RÉSUMÉ

Il est fondamental d'identifier des coûts et des approches rentables permettant de contrôler la sécurité alimentaire et la nutrition en vue d'accroître l'utilité et la viabilité des programmes y relatifs. Les systèmes de contrôle de la sécurité alimentaire et de la nutrition devraient être périodiquement évalués afin d'assurer que leurs objectifs sont atteints et réorientés vers les besoins changeants en informations pour des interventions de politique alimentaire. Le lien entre les informations générées par ces systèmes et les processus de planification et d'ordre politique pour améliorer la sécurité alimentaire demeure faible dans de nombreux pays de l'Afrique subsaharienne. Le présent exposé met au point un cadre permettant d'évaluer les systèmes de contrôle de la sécurité alimentaire et de la nutrition. Parmi les critères utilisés dans l'évaluation figurent la capacité de traiter les données, l'analyse et la présentation des rapports, la production d'informations à temps, l'engagement des décideurs et la rentabilité. En utilisant ce cadre, cet exposé évalue les programmes de « Système national d'alerte avancée » et de Système national d'alerte rapide en cas de famine » en Ouganda afin d'en tirer des leçons utiles pour d'autres pays de l'Afrique subsaharienne. Les auteurs pensent qu'un système fiable de contrôle de la sécurité alimentaire et de la nutrition est simple et orienté vers l'utilisateur, et qu'il a l'engagement des décideurs concernés qui utiliseront ces informations lors de la planification et la conception de la politique sectorielle. Cet exposé souligne qu'une l'évaluation fréquente des méthodes et des systèmes de contrôle de la sécurité alimentaire est essentielle pour sous-tendre l'engagement des décideurs. En outre, des évaluations continues faciliteront les systèmes de contrôle qui sont déjà en place pour qu'ils deviennent des systèmes qui englobent la sphère de la sécurité du bétail. L'utilisation de l'approche relative au bétail dans le contrôle de la sécurité alimentaire appuiera des solutions préventives et dynamiques plutôt que des approches curatives en cas d'insécurité alimentaire.

Mots-clés: Alerte rapide, sécurité alimentaire, nutrition, contrôle, évaluation

The main objective of an early warning system is to monitor food security and nutritional status so to inform decision makers of impending food shortages at national and local levels. The response that an early warning system triggers is an important indicator of its success. The appropriateness of this response, in turn, depends on the links between the information generated by the system and the policy making process. This linkage remains weak in several sub-Saharan African countries. The recent food shortages experienced in Southern Africa and countries like Sudan and Ethiopia, are partly due to the failure by governments and international organizations to take public action of ensuring food security despite the presence of early warning systems in the countries [1, 2]. This is partly because food security monitoring systems, which are mostly established by external or donor funding, have not been fully integrated into the decision making processes of the government. Even though there has been limited analysis on the impact of monitoring systems on food security decision making, there is increasing interest on the part of African policy makers to improve the use of information provided by these systems. The nature and role of food security monitoring systems in meeting the information needs required by decision making processes vary, depending on the type and objective of these systems. However, a carefully outlined conceptual framework allows us to evaluate the performance of food security monitoring systems, and helps us move towards re-linking them to the policy making process.

Few attempts have been made to define the criteria needed for understanding the impact of the monitoring systems [3]. More importantly, there has not been any systematic evaluation of the role of monitoring systems in the decision making process. Continuous evaluation of the benefits of food security monitoring systems is necessary to ensure sustainability and allow them to alleviate food insecurity. More work needs to be done in assessing the operational aspects of food security monitoring systems. The quality and accuracy of data, appropriateness of data processing procedures, effectiveness of information disseminated, and its use for planning and policymaking purposes all need to be assessed. Such evaluations will re-orient food security monitoring systems towards the changing needs of its clients.

This paper develops a framework for evaluating food security and nutrition monitoring systems in sub-Saharan Africa. Using Uganda as a case study, it evaluates the National Early Warning System (NEWS) and the Famine Early Warning System (FEWS) in order to gain useful lessons for other countries in sub-Saharan Africa.

NEWS are generally implemented by the governments with the help of external donors and executing agencies, such as the Food and Agricultural Organization (FAO) of the United Nations (UN) or the United States Agency for International Development (USAID), and focus on monitoring food security at the national level. FEWS on the other hand are initiated and implemented by external agencies independently or in collaboration with the host country governments. They monitor food needs of the vulnerable population to prevent famine-related disasters [4, 5].

