National Food Security Information Coordinator
Office of the President
Let me begin by thanking the organizers of this Symposium, IAWG FIVIMS Secretariat and the authors who have presented various papers on the Measurement and Assessment of Food Deprivation and Undernutrition.
In my presentation I will give an overview of the current food insecurity situation in Kenya, the information systems that generate data for food security and the utility of the methods that have been presented here. Lastly I will give a brief outlook on the way forward in improving existing food security information system in Kenya.
Kenya has a total surface area of approximately 587 300 Km2, of which 576 300 Km2 are land area. More than 84 percent of the land area is arid or semi-arid and only 16 percent is high and medium potential. Approximately 80 percent of the population resides in the medium and high potential areas, deriving their livelihood from rain-fed agriculture.
The food security situation in Kenya
Kenyas major staple food crops are maize, wheat rice and beans. The total consumption of cereals and legumes outstrips local supply. Over the last decade, the production of maize has been below consumption levels. During the same period, the country has been self sufficient in beef, mutton and milk.
Kenya relies heavily on the agricultural sector for economic growth, foreign exchange generation and employment creation. Agriculture accounts for 30 percent of the Gross Domestic Product and is also the main source of livelihood for 80 percent of the population. The growth in the sector has shown a declining trend over the years. Other indicators of well being of the population, including food security, have shown a deteriorating trend. Recent estimates based on the Welfare Monitoring Survey (1997) show that 53 percent of the rural population and 49 percent of the urban population live below the poverty line.
There are two categories of food insecurity in Kenya - chronic and seasonal (transitory). Chronic food insecurity, whose underlying cause is low access to production resources, is widespread in the medium and high potential regions where population densities are high and productive land is not accessible. Seasonal food insecurity, on the other hand, results from environmental and economic shocks. It largely affects the vast arid and semi-arid areas that experience frequent droughts, resulting into high livestock losses. It is estimated by the Ministry of Agriculture that about 2 million people suffer from chronic food insecurity.
According to the FAO estimates published in the State of the Food Insecurity (SOFI), around 13.4 million of Kenyas population is undernourished. A comparison of these estimates with the prevalence of food poverty that affects half the Kenyan population, approximately 15 million persons using the 1999 population census data, would at first sight show some correlation. The FAO national estimates of undernourishment, however, mask the regional disparities and other localized areas of high prevalence of undernutrition in the country. There is a need to develop methodologies to capture subnational levels of chronic undernutrition.
The prevalence of food insecurity increases drastically during times of environmental shocks. During the years 1999-2001 and part of 2002, nearly 3 million people in the country suffered from food insecurity and received food aid, especially in 25 districts in the arid and also in medium and high potential regions. Previously, the medium and high potential regions of the country had not experienced extreme drought; however, the drought of 1999-2002 the affected a considerable number of people in the agriculturally potential regions.
Data on nutritional status in Kenya mainly capture children below 5 years of age. The anthropometric indicators used are stunting, wasting and underweight. As indicated, the nutritional status of the under fives, a proxy for well being, has deteriorated over the years. Children classified as nutritionally stunted declined from 37 percent in 1982 to 33 percent in 1994. According to the Multiple Cluster Indicator Survey conducted by Government of Kenya and UNICEF, about 35 percent of the children are classified as nutritionally stunted. I note that the country is losing ground on the improvements observed between early 80s and early 90s. This increase is worrying because looking at the differentials in nutritional status in country, there are particular areas with prevalences that are higher than the national averages.
The impact of the HIV/AIDS pandemic on food insecurity is not well documented in the country. However, it is apparent that the ability of households to produce will be compromised when adult members fall sick or die. As the pandemic spreads, we can only imagine the impact it is likely to have, especially in the western parts of the country where the prevalence is high for both HIV/AIDS and poverty.
Causes of food insecurity
Some of the causes of food insecurity as highlighted by the Inter-Agency Task Force on the food security in the Horn of Africa include:
low productivity of agriculture;
weak knowledge and information systems;
narrow livelihood base;
low educational standards.
To address these underlying causes, an integrated information system is required.
Existing information systems
I will give a brief overview of the type of information systems in Kenya that use some of the methodologies discussed in this symposium.
The agricultural information system is a well-established data collection framework that collects regular data on crop produc- tion of the main agricultural crops, livestock production, rainfall, market prices and other information. The production data are used to project national food consumption and the national requirements.
The child health and nutritional information system within the Ministry of Health (CHANIS) is a facility-based information system that collects data on anthropometry, especially of children. The nutritional indicator used is weight-for-age. Other surveys are conducted, but not on a regular basis, therefore CHANIS forms a very good basis for us, especially for the development of Food Insecurity and Vulnerability Information and Mapping Systems (FIVIMS).
