues in the Food Crisis Management Group have kindly commented on the manuscript.
ANNEX: Table of Contents from a South Asian Provincial
Famine Code - Bengal, 1913 (28)
I. Standing preparations
1. System of intelligence
2. Programmes and estimates
3. Special rules for railways
4. Circle and charge organization
5. Reserve of tools and plant
II. Preliminary measures of enquiry and preparation when there
is likelihood of distress
III. Observation and test, and prolonged scarcity not amounting to famine
IV. Declaration of distress and commencement of relief
V. Powers and duties of District Boards and officer' of government, in time of scarcity or famine
Vl. Relief works
General remarks applying to all works
Class I. Public works
1. Organization and control
2. Admission and classification of workers
3. Charges and gangs
5. Payment and measurement
6. Dependents on works
Class II. Village works
1. Non-departmental works
2. Private works (aided and unaided)
VlI. Wages and allowances
VIII. Gratuitous relief
IX. Poor houses and kitchens
X. Distress caused otherwise than by drought, e.g., by flood
or sea wave
XI Rains policy and closure of relief
XlIl. Aboriginal tribes
XIV. Special relief to weavers and other artisans
XV. Charitable relief fund
XVII. Duties of Medical Officers in times of distress and famine
XVlll . Accounts
1. Provision of funds
2. Method of drawing and accounting for famine charges
3. Control and audit
II. Public works
1. Provision of funds
2. Method of accounting for famine charges
3. Control and audit
1. Classification of expenditure
Appendix I. Reports, returns, etc.
I. By the local Government
II. By the Famine Commissioner
III. By the Commissioner of a Division
IV. By the Director of Agriculture
V. By the Accountant-General
VI. By the District Officer
VlI. By the Subdivisional Officer
VlII, By the Superintendent of Police
IX, By the Chief Engineer
X. By the Superintending Engineer
Xl. By the Executive Engineer
Xll. By the District Engineer
XlIl. BY the Inspector-General of Civil Hospitals
XIV. By the Sanitary Commissioner
XV. By the District Medical Officer
XVI. By the Charge Superintendent
XVII. By the Circle Officer
XVIII. By the officer in charge of public or non-departmental works
XIX. By the Superintendent of a poor house
XX. By the Superintendent of a kitchen
XXI. By the officer in charge of orphan relief
XXII. By the officer in charge of relief of weavers and other artisans
XXIII. By the Medical Officer on a relief work charge
XXIV. By the Police Officers
XXV. By the Treasury Office
Appendix II. Note on the organization and management of public works
I. General organization and establishment
II. Equipment of a charge
lII. Funds and supply of coin
IV. Drinking water and disinfecting water supply
VI. Field hospitals
IX. Organization of relief labour: the gang
X. Tasks and systems of work
Xl. The daily routine of work: reports and accounts
Xll. Checking work, numbers and payments
XIII. Duties of Military Officers appointed to supervise departmental works
Appendix lII Note on the organization and management of poor
Appendix IV. Note on the organization and management of kitchens
Appendix V. Central Provinces Administration Resolution on the operations undertaken for the relief of weavers in the Central Provinces during the famine of 1907-08
Appendix Vl. Estimate of the number of forms required for:
(a) Officer in charge of a relief work of 3,000 labourers
(b) Circle Officer
(c) Superintendent of poor house
(d) Superintendent of kitchen
(e) Officer in charge of State orphanage
1. World Food Programme, Annual/ Report of the Executive Director on the Development of the Programme, 1982 Agenda Item 4, Fifteenth Session, Rome, 16-27 May, WFP/CFA 15/4 (1983).
2. World Bank, World Development Report (Oxford University Press, New York, 1982).
3. B. P. Acquaah, "Development of an Early Warning System on Basic Food Supplies with Particular Reference to Crop Forecasting," report to the Government of Somalia, FAO 371324 (1977).
4. Food Information Group 83, Special/ Reports: Food Crops and Shortages, Global Information and Early Warning System on Food and Agriculture (FAO, Rome, 1983).
5. M. Frere and G. F. Popov, "Agrometeorological Crop Monitoring and Forecasting" (FAO, Rome, 1979).
6. C. M. Sakamoto, "The Technology of Crop-Weather Modelling," in W. Bach and S. H. Schneider, eds., Food Climate Interactions (Reidel, Dordrecht, 1981).
7. Office for Special Relief Operations, Report of the Food Assessment Mission to Kampuchea ( FAO, Rome, 1980).
8. FAD/UNICEF/WHO Joint Expert Committee, Methodology of Nutritional Surveillance, WHO Technical Report Series, no. 593 (WHO, Geneva, 1977).