The Ugandan case study presented in this paper is based on a historical review of patterns of food security monitoring - on personal interviews with program managers of the monitoring systems, - and users of information in government and non-government organizations. The authors have done their best to reflect the changing nature of NEWS and FEWS over the years and that the issues identified in this paper remain valid.

The paper is organized as follows: First, we provide a conceptual framework for evaluating food security and nutrition monitoring systems. Next, based on this framework, the two parallel systems of food security monitoring in Uganda are evaluated. Consequently, lessons are drawn from these systems in designing and modifying food security monitoring systems to maximize information generation capacity to use in alleviating food insecurity. Finally, concluding remarks are presented in the last section.

Periodic monitoring of household food security and nutritional status is fundamental to generating adequate information on the its state, as well as the factors that influence it. The information collected through monitoring systems is a key ingredient for designing appropriate actions that will improve the food security and nutritional status of households. It is also important to assess the impact of these interventions with regard to improvements in food security and nutrition. Thus, food security and nutrition monitoring is defined as a process of monitoring, analysis, and interpretation of indicators and causal factors associated with household food security and nutrition, in order to make appropriate decisions that will lead to effective interventions to improve the food security and nutritional status of the population.

Lack of evaluations on the performance of food security and nutrition monitoring systems has been frequently suggested as the reason for the unsustainable nature of monitoring systems, which continue to be heavily donor-driven and dependent on external aid and technical assistance [6]. Monitoring systems should be evaluated on two levels: first, the performance of monitoring systems should be examined to determine the quality of information and whether it meets the information needs of users; second, the impact of information on policy decisions should be measured.

A sound food security and nutrition monitoring system is simple, user-driven, based on existing institutional structures (which increases the capacity for analysis and interpretation) and has the commitment of relevant decision makers who will use the information in planning and policy design. Tracking the deviations in implementing food security monitoring systems from these criteria helps to re-orient their activities towards the ultimate goal of forming informed food security decisions.

Figure 1 lays out a conceptual framework for evaluating food security and nutrition monitoring systems. We view this evaluation as a cyclic one, with the criteria used to evaluate the performance of the monitoring systems presented in the circles on the periphery.

A fundamental factor that determines the sustainability of food security monitoring systems is the user-driven nature of the objectives towards which information is generated. Operational linkages between the monitoring system and the institutions using the generated information need to be in place for the system to be successful in triggering appropriate policy response [7, 8].

The quality of data and the speed with which it is generated is determined by the simplicity of the instruments used for gathering information. The use of existing infrastructure for collecting and compiling information has proven to be more successful, than the creation of new institutional structures for purposes of food security monitoring. Early warning systems can [and should] utilize traditional methods used by local communities for early warning purposes [9].

Early warning systems should be evaluated on the accuracy of the data they produce as well as their capacity for data processing, analysis and interpretation. It is important that the information produced arrives in a timely fashion, giving decision makers enough lead way to produce the necessary policy responses. The same decision makers should be committed to using the data produced by these systems.

The cost of information generated should be weighed against the benefits of policy impact. Even though it is generally argued that such benefits are not readily quantifiable, one approach is to document the use of information from food security and nutrition monitoring systems, for various planning and policy making purposes on a case-by-case basis, and estimating cost-savings compared to generating the information separately. Even though the donors frequently raise the cost of food security and nutrition monitoring as a concern, very few attempts have been made to measure the actual costs and benefits of these systems. Empirical evidence from Malawi indicates that the internal rate of return to food security and nutrition monitoring ranges from 66 to 75 percent [10]. Another study shows that every dollar spent on mitigation saves approximately $4-10 in recovery costs [11].

It is conceivable that donor agencies use part of the information collected by monitoring systems in their planning exercises and save donor resources, that would otherwise go towards data collection. Such benefits are over and above the benefits of information for which the monitoring system were originally intended. Furthermore, transparency in the use and sharing of data will enable faster response to food emergencies and limit the replication of data collection efforts [12].