The Household Welfare Monitoring Surveys, conducted within the National Sample Survey Programme by the Central Bureau of Statistics, collect information at the household level. Anthropometry is collected through the nutrition module, providing information on height, age and weight of children, especially those under five years, in order to determine the nutritional status of the population. In addition, data are collected on income, expenditures and agricultural production. The National sample also forms a frame for other studies such as Kenya Demographic and Health Survey, which was conducted in 1998 and collected information on BMI to determine the nutritional status of adults. BMI is generally not collected in the welfare surveys; however, its usefulness in determining adult nutritional well-being is becoming apparent.
Early Warning Systems present in the arid and semi-arid lands by the Arid Lands Resource Management Project (ALRMP), Office of the President, collect information on livestock and agricultural production, income-generating activities, food purchases and nutritional status of children under five years as assessed by the mid-upper arm circumference method. The existing information system provides data that can be used to analyze the causes of food insecurity based on the FAO FIVIMS Conceptual Framework, namely, socio-economic characteristics, food availability trends and supplies, stability of food supplies and access, as well as health status and nutritional status.
Utility of the symposium methodologies
To address food insecurity in Kenya and to design appropriate interventions, it is important to identify who the food-insecure or nutritionally stunted are, where they live, why they are vulnerable, the risks they face and how they cope with these risks. The methodologies described in this Symposium are useful, but they must be used in combination in order to answer key questions both at the national and the subnational level. With decentralization and the poverty reduction strategy, most of the information is being demanded at the subnational level to enhance district level planning.
The FAO methodology is useful at the national level as well as for regional level comparisons. On the other hand, it is not useful at subnational levels because all the data is aggregated at the national level.
Anthropometry is useful, both at the sub-national and the household level, but also at the national level. In Kenya, the Central Bureau of Statistics has collected nutritional anthropometry data since the 1970s. District-based information has been collected since 1982, enabling the analysis of trends for the nutritional status of children, which serves as a proxy of well-being of the society. Developing these trends have been useful in identifying districts and subdistricts where nutritional status is deteriorating, necessary information for designing and evaluating possible interventions. In addition, data on nutritional anthropometry and household characteristics collected in the 1997 Welfare Monitoring Survey have been used to assess the association between food poverty, malnutrition and other socio-economic factors such as household size, water and sanitation, and level of education.
Qualitative methods are used in area-specific surveys at the subnational and household level. The ALRMP Early Warning Systems collect quantitative data on household food consumption based on a three-month recall. The surveyed population is asked about consumption of main cereals, milk, meat and wild food, and also about any unusual changes of diet in the last three months. Combined analyses are then performed to evidence household food consumption patterns using the dietary information and other collected data such as cereal purchases, income sources, human diseases and relief food distribution. The information is used to separate the food secure from the food insecure for the purpose of food aid interventions.
Improving information on food security.
Kenya is in the start-up phase of the Food Insecurity and Vulnerability Information and Mapping Systems. The national FIVIMS has several objectives. The first is to increase food security issues in the development planning process and to facilitate collaboration on information systems, especially among donor technical agencies and government departments that collect food security-related information, in order to avoid duplication and to improve performance. Other objectives are to improve the performance of the national food security information systems through networking across food, health and nutrition aspects, covering the entire FIVIMS conceptual framework for causal analysis of food insecurity and undernutrition, and to improve the quality of subnational food security data analysis and mapping. It is important to map at subnational levels, as pockets of high prevalence of malnutrition exist where district aggregates indicate low levels of undernutrition. Therefore, with FIVIMS, we hope to map out the country in such a manner that it will be possible to target the most vulnerable areas in the country.
A national food security information coordinating office has been set up within the Office of the President, Arid Lands Resource Management Project, and is currently working within the existing food security information structures. The Kenya Food Security Steering Group and the Kenya Food Security Meeting are the designated focal institutions for FIVIMS. The Kenyan Food Security Steering Group brings together all the stakeholders in food security, the government and international partners that operate systems relevant to FIVIMS. The role of FIVIMS is to strengthen and consolidate the national network. The existing food security structure currently produces a consolidated report that puts together information from all the stakeholders, thus having one single source of information. This was useful during the 1999/2000 drought because information came from one source and was agreeable to the Government, international community and the NGOs. Thus, most active and technically advanced food security information structures in the country were linked to the 1999/2000 drought emergency, which involved information sharing, coordination and decision-making. Lessons learnt from the Arid Lands Early Warning Systems are currently being piloted in the semi-arid areas of the country because the semi-arid, the medium and high potential have large populations and a high prevalence of poverty, yet have no coordinated food security information system operating at the subnational and household level that can provide answers to key questions needed for designing and eventually evaluating possible interventions.
Mr Chairman, this is what we have for the Kenyan perspective.
