The Rockfeller Foundation
New York, NY, USA
In my brief comments this evening, I would like to speak to you from the perspective of my participation in FIVIMS since 1997 and, in particular, to draw from my experience as chair of the FIVIMS Inter-Agency Working Group (IAWG) during the last three years. From that perspective, I wish to address some of the policy and institutional factors that are likely to influence and, in many cases, block efforts to apply the more refined measures of food and nutritional security that you are exploring during this Symposium. Then, I shall try to draw some implications for the work of the Symposium this week.
But first, for those of you who are not familiar with FIVIMS, let me provide a very brief background. The 1996 World Food Summit recognized that knowledge of the dimension and nature of food insecurity was poor, owing to weak national information systems, and that this knowledge was critically important to design and target effective policies and interventions. Even the most basic questions could not be answered in most developing countries and in many developed countries as well. These questions included:
How many people are food insecure?
Where are they located?
What is the nature of their food and nutritional insecurity?
What are the contributing factors?
These same questions could not be answered for those groups who were vulnerable to falling into food insecurity. In addition, there was little information on the shocks that can make vulnerable households food insecure, on which households are most exposed to those shocks and on their coping mechanisms to manage and mitigate shocks. Because information was lacking to answer most of these questions, the WFS - at the strong urging of the group of Latin American countries in particular - included the commitment to strengthen national information systems in the final Summit Plan of Action. At the same time, UN agencies also committed themselves to cooperate more closely in assisting developing countries to build national capacities to collect, manage, map and analyse information on food security. The UN agencies also committed themselves to monitor more closely, and with greater harmony, trends and patterns in world hunger in order to develop and deliver more effective global programmes.
FIVIMS and the IAWG are the concrete manifestations of those commitments. With its Secretariat in FAO, the IAWG includes all UN agencies working on food security issues as well as key bilaterals and NGOs. The most active membership includes, in addition to FAO, UNICEF, WFP, WHO, UN Development Programme, International Fund for Agricultural Development, World Bank, IFPRI, USAID, DFID, Helen Keller International and, very recently, the Rockefeller Foundation.
FIVIMS works at two levels, global and national. At the national level, rather than introducing new information systems, it works to build upon, strengthen and rationalize the systems that already exist. In 1998, FIVIMS embraced a holistic conceptual framework that encompassed the range of determinants of food and nutritional insecurity and vulnerability. These included not only food availability, but also personal income that drives effective demand and access, education, health care, access to water and sanitation, and human health dimensions that drive biological utilization. More recently, FIVIMS has firmly embraced the sustainable livelihoods approach to better characterize and understand the determinants of household vulnerability and coping. Since its launch, FIVIMS has had many undeniable achievements, including work in more than 50 countries. But it is fair to say that we still have a long way to go to fully meet the ambitious goals set for FIVIMS in 1996.
What are some of the factors that have limited our progress? And what have we learned from our work that might be useful to the challenges being addressed in this Symposium? Let me focus on several such factors. First, I think an important starting point is to recognize honestly that food security and indeed agriculture are low priorities in most developing countries. This is reflected in turn in a disproportionately low share of government development resources allocated to food security programmes. As a result, there is a low demand for accurate, timely and spatially disaggregated information, and there are few government resources devoted to data collection and analysis. Where food security information systems exist, the components are almost always split among different ministries and designed to address different objectives. Data sharing across these agencies is the exception rather than the rule, and as a result, there is little or no integrated cross-sectoral analysis. Ministries of agriculture, health, land and water, and bureaus of statistics and census are often all active collecting potentially complementary data, but they generally use different spatial coverage, sampling procedures, survey frequencies and data management systems. All of this is generally characterized by an official unwillingness to share or pool data or to collaborate in data analysis.
Unfortunately, the behaviour of most of our agencies has generally made matters worse. Rather than promoting integrated and more cost-efficient systems, specialized agencies typically continue to work to strengthen the independent capacities of the ministries with which they are affliliated. And this situation prevails, despite the fact that FIVIMS was established to promote greater cross-agency cooperation and integration. Basically, old habits die hard. Agency networks and pro-grammes create their own momentum and resist reform. This is especially true at the country level to which most UN agencies have devolved considerable decision-making authority in the last decade.
The situation is made even worse by shrinking agency budgets, which in turn put greater pressure on individual agencies to try to distinguish their programmes by developing - and defending the independence of - what we might call information brand products. Agencies generally promote different definitions of food security and vulnerability and use vastly different indicators and methods to measure those indicators. Thus, we have the WFP Vulnerability Assessment and Mapping System, the USAID Famine Early Warning System, a range of nutritional surveillance systems and the infamous FAO method. What this means is that rather than being a force to consolidate methodological harmony within governments, we often serve only to encourage even greater fragmentation.