Food security monitoring should eventually lead to institutional development, so that the information generated can be effectively used in decision making. Past experiences in food security monitoring systems indicate that there is a tendency towards using shortterm technical assistance in generating data from the field. Even with long-term projects, such as the ones implemented by FAO, there have not been adequate resources to develop institutional and human capacity to sustain and justify the existence of monitoring systems.

Monitoring the use of information in designing policies and intervention programs and evaluating the impact on policy decisions is important for identifying new channels for information dissemination. Information obtained from evaluating the benefits of monitoring systems in influencing policies can be a useful instrument for raising resources and support for sustaining them. Monitoring and evaluation of the impact of the output from monitoring systems in influencing policy decisions requires constant follow up of the flow of information and the documentation of this information, at various stages of decision making. This requires special efforts on the part of the monitoring systems' program managers. The benefits of such documentation, however, outweigh the time and cost involved.

Although food security early warning systems have been implemented for the past 20 years in several sub-Saharan countries, only a handful of these countries can be presented as success stories in using early warning information for designing effective interventions in alleviating food insecurity and malnutrition [4, 13]. Even in those countries that have established an operational early warning system, several issues relating to the principles of monitoring food security need to be addressed. Some emerging lessons from the evaluation of the early warning systems in Uganda, along with similar experiences in other sub-Saharan African countries are presented in this section.

For successful implementation of early warning systems, having a critical mass of people trained in food security issues from various sectors, including universities and other academic institutions, is a prerequisite. Lack of such capacity to translate the data collected by monitoring systems into policy decisions and to design interventions poses a formidable challenge in all sub-Saharan countries [16]. Strengthening the capacity of subject matter specialists, such as agronomists, livestock specialists, and home economists in national research organizations dealing with food security and nutrition issues, is essential for fostering policy debate and dialogue at the local and national level. This requires a multi-disciplinary approach to capacity strengthening in food security and nutrition policy analysis. Past experiences in such an approach indicate that it is possible to train social scientists in food and nutrition issues, and nutritionists and food scientists in policy and planning methods [17]. This is important, given the overall limited capacity, for designing interventions in the region.

Despite the capacity strengthening efforts of several donors and institutions, the number of experts on food and nutrition monitoring in sub-Saharan Africa has remained very low [16]. In Uganda, even though physical capacity for data collection is adequate, it requires further strengthening to ensure improved quality of output from NEWS. In addition, food security analysis capacity has to be strengthened at decentralized levels [9]. Experiences from Malawi indicate that through constant monitoring and training of nutritionists and social scientists at various levels over a period of three years, it is possible to place necessary capacity in place for information generation and analysis [18].

Information obtained through early warning systems can be used for emergency planning, program design, and policymaking objectives in a wide range of sectors including; agriculture, health, economic planning, social welfare, labor, and famine relief. To meet the diverse information needs and the various objectives of food security interventions, early warning systems should be flexible. Such flexibility does not exist in the currently implemented monitoring systems in Uganda. The case with the food security monitoring systems in Botswana [19], Zambia [20], or Tanzania [21] is similar. For example, the flexibility in Malawi's early warning system enabled a quick transformation of an information system originally designed for policy planning into a system supporting the information needs for drought management as demonstrated by the recent efforts to prevent famine in Southern Africa [22].

The immediate need in sub-Saharan Africa is the establishment of multi-sectoral committees on food security and nutrition that involve agriculture, health, transport, trade and industry, natural resources ministries, as well as academic and research institutions. Such committees may also include donor representatives as observers. Although various sectoral ministries could implement early warning systems to meet their planning and policy making objectives, it is important to have a focal office at the national level to bring together the outputs of the various systems operating in a country. Without a central committee to coordinate the activities of early warning systems, messages from different sources can lead to confusion and delay effective response [8]. For example, during the periods of drought management in Malawi, the Disaster Preparedness Committee, coordinated by the Office of the President and Cabinet, acted as the decision making group and was responsible for dissemination of information to donors and other international organizations. This substantially reduced the duplication of informationgathering efforts by several independent and non-governmental organizations [22].

Experience from other sub-Saharan African countries shows that when food and nutrition information systems are placed at higher levels such as the Office of the President or the prime minister, developing an ongoing advocacy campaign to promote a common understanding of food security and nutrition issues, as well as keeping them at the forefront of the policy dialogue becomes much easier [23]. Unfortunately, in several countries, including Kenya, Zambia, Zimbabwe, and Swaziland, the efforts to organize information generation and use have been hamstrung by bureaucratic barriers against coming together through multi-sectoral food and nutrition committees [6].