Shyam S. Dubey
Ministry of Food
New Delhi, India
Food and nutrition security
Ensuring food security for India has been a major task of the government since independence. Over the last few decades, policies and programmes have been designed to ensure availability of food grains to all sections of the society, particularly the weaker sections. As a result of a new agricultural strategy adopted in the 1960s, the production of food grains increased from 50.8 million tonnes in 1950 to about 210 million tonnes in 1999-2000. The total per capita availability of food grains per annum amounts to 210 kg/capita/year, which is more than the normative requirement. The per capita availability of food grains can be a good indicator of self-sufficiency and adequacy at the national level, but the challenge that India faces today is to translate this macro-level availability into micro-level availability by ensuring physical and economic access to adequate and nutritious food at the household level. The root cause of household food insecurity is poverty. Various government programmes have been launched in India to remove poverty and to achieve food security. The Targeted Public Distribution System, through a vast distribution network of around half a million fair price shops, is one of the most effective instruments for household food security. Through this programme, 35 kg of food grain per family per month are being provided at affordable prices, particularly to weaker sections, i.e. families below the poverty line. Further, for around five percent of the population who are hungry, a new scheme called Antyodaya Anna Yojina has been launched, wherein 10 million families identified by the State Government are being provided with 35 kg of food grains per family at very low prices.
Current status of food security and nutrition
Owing to various government programmes and their proper implementation, there is a positive trend in terms of increases in the per capita income and per capita food availability, reduction in death rate, increase in life expectancy, increase in literacy levels and, above all, reduction in the percentage of population liv- ing below the poverty line. As indicated in the District Nutrition Profile, the average intake of nutrients (per consumption unit per day) was 2 308 kcal, which is close to the recommended dietary allowance. The prevalence of underweight among preschool children has decreased from 68.6 percent in 1988-90 to 47 percent in 1998-99. The percentage of stunting in children has also decreased from 65.1 percent in 1990 to 45.5 percent in 1998-99. Chronic energy deficiency, defined by the BMI in adult males and females, in 1975-79 was 55.6 percent for men and 51.8 percent for women, which came down to 28.6 percent and 36.2 percent respectively in 1995-96 (India Nutrition Profile 1998). The infant mortality rate has declined from 146 per 1 000 live births in 1950-51 to 68 per 1 000 live births in 2001. The child mortality rate in children under five has declined from 236 per 1 000 in 1960 to 95 per 1 000 in 1998-99. However, despite tremendous economic growth and spectacular increases in food grains production in the last 50 years, the problem of malnutrition continues to exist, especially among children and women, as well as problems relating to low-birth-weight babies and the prevalence of anaemia.
Poverty as a broad indicator of food deprivation and undernutrition
Poverty and undernutrition are closely related. Undernutrition in simple terms can be defined as a state of low dietary intake. Poverty is a state of deprivation and, in absolute terms, reflects the inability of an individual to satisfy certain basic minimum needs for a sustained, healthy and reasonably productive life. Poverty and lack of purchasing power have been identified as two major factors responsible for low dietary intake in India. The concern over the economic factors resulting in chronic undernutrition led to the use of energy intake as the basis of estimating poverty.
The Planning Commission is the nodal agency in the Government of India for esti- mation of poverty. The brief methodology followed by the Planning Commission in estimating the poverty is as follows:
The Planning Commission estimates the percentage and number of people living below the poverty line (which is also known as the poverty ratio or the head-count ratio) using the Expert Group method for the years for which large sample survey data on consumer expenditure are obtained from the National Sample Survey Organization (NSSO).
The poverty in the country is estimated from the Expert Group method. It uses state-specific poverty lines and state-specific consumption expenditure distributions from the large sample surveys conducted by the NSSO. This method has been used to estimate poverty at the national and state level since March 1997.
Poverty in India is calculated on the basis of consumption expenditure and not income. This is because expenditure data reflect more accurately peoples actual level of living, while income data are more concerned with peoples potential level of living.
The Planning Commission currently uses minimum consumption expenditure anchored in an average food energy adequacy norm of 2 400 kcal (rural) and 2 100 kcal (urban)/capita/day to define the states specific poverty line. These energy norms are based on the recommendation of the Indian Council of Medical Research on per capita daily energy requirements for various categories of people, based on age, sex and activity level.
The national poverty lines are expressed as the monthly per capita consumption expenditure of Rs49.09 in rural areas and Rs56.64 in urban areas at 1973-74 prices, corresponding to a basket of goods and services anchored on a norm of per capita daily energy requirement of 2 400 kcal in rural areas and 2 100 kcal in urban areas. The money values of these poverty lines in 1999-2000 are monthly per capita consumption expenditure of Rs327.56 in rural areas and Rs454.11 in urban areas.
The National Sample Survey Organization also conducts consumer expenditure surveys every year with a reduced sample (one-fifth); however, the Planning Commission uses only the large sample survey data.