One of the experiences of FIVIMS is instructive in this regard. Over the last three years, a working group has tried to develop a single approach to define and measure vulnerability. After several meetings over two years, efforts essentially ceased. We simply could not agree on a single comprehensive approach. Instead, we agreed to disagree on the specifics but to try to define common ground in an over-riding framework within which different methods could be situated and interpreted.
And maybe this is the most realistic second-best solution. After all, each agency is driven by its specific programmatic mandate, within which its unique approach makes eminent sense. The problem, however, is when we advise governments that cannot afford the luxury of competing approaches and that need efficient and cost-saving methods. The primary impact at the country level is one of confusion: different conceptual frameworks, different methods, different units being created within national institutions ... and all chasing the same shrinking national resources.
So what is the way out? What might be some of the implications for your work at this Symposium? Let me make three closing observations on what might be qualities of a sound national system, and therefore criteria to use in selecting and adapting approaches to national conditions.
My first point is the critical importance of a sensitive and unbiased assessment of national demand for information. It seems obvious that if we want a method to be used, it has to be useful - in particular, useful to help policy-makers in the daily decisions that they themselves consider important. This means that one has to begin by asking those same policy-makers what is the minimum priority information they need and would actually use. And given government programmes, how timely does the information have to be? And disaggregated to what geographical level? One also has to ask what is the capacity of national agencies to absorb and use information. Answers to these questions will, of course, differ across countries and within the same countries over time. This suggests that a single blueprint or cookbook approach is unlikely to be helpful. Rather, conceptual frameworks and methodological principles - all of which can be adapted to local needs, conditions and resource realities - are likely to be most useful.
My second point is the importance of cost-effectiveness or, more simply: keep it simple and keep it cheap. The message here is that we need to internalize the fact that national budgets for information are, and will continue to be, grossly suboptimal for some time. There is no point in promoting an elegant and insightful method that exceeds the treasurys capacity to sustain financially supporting the use of the method.
But the other side of that constraint is the opportunity to design and promote methods that cut across, piggyback upon and rationalize existing information systems. In most countries, large budget efficiencies could be obtained by reducing duplication and integrating efforts across government units. However, this requires much greater coordination and much less turf consciousness and branding among development agencies and government institutions, as well as among researchers.
This brings me to my final point. The situation could be much improved if there were greater transparency and clarity in stating what questions particular approaches can answer, and even greater honesty in admitting what questions they cannot answer. Rather than hearing competing claims that sell one institutions approach while discrediting others, what developing countries really need is a fuller disclosure of what methods are cost-ef- fective for what purposes, for what questions that need to be answered, with what precision and, as importantly, at what cost, both financially and in terms of human resources. In short, our collective responsibility is to help governments make an informed choice and to help them adapt the most appropriate set of methods to their particular circumstances.
Fortunately, there are a number of international initiatives that promise greater attention to, and demand for, good information. These initiatives also offer the possibility of real coordination and of additional resources at the country level. Over the past three years, as part of the UN reform effort, the UN system has introduced the Common Country Assessment (CCA) process. This calls for UN country teams to come together and work collaboratively with governments to diagnose their development challenges and priorities. The CCA process provides an opportunity to tailor a sustainable national information system to help guide and monitor the impact of development programmes. Similarly, the World Bank has launched the Poverty Reduction Strategy Process aimed at guiding the formulation of strategies to reduce human poverty. The starting point is again a national diagnosis that relies upon a strong information system, within which food security must be a core element. And finally, within the Millennium Development Goals process, national Millennium Development Reports will be prepared. These national reports can, and indeed must, go beyond the simple indicators included in the MDGs to provide a more nuanced and locally relevant profile of human poverty and food and nutritional insecurity. Now, considerable efforts are being made to integrate and harmonize these three processes, and I for one am optimistic that they will succeed.
The discussions at this Symposium can contribute to this process by providing greater clarity and, hopefully, an emerging consensus on the strengths and weaknesses of currently competing methods. My sense is that the methods that are finally adopted, and adapted, by developing countries will be those that are:
truly demand-driven at the country level;
those that promote cross-sectoral integration;
those that are simple and cost-effective;
those that are transparent about the sets of questions they can answer efficiently, as well as those they cannot.
I very much look forward to see how far we have come this week towards these ends.