One of the major constraints in developing comprehensive early warning systems in sub-Saharan Africa has been the limited infrastructure to collect and process the data from the field. While such an infrastructure exists in Uganda, it is poorly developed to meet all data generation needs. In countries with a well-developed network of data-collection systems, it is generally suggested that the existing infrastructure should be used effectively [24]. In countries where there are no organized data-collection systems for food-security monitoring, such as Eritrea and Namibia [25], it may be necessary to create new infrastructure. Elsewhere, it is possible to combine the existing but poorly organized structures and reorient them towards specific goals of food security monitoring. In Uganda, as in Malawi, Botswana, and Kenya, the use of data-collection institutions that were in place for the collection of information in farm management and agricultural surveys could make possible cost-effective food security monitoring systems [10].

Data for Who?
Collecting and analyzing information will not by itself lead to successful policy implementation. According to Habicht [2000] all evaluations should start with the question, "Who will use the information and for what purpose?" The lack of continuous assessment of the user needs rendered the output of NEWS "data-driven" and thus, less useful to the policy makers in Uganda. One way of preventing the efforts spent on data collection and analysis from becoming futile is to involve the decision makers in the generation of information. Another option is to perform a user needs assessment [27]. For example, during the 1991 to 1992 drought in Malawi, the designers of the monitoring system constantly evaluated the information needs for drought management planning by working closely with the Disaster Preparedness Committee. This increased user participation in the monitoring system and reduced the probability of its becoming a process-driven system [22]. Another lesson learned from Namibia and Zambia is that reducing the number of data variables will increase the use of information in decision making [20].

The early warning system established in Uganda has the potential of generating situation analysis reports within four to six weeks from the time of data collection if the human capacity of NEWS is improved with additional resources for field operations. This is possible in Malawi, because of the trained human capacity and resources for data processing, analysis, and preparation of food security working papers at the district level [28].

In addition, the availability of timely information for making appropriate decisions and the quality of information presented to the decisionmakers play an important role in strengthening the information-action linkages. For instance, the quarterly nature of NEWS bulletins in Uganda make them less timely for meeting immediate information needs. The time gaps between collection of data and analysis, and between analysis and reporting, have been so wide in the past that when the information is given to the decision makers, it is too late for them to make effective use of it. To minimize this time lag, it is essential to decentralize the data processing and analysis systems. In doing so, attention should be paid to developing appropriate but flexible computer data processing systems at the district levels.

One of the major criticisms of the currently operating monitoring systems in sub-Saharan Africa is that of the large volume of data that is collected, only few are analyzed, and even fewer are reported from the subset that have been analyzed. Another issue to focus on is the presentation of the material. The analyses should strive to be as transparent as possible and clearly communicate to policymakers through different media outlets, including printed press, radio, town meetings, and computer networks [27]. For instance, one of the reasons identified for the recent food shortage experienced in Malawi is the failure by the media to identify the government and donors in time to avert the crisis [29].

Prior to the famine of 2002 in Malawi, two different systems predicting the situation of food security in the country reached different conclusions. While an assessment by FEWSNET Malawi - a sister program of the FEWS network in Uganda - did not clearly identify the famine, another assessment by Save the Children-UK (SC-UK), predicted that barring lower prices of maize by 2002 was inevitable. Why was the SC-UK able to foresee the crisis? As pointed out by Seaman [2], this was possible because SC-UK, through a vulnerability analysis [using the Household Economy Approach (HEA)], was able to gather information on people's livelihoods in an accurate and geographically disaggregated fashion. The FEWSNET assessment, on the other hand, despite identifying a potential problem, was not able to break down aggregate food supply to the level of household access, to food.

Although, NEWS in Uganda is helpful in gathering information at the district level, it falls short of providing information on household food security and on vulnerable population groups. Given that there is food self-sufficiency at the national level, at least during the years of normal rainfall, and that the major food insecurity threats are at the household level, there is a need to go one step further to the household level to collect information and formulate programs and policies to mitigate the household food insecurity.

Malawi was the first country in southern Africa to establish a food security and nutrition monitoring system at the household level [28]. Based on lessons learnt from the Malawi experience [30], a Regional Food Security and Nutrition Information System has been established by the Southern Africa Development Community (SADC) in selected member countries (Mozambique, Swaziland and Zimbabwe) with a focus on identifying probable causes of household food insecurity [31].

There has been a recent shift in early warning systems towards using livelihood approaches similar to that of SC-UK's in identifying impending food-related disasters. For instance, the Famine Early Warning Systems Network (FEWSNET) now employs "… food economy principles and techniques to ensure a quantifiable, and therefore comparable output from its livelihood baselines. These baselines are used as the reference point against which to judge the effects of current shocks." [32] To do so, FEWSNET has started working in collaboration with organizations like Save the Children and OXFAM. Even though this shift towards utilizing livelihood approach deserves applause, such baseline reports are limited to only a few countries and subregions and FEWSNET does not currently have the capacity to start using these reports in actual food and nutrition monitoring. Also, the decision to establish Food Insecurity and Vulnerability Information and Mapping Systems (FIVIMS) following the World Food Summit of 1996, is promising for the identification of food insecurity and malnutrition at the household level [27].

Recently, there has been considerable interest among planners and policy makers in identifying cost-effective and time-efficient approaches to monitoring the food security status of developing-country households. Enhancing the effectiveness of food security and nutrition monitoring systems requires the use of recent innovations in data collection, processing, analysis and interpretation by planners, designers and field staff. This can be made possible through the evaluation of monitoring systems operated for various objectives and under different circumstances. The outcome of such evaluations should be constantly reinforced by additional lessons from the implementation of food security and nutrition monitoring systems from other countries. The evaluation of NEWS and FEWS in Uganda, presented in this paper, is an attempt in this direction.

The focus on the two programs has brought to light some important issues that need to be addressed in food and nutrition monitoring systems. These include the need for establishing multi-sectoral committees on food security and nutrition issues in Africa, building human capacity in food security, ensuring the flexibility of early warning systems so that they can be used in emergency planning, program design, and policy making in various sectors, strengthening the infrastructure for data collection, identifying and acknowledging the users of the data, and generating timely information that will be analyzed and reported accurately.

Frequent evaluation of the methods and the systems of monitoring food security is essential for generating continued interest by decision-makers. This will also provide an opportunity for the existing food security monitoring systems to evolve into those that encompass spheres of livelihood security and bring preventive and proactive solutions rather than curative approaches [13]. A majority of food security monitoring systems continues to be based on macro-level information and operated at the national level. Future challenges in addressing household food security will require devolution of the process of food security monitoring and intervention systems to decentralized levels. In meeting these challenges, the role of continued evaluation of food security monitoring systems for their impact on food security decisionmaking cannot be overemphasized.

1. Deng BL Famine in the Sudan: causes, preparedness and response: a political, social, and economic analysis of the 1998 Bahr el Ghazal famine, IDS Discussion Paper 369, 1999.
2. Seaman J Rethinking early warning systems, humanitarian exchange, 2002; 22:1013, http://www.odihpn.org/pdfbin/newsletter022.pdf.
3. Habicht JP and P Pinstrup-Andersen Principles of nutritional surveillance. Pew/Cornell Lecture Series on Food and Nutrition Policy, Cornell University, Ithaca, New York. 1990.
4. Buchanan-Smith M, Davies S and R Lambert A Guide to famine early warning and food information systems in the Sahel and Horn of Africa. Research Report No. 21, Vol. 2. Institute of Development Studies, University of Sussex, 1991.
5. Davies S, Buchanan-Smith M and R Lambert Early warning in the Sahel and Horn of Africa: the state of the art: A review of the literature. Research Report No. 20, Vol. 1. Institute of Development Studies, University of Sussex, 1991.
6. Quinn VJ and E Kennedy Food security and nutrition monitoring systems in Africa, Food Policy, 1994;19:234-254.
7. Habicht JP Nutritional surveillance in China: results of user survey. Mimeo, Cornell Food and Nutrition Policy Program, Ithaca, New York, 1991.
8. Report for the Second International Early Warning Conference (EWC II). Synthesis of the findings of the early warning regional consultations in Africa, Asia, the American Hemisphere and Europe. Bonn, Germany, October 16-18, 2003.
9. IDNDR. International Decade for Natural Disaster Reduction. Guiding principles for effective early warning, The Convenors of the International Expert Groups on Early Warning of the Secretariat of the International Decade for Natural Disaster Reduction IDNDR Early Warning Program, Geneva, October 1997.
10. Babu SC and GB Mthindi Costs and benefits of informed food policy decisions - a case study of food security and nutrition monitoring in Malawi, Quarterly Journal of International Agriculture, 1995;34-3:292-308.
11. Benfield Creig Hazard Research Center Center for Research on the Epidemiology of Disasters and Actionaid, 2002, Development at risk: The brief for the World Summit on Sustainable Development, Johannesburg, South Africa, 26 August- 4 September, 2002.
12. Pelletier D 1993 Towards a UNICEF nutrition information strategy: Report of a meeting. Nutrition Section and Evaluation and Research Office, UNICEF New York, May 1993.
13. Davies S Adaptable livelihoods: coping with food insecurity in the Malian Sahel Institute of Development Studies, University of Sussex, 1996.
14. Babu SC and B Chaura Facing donor missions with informed policy decisions - lessons from food security and nutrition monitoring in Malawi," Africa Development, 1997:22.
15. IGAD Inter-Governmental Authority on Development. Assistance for strengthening the national food information system in Uganda - project document. Inter-Governmental Authority on Development, Djibouti, 1994.
16. Babu SC, Mthindi GB and D Ng'ong'ola Developing decentralized capacity for development policy analysis - lessons from food security and nutrition monitoring in Malawi. African Development Review, 1996;8:127-145.
17. Babu SC Rethinking training in food policy analysis: how relevant is it to policy reforms? Food Policy, 1997;22:1-9.
18. Babu SC Multi-disciplinary capacity strengthening for food security and nutrition policy analysis - lessons from Malawi. Food and Nutrition Bulletin forthcoming, 1997.
19. Belbase K and R Morgan Food security and nutrition monitoring for drought relief management. Food Policy. 1994;19:285-300.
20. Eele D Indicators for food security and nutrition monitoring. Food Policy. 1994;19:314-328.
21. Pelletier D and U Jonsson The use of information in the Iringa nutrition program: some global lessons for nutrition surveillance. Food Policy, 1994;19:301-313.
22. Babu SC and E Chapasuka Mitigating the effects of drought through food security and nutrition monitoring: lessons from Malawi. Food and Nutrition Bulletin, 1997;18:71-83.
23. Quinn VJ A history of the politics of food and nutrition in Malawi: the context for food and nutrition surveillance. Food Policy, 1994;19:255-271.
24. Arnauld J, Alarcón JA and MDC Immink Food security and food and nutritional surveillance in Central America: the need for functional approaches," Food and Nutrition Bulletin, 1990;12:26-33.
25. Botes A Household food security and nutritional surveillance in Namibia: the methodology and its use in development and policy interventions. Paper presented at the Southern Africa Regional Conference on Household Food Security and Nutrition Monitoring for Development Planning and Policy Interventions, SADCC/UNICEF/USAID/UNIFEM/CFNPP, Mangochi, Malawi, 7-9 April 1992.
26. Habicht JP Evaluation and monitoring: Who needs what information and why do they need it? Food and Nutrition Bulletin 2000;211:87-90.
27. IFAD. International Fund for Agricultural Development. Guidelines for National FIVIMS: Background and Principles, Second Meeting of the Inter-agency Working Group (IAWG) on Food Insecurity and Vulnerability Information and Mapping Systems (FIVIMS), April 15-17, 1998.
28. Babu SC and GB Mthindi Developing decentralized capacity for disaster prevention: lessons from food security and nutrition monitoring in Malawi," Disasters 1995;192:127-139.
29. SARPN. Southern African Regional Poverty Network, Poverty Brief, March 2003.
30. Babu SC Improved policies through food security and nutrition monitoring. Food Policy. 1992;5:384-386.
31. Mugwara R Linking relief with development in Southern Africa: A SADC perspective on the 1991/92 drought emergency, IDS Bulletin, 1994;25:92-95.
32. FEWS NET. Famine Early Warning Systems Network http://www.fews.net/livelihoods/about.cfm.

Copyright 2005 - Rural Outreach Program