As a result of various government policies and programmes at the national level, the incidence of poverty using the Head Count Ratio declined from 44.48 percent in 1983 to 36 percent in 1993-94 and has declined further to 26.1 percent during 1999-2000. In absolute terms, the number of poor declined from about 320 million in 1993-94 to 260 million in 1999-2000.
Household consumption expenditure surveys
The NSSO has been carrying out All-India Household Surveys on Consumer Expenditure with a large sample usually once every five years and also with a reduced sample every year. The latest large survey on household consumer expenditure relates to the period 1999-2000 (55th round). Data on household consumption expenditure through the NSSO are available on a regular basis at national and state levels, separately for rural and urban areas.
Anthropometric indices of nutritional status in children
The nutritional status of children in various surveys is measured in terms of weight-forage (underweight), height-for-age (stunting) and weight-for-height (wasting), and is compared with the nutritional status of an international reference population recommended by WHO. A scientific report from Nutrition Foundation of India has concluded that the WHO standard is generally applicable to Indian children. The findings of the National Family Health Survey - II (1998-99) reveal that the prevalence of underweight children decreased from 52 percent in 1992-93 to 47 percent in the year 1998-99. The survey also reveals that 45.5 percent of children under three years were stunted, while 15.5 percent were wasted.
Food insecurity atlas for rural India
The World Food Programme in coordination with the M.S. Swaninathan Research Foundation has prepared a Food Insecurity Atlas and has mapped the Indian states on the basis of 19 indicators of food insecurity and vulnerability. In addition, all the states have been mapped on a composite index of food insecurity.
Establishment of FIVIMS in India
In order to achieve the goal of food security, accurate and timely information on the frequency, nature and underlying causes of food insecurity and vulnerability is essential. Policy-makers at national and other levels need this information to formulate correct policies and effective implementation of various welfare programmes. The role of FIVIMS therefore assumes great importance as it provides a mechanism for proper linking and networking of various information systems that collect, analyse and regularly update the information related to food insecurity and vulnerability. Considering that food security has multisectoral dimensions, there are several key players in India among departments and ministries, at both the central and state level, whose functions and activities have a direct bearing on efforts to ensure food and nutrition security.
Existing information sources and databases in India
There are varied and rich sources of useful information and databases relating to food insecurity and vulnerability assessment that are regularly updated and revised. Such diverse and rich databases can be used to establish FIVIMS in India, and the major relevant information sources and database are as follows:
CENSUS DATA: The census data are collected once every ten years and are based on an enumeration basis, i.e. house-to-house data. This is the most comprehensive and accurate database that the country has. Certain information like birth and death rates, infant mortality rate, etc. is updated every year.
NATIONAL SAMPLE SURVEY ORGANIZATION: The National Sample Survey was set up in 1950 by the Government of India to collect data on various facets, i.e. household consumption expenditure, unemployment, landholding, health statistics, etc. Household consumer expenditure surveys provide useful information on the number of meals consumed per day and average intake of energy, protein and fat per consumption unit per day, which is updated every year on a smaller sample basis. Information on the perception of households regarding sufficiency of food has also been collected by NSSO since 1983.
AGRICULTURAL STATISTICS: India has a well-established system of collecting agricultural statistics. Statistics of land use have been available since 1884. Agricultural statistics are generated by the Ministry of Agriculture and are updated on a regular basis.
NUTRITION-RELATED DATA: The National Nutrition Monitoring Bureau has units in ten states and reports data in the form of survey reports at the district level for each state. The Indian Council for Medical Research also collects data on health and nutrition. The National Family Health Survey provides useful information on health status and nutrition parameters like child nutrition, nutritional status of women, etc. The last round in 1998-99 collected data from 22 states.
MANAGEMENT INFORMATION SYSTEM OF THE INTEGRATED CHILD DEVELOPMENT SERVICES (ICDS): This system was launched in 1975 to improve the health and nutrition status of children below six years and pregnant and lactating mothers.
DISTRICT-LEVEL DIET AND NUTRITION PROFILE: The Food and Nutrition Board and the Department of Women and Child Development sponsored studies in 1994 to develop district-level diet and nutritional profiles. The District Nutrition Profile of 187 districts from 18 states/Union Territories has been analysed and compiled into a national document entitled India Nutrition Profile. This is a very comprehensive database that extends to the district level, providing information on dietary intake and nutritional status of not only the vulnerable groups but also all age groups and nutritional inadequacies in certain geographical areas and in certain groups of the population. Similar district-wide profiles have been conducted in three more states, i.e. Orissa, West Bengal and Uttar Pradesh.
THE NATIONAL HUMAN DEVELOPMENT REPORT - 2000: The Planning Commission produced this report, which is an attempt to map the state of human development in the country in terms of quality of life on the basis of various indicators of human development. Some of the indicators include: the proportion of the population below the poverty line; the proportion of the population without access to safe drinking water, sanitation, electricity, medical attention at birth, or vaccinations; the proportion of the population living in kutcha houses; the proportion of the population who are illiterate; and the proportion of children not enrolled in school.
Progress in establishment of FIVIMS
In order to coordinate efforts in developing FIVIMS, an interdepartmental committee on FIVIMS was created in 1998 in the Department of Food and Public Distribution, which had representatives from the Planning Commission, Department of Agriculture and Cooperation, Ministry of Rural Areas and Employment, Ministry of Health and Family Welfare, Department of Women and Child Development, Department of Agricultural and Cooperation and Department of Statistics. The National Informatics Center with the largest computer network up to the district level in the country was also associated with the establishment of FIVIMS. The Center has already developed software to map a GIS database up to the district level. A pilot project on the establishment of FIVIMS in two states, i.e. Himachal Pradesh and Orissa, will be initiated with FAOs technical and financial assistance under the Technical Cooperation Programme.
In the proposed pilot project, a FIVIMS unit will be created at the national level in the Department of Food and Public Distribution and at the state level for the establishment of FIVIMS. The FIVIMS unit in the Department of Food and Public Distribution will provide technical advice, receive and analyse data, and coordinate the activities of several ministries or departments for integrating relevant information. A comprehensive database with maps up to the district level using existing information relating to health, malnutrition, food distribution and food production will be generated and constantly updated for the concerned state. Such detailed information, together with mapping of food insecure and vulnerable districts in the state, can then be used for policy formulation. Establishment of FIVIMS in India will facilitate inter- and in-traministerial analysis with a view to promote holistic approaches to the entire question of food and nutritional insecurity. Further, a comprehensive, well-integrated and cohesive information system will be developed for use by policy-makers for the purpose of achieving food and nutrition security.
I will present the national perspectives based on the experience gathered in Colombia and other Latin American countries in the use of different measures for guiding the assessment of need and for targeting programmes and policy responses.
Let me begin with the methodologies that I know best and with which I have some experience, in particular, anthropometry. Figure 1 shows the levels of underweight, stunting and wasting in children under five from four nationwide surveys from the 1970s, 1980s, 1990s and 2000.
Clearly, it is the trend across time periods that draws our attention, as we see a tendency towards the reduction of malnutrition. However, the last survey showed that malnutrition is not evenly distributed, with some areas that have rates that are twice the national average. Anthropometric data that are disaggregated even down to the township level provide valuable information on where the problems are located - we have found that malnutrition prevalence may be as high as 40-60 percent in some areas. Therefore, my comment is that anthropometric surveys are a very useful methodology and can be carried out in countries with the level of development similar to Colombias and even can be carried out in difficult situations like that at present.
Anthropometric surveys are used in Colombia to identify country trends, to single out regions or large districts within the major cities and to identify families and persons with needs. In our experience, anthropometry is used at all levels of the administration, and recently, BMI calculations have been included for pregnant women, adolescents and schoolchildren. The existence of Epi Info software and its supporting networks has made it much simpler to implement this methodology.
In the following sections, I discuss the relevance of the five measures of food deprivation and undernutrition, discussed in the Symposium, to the assessment of food insecurity in Colombia.
FIGURE 1. PREVALENCE OF MALNUTRITION IN CHILDREN <5 YEARS OF AGE, COLOMBIA
FAO methodology for estimating the prevalence of undernourishment
According to statistics derived from the FAO method of estimating undernourishment in Colombia, during the last 50 years, our country has experienced an increase in dietary energy supply (DES) from 1 750 kcal/person/day to 2 580 kcal/person/day. This trend is in good agreement with the reduction of malnutrition presented in the figure. In addition, comparison of the percent of stunted children with the percent of undernourished children estimated by FAO methodology and published in SOFI 2000 also shows a good agreement. Therefore, I can say that the FAO data are consistent with nationally produced data on malnutrition and have the advantage of being available to us free of cost. However, in spite of the great efforts made by FAO to communicate with countries, the DES calculation and, even more so, the percent of malnourished, are not widely known or used. Therefore, I suggest more dissemination of this data, and of course, this should be included as one of the suites of indicators.
Household expenditure data
Colombia has been using this methodology since the late 1970s. It is widely known and used by government officials of the economic and agricultural sector and is also used by the private agricultural sector. However, the interpretation of the results in terms of food security is not straightforward. The use of other information is necessary to explain the results. For example, the percent of expenditure dedicated to food has been decreasing in Colombia for the past 30 years. As mentioned by Dr Smith, a good measure of vulnerability is the percent of a households total expenditures on food. Colombia, therefore, has decreased its food vulnerability. This information adds to the previous data gathered with the other methodologies and gives us a better picture of the food security in the country.
Household expenditure surveys are also used to identify changes in preferences. Between 1984 and 1995, expenditures for meals outside the home have increased 18 percent nationwide, while in the capital, the increase has been only 8 percent. What this means in terms of food security needs to be further explored, perhaps with qualitative methodologies. In comparing food expenditures across three income levels, people spend similarly a substantial amount of money on meat and milk (approximately 20 percent), and the proportion of expenditures for eating outside the home varies little by income class (from 17 percent in the lowest income class to 24 percent in the middle and high income class). These data, however, are somewhat surprising and should be studied further in order to understand their implications fully.
Data from household expenditure surveys are used extensively in Colombia to evaluate food security and are linked to food prices; they are also used for monitoring purchasing power. Perhaps, given the potential for this methodology to assess food security, software or applications based on platforms such as Epi Info could be developed, aimed at boosting the use of household expenditure surveys in smaller segments of Colombia.
Individual food intake survey methods
Colombia has not performed such surveys nationwide since the mid-1980s. In the country, there is a generalized consensus on the need for such survey data not only for food security but also for general health status. However, economic and logistical constraints within the country, and the reluctance and lack of enthusiasm of the funding agencies, have made it impractical to initiate such an effort.
Qualitative measures of food insecurity and hunger
Discussion among academics and interested professionals places great hope in this methodology. The measures are directly tied to peoples needs, and the hope is that the results could be rapidly translated into actions. Given the relative simplicity of the method and the possibility of making changes that alter the significance of the answers, I suggest the development of a core set of questions in standard language and further adaptation to the linguistic and other particularities of the target population. We have recently begun a pilot project on the use of qualitative measures for measuring food insecurity.
In summary, the six keynote papers for this Symposium were very helpful for understanding the different aspects of measuring food security and their strengths and problems. It appears that none of the methods alone could be considered sufficient for the description of food security, and so each country or region should balance the amount of effort devoted to each method. Qualitative measures of food insecurity and hunger should be included in every effort to describe food security. Software for household income and expenditure surveys as well as qualitative measures should be developed in similar ways to that using Epi Info for anthropometry. Individual food intake surveys should be conducted at least every ten years because they provide valuable data and also because they may be used as a gold standard to calibrate the other methods. Coordination with other UN and bilateral agencies to support the initiative and to lobby funding agencies is necessary.
Rita Bhatia and Annalisa Conte
World Food Programme
This study is a collaborative effort between the two divisions of WFP - Policy and Operations. This presentation will focus on the findings of a community and household food security assessment carried out by the WFP Vulnerability Analysis and Mapping Unit in the northern districts of Mali. We show the results of an analysis of primary data collected by the Vulnerability Analysis and Mapping Unit that attempts to identify a relationship between food availability and accessibility with nutrition. This work was based on a number of initial assumptions:
(1) Undernutrition is an outcome of food insecurity.
(2) Undernutrition is likely to be associated to food consumption problems.
(3) Food consumption problems are likely to be due to limited availability and/or accessibility to food.
We carried out an initial situation analysis based on secondary data derived from the Demographic and Health Survey on childhood nutrition, and from this situation analysis, we concluded that the populations living in the northern districts of the country were affected by chronic food insecurity or were highly vulnerable to becoming food insecure. We decided to concentrate our primary data collection effort in this northern area. A random selection of 63 villages that were representative of the population living in this area was the basis of the sample. We developed several working instruments to collect qualitative and quantitative information, including participatory qualitative tools to access food security at the community level and quantitative techniques to access food security at the household level, as well as anthropometric data collection.
Since our objective was to assess food security at both the community and household level, we conducted household sampling for each of the 63 villages. Twelve households were randomly selected in each village, applying criteria based on the initial assumptions. Therefore, we selected four households with stunted children, four households without stunted children, two households headed by a woman and two households with no children less than five years.
TABLE 1. AGGREGATE HEALTH AND NUTRITION FINDINGS FROM MALI
A. Prevalence of undernutrition
B. Vaccination coverage
C. Two week prevalence of childhood diseases
A total of about 3 000 children between six and 59 months of age from 63 villages in the selected households were systematically measured. Analysis of the anthropometric data revealed a high prevalence of malnutrition, and, as Table 1 indicates, wasting was almost 15 percent, while underweight and stunting were both above 30 percent. We assessed the level of vaccination coverage, which was quite low, as well as the prevalence of malaria, diar-rhoea and measles among children during the previous two weeks (see Table 1).
We collected information addressing food security availability and accessibility for all the communities within the surveyed area. From this analysis, we found that 15 percent of the communities were chronically food insecure, in the sense that they had a very low availability and accessibility to food, and 13 percent were vulnerable to food insecurity in the sense that they were better off than the first group but still had a food gap that they were not always able to fill. Forty-three percent of the communities had a food gap that they were able to successfully fill and 29 percent were described as secure, as they did not have difficulty accessing sufficient food. We next looked at the prevalence of malnutrition within each of these categories (Table 2). We found a very high prevalence of undernourished children in the chronically food insecure and vulnerable communities. Both of these groups were confronted on a regular basis or on an irregular and temporary basis with difficulties in accessing food, and in general, underweight and wasting in these communities were higher than the average (as shown in Table 1). Among those communities that had successfully coped with food gaps, the levels of wasting were lowest in children from communities that relied on agriculture plus fishing, compared with those in communities relying mostly on agriculture and livestock. The prevalence of malnutrition was still quite high even among communities that had successful agriculture, surplus production and adequate food intake. We concluded that, given the high prevalence of chronic malnutrition in all community groups, even those with good food availability and accessibility cannot really be considered as food secure because there remains evidence of malnutrition.
TABLE 2. COMMUNITY MALNUTRITION ASSESSMENT BY FOOD SECURITY STATUS PREVALENCE OF UNDERNUTRITION
Community food security status
Chronically food insecure
Successfully coping, agro-fi shing
Successfully coping, agro-pastoral
The results from this study confirm that undernutrition is a multidimensional problem. It also confirms that food alone and undernutrition are not necessarily part of a direct cause and effect relationship. Undernutrition is definitely a manifestation of overall degraded living conditions, and therefore, it is essential to identify the underlying causes of malnutrition and undernutrition that are not as evident as we may believe.
In conclusion, we recommend that in countries where undernutrition is so high as to be considered a public health problem, the use of malnutrition prevalence as the leading indicator or the unique indicator for food-aid targeting at the community and household level should be carefully scrutinized. The identification of the underlying causes of undernutrition is necessary to design and implement activities that effectively address the problem according to the characteristics of the different population groups, for whom food availability and access may be just one of a complex of problems contributing to undernutrition. Therefore, the use of food aid in these cases would be only a small input and would have little impact on addressing the real causes of malnutrition if not integrated into a comprehensive programme including, among others, education, hygiene, access to sanitation and clean water.
World Food Programme
My presentation will cover mainly the indicators that are commonly used in the field and will look at some of the ways in which the United Nations Childrens Fund (UNICEF) uses the information for programming purposes.
What do UNICEF-assisted programmes measure?
We measure the following:
overall nutrition through anthropometry;
specific nutrients, particularly three micro-nutrients deficiencies- vitamin A deficiency disorders, iron deficiency anaemia and iodine deficiency disorders.
individual food deprivation, particularly infant and young child feeding practices, which are very important for child survival, growth and development; and
food intake, particularly of pregnant and lactating women. In recent years, UNICEF has also been concerned with maternal and adolescent nutrition.
What indicators are used to measure overall nutrition?
Underweight, measured by weight-for-age below -2 Z-score, is the most commonly used indicator to assess childrens nutritional status. The information on underweight is readily available because growth monitoring and promotion activities use this indicator.
Wasting is used mainly in emergency situations because the focus is on acute malnutrition and immediate action to ensure childrens survival. In some emergency situations, UNICEF-assisted programmes use the mid-upper arm circumference for rapid screening to determine what immediate attention the child needs.
Although stunting is a very important indicator, it is not very widely used in the field. Stunting is mostly used in situation analyses conducted every five years for the formulation of a new country programme. In those instances, the data from DHS and MICS surveys are utilized. Hence, stunting is not used as an indicator for regular monitoring of pro-gramme implementation.
UNICEF is concerned mainly with process and subclinical indicators of the following three micronutrients.
Vitamin A deficiency disorders
Adequate vitamin A coverage with six-monthly supplementation is the process indicator of choice. Thanks to national immunization days, vitamin A coverage with one round has dramatically increased. The focus now in the field is on identifying strategies to ensure adequate coverage with two capsules per child per year. In women, we look at the prevalence of night blindness during the last pregnancy ending in a live birth within the preceding three years. This is the only clinical indicator that is being used now. In countries where the coverage is high, that is over 90 percent, the recommendation is to use serum retinol, which is done through surveys. Vitamin A deficiency in children under five years is a public health problem in four countries in West and Central Africa where vitamin A deficiency surveys have been conducted: Liberia (58 percent), the Democratic Republic of Congo (61 percent), the Gambia (64 percent), and the Central African Republic (68 percent).
FIGURE 1. ANAEMIA PREVALENCE BY CHILDS AGE
Source: National Micronutrient Surveys (1999)
Iodine deficiency disorders
The process indicator used is the percentage of households consuming adequately iodized salt over 15 parts per million. When the coverage goes above 90 percent of the households consuming iodized salt, urine iodine surveys among schoolchildren are carried out.
Iron deficiency anaemia
The indicator of choice is haemoglobin. The 1990 World Summit for Children goal was to reduce anaemia by 30 percent in women of child-bearing age by the year 2000. There was no goal set for children. However, we now see a high prevalence of anaemia in children, which emphasizes the need to develop strategies for anaemia prevention and treatment in young children. Figure 1 shows iron deficiency anaemia in children under 36 months of age in Liberia and the Central African Republic. Infants of 6-11 months have the highest anaemia rates, probably passed on to them by their anaemic mothers.
Facilitating factors in data collection
UNICEF monitors other child survival-related indicators besides anthropometry and micronutrient deficiency, for instance under-five mortality rates and prevalence of infectious diseases such as diarrhoea, measles, malaria, acute respiratory infections and HIV/AIDS.
The development of low-cost field-appropriate technologies such as HemoCue, dipsticks and dried blood spot have facilitated the implementation of subclinical micronutri-ent surveys, which otherwise would be very difficult to implement in countries with infrastructure and logistical problems.
Breast milk is the perfect food for a babys first six months of life, benefiting children no matter where in the world they live. UNICEF is well known for its support of breastfeed-ing promotion in order to save and enrich the lives of children everywhere.
Indicators used to monitor breastfeeding practices are:
initiation of breastfeeding within one hour after delivery;
frequent, on-demand feeding (including night feeds);
exclusive breastfeeding in the first six months of life - very important for child survival, particularly since the advent of HIV/AIDS;
breastfeeding complemented by locally available and hygienically prepared, appropriate foods;
increased breastfeeding during illness and recovery; and
continued breastfeeding up to two years of age and beyond.
A womans nutritional status has a direct bearing on her chances of surviving pregnancy and of her childs likelihood of leading a healthy life. UNICEF works in many areas of maternal nutrition, but an area of major focus in the coming years is the prevention of low birth weight. Improving womens nutritional status, not just during pregnancy but from adolescence onwards, is key to reducing maternal death and improving child health. This process will reduce not only maternal mortality but also the prevalence of low birth weight, the risk of birth defects, and the rates of stillbirths and infant mortality.
Data on low birth weight rates tend to be scanty and unreliable, and so it is difficult to assess the magnitude of the problem. Low birth weight occurs because of poor maternal and infant health and nutrition as well as a number of other risk factors, such as infection and malnutrition during pregnancy. Cultural practices that discourage women from gaining weight during pregnancy, such as long hours of physical work, also cause low birth weight. All of these factors are compounded when the mother is herself still a child and has not yet completed her own growth.
UNICEF is focusing on a Low Birth Weight Prevention Initiative that is being piloted in 11 countries: Pakistan, India, Bangladesh, Nepal, Indonesia, Philippines, Viet Nam, China, Tanzania, Madagascar and Mozambique. The initiative includes the use of multiple micro-nutrient supplements for pregnant women and will complement UNICEFs Care for Women and Children Initiative, which focuses on womens education, workload, physical health and nutrition status, emotional wellbeing, reproductive health and care during pregnancy and lactation.
Other indicators that we use for maternal nutrition include maternal mortality and the percentage of pregnant women who make at least three antenatal care visits. The number of antenatal care visits is an important indicator for nutrition because it looks at the aspect of care for women. If the family were caring, they would ensure that a pregnant woman made prenatal care consultations and received care from the health services. Weight gain during pregnancy, which is a major determinant of foetal growth, is not readily available because in Sub-Saharan Africa, only about 50 percent of pregnant women make at least one prenatal visit.
There is a lack of clear guidelines for assessing the nutritional status of adolescents. In addition, variations in the rate of maturation and growth make the interpretation of adolescent growth patterns particularly difficult. During the anthropometry discussion group on the first day of this Symposium, we discussed the challenge of nutritional indicators for measuring adolescents nutritional status. What should be recommended to countries that want to assess this? The indicator we have been using is the percentage of adolescents who give birth before the age of 18 years. This is a proxy that does not really say anything about the nutritional status. Besides, after childbirth, it is too late to worry about the adolescents nutritional status, as it is already compromised. This is because the adolescent mother, who has not completed her own growth, has to compete for nutrients with the baby she is carrying. The rate of early child-bearing in West and Central Africa ranges from 25 percent in Ghana to 57 percent in Mauritania. This emphasizes the urgent need to identify effective strategies to address adolescent nutrition.
Where does the information come from?
In the last decade, routine monitoring has provided a good source of data, which however are not always reliable in some countries. DHS have provided very good data nationally and, in recent years, have also begun to provide regional estimates, which are very important for identifying regional disparities and for decentralized planning. UNICEF works very closely with DHS at the field level and finds them very open to suggestions. The UNICEF-assisted MICS were conducted in many countries between 2000 and 2001 for the end-of-decade evaluation of progress on the implementation of the World Summit for Children goals. The MICS methodology has been standardized with that of the DHS to ensure data comparability. A few countries have also carried out national micronutrient surveys, particularly in the West and Central African regions.
How do we use the information?
We use the information particularly for advocacy purposes, e.g. advocacy for policy formulation - development of nutrition