9. M. F. Lechart and C. Ville de Goyet, "Early Warning Alarm Systems of Nutritional Deterioration in Emergency Periods," Arch. Latinoamer. Nutr., 27 (2): 64 (1977).
10. M. Gebre-Medhin, "Famine in Ethiopia. ll: Sequence of Events Model of a Consolidated Food and Nutrition Information System (Early Warning System)," J. Trop. Paediat. Environ. Child Hlth., 23 (1): 23 (1977).
11. J. Kreysler, "Nutritional Surveillance in Botswana as a Method of Planning the Prevention of Hunger Situations," in Proceedings of the Symposium on Drought in Botswana (University Press of New England, Hanover, N.H., USA, 1979).
12. D. S. Mileti, Natural Hazard Warning Systems in the United States: A Research Assessment (Institute of Behavioral Sciences, University of Colorado, Boulder, Colo., USA, 1975).
13. R. N. Nag and P. Ray, "Famine Codes: Their Implications," Food Systems and Society in Eastern India Working Paper No. 1 (CRESSIDA, Calcutta, India, 1979).
14. N. Sengupta and P. R. Ray, "Relief Manuals and Codes of West Bengal and Orissa after Independence," Food Systems and Society in Eastern India Working Paper No. 25 (CRESSIDA, Calcutta, India, 1980).
15. H. Sheets and R. Morris, Disaster in the Desert: Failures of International Relief in the West African Drought (Carnegie Endowment for International Peace, Washington, D.C., 1974).
16. J. Shepherd, The Politics of Starvation (Carnegie Endowment for International Peace, New York, 1975).
17. F. Buttet, "De la Semence au Repas," Developpement (Berne, Switzerland), no. 8 (1982).
18. L. Bondestam, "Understanding Hunger and Predicting Starvation," Food and Nutr. Bull., 3 (4): 1 (1981).
19. R. B. Outhwaite, "Food Crises in Early Modern England: Patterns of Public Response," in Proceedings of the Seventh International Economic History Congress (Edinburgh University Press, Edinburgh, UK, 1978).
20. E. Waddell, "How the Enga Cope with Frost: Responses to Climatic Perturbations in the Central Highlands of New Guinea," Human Ecol., 3 (4): 249 (1975).
21. B. Currey, Mapping Areas Liable to Famine in Bangladesh (University Microfilms, Ann Arbor, Mich., USA, 1979).
22. N. S. Jodha, "Famine and Famine Policies: Some Empirical Evidence," Economic and Political Weekly, 10 (41): 1609 (1975).
23. R. Chambers, Health, Agriculture and Rural Poverty: Why Seasons Matter, IDS Discussion Paper 148 ( Institute of Development Studies, University of Sussex, UK, 1979).
24. S. Tabor and S. Untung, An Agricultural/Calendar for Central Lombok (Puslitbang Gizi, Bogor, Indonesia, 1981).
25. R. Baker, "The Administrative Trap," Ecologist, 6 (7): 247 (1976).
26. K. S. Singh, The Indian Famine 1967: A Study in Crisis and Change (Peoples Publishing House, New Delhi, 1975).
27. V. Subramanian, Parched Earth: The Maharastra Drought 1979-73 (Orient Longman, Bombay, India, 1975).
28. Government of Bengal, Bengal/ Famine Code (Bengal Secretariat Book Depot, Calcutta, 1913).
Additional Bibliography (not specifically cited in text)
Cornell University, Report of Meetings about the Indonesian Nutrition Surveillance System Project (Cornell University, Ithaca, N.Y., USA, 1981).
Currey, B., "Fourteen Fallacies about Famine," Ceres, 14 (2): 20 (1981).
Eusuf, A. N. M., and B. Currey, The Feasibility of a Famine Warning System in Bangladesh (Ministry of Relief and Rehabilitation, Government of Bangladesh, Dhaka, 1979).
Horwitz, A., et al., Third Interim Evaluation of the Indonesian National Nutritional/ Surveillance System (USAID, Jakarta, 1983).
Sisler, D. G., and D. O. Dapice, Report of Consulting Visit for the Indonesian Nutrition Surveillance Project (NIHRD, Jakarta, 1981).
Wilson, D., et al., First In-Progress Evaluation of the Indonesian National Nutritional Surveillance System (NIHRD-USAID, Jakarta, 1981).
Wilson, D., et al., Second In-Progress Evaluation of the Indonesian National Nutritional/ Surveillance System (NIHRD-USAID, Jakarta, 1982).
Contents - Previous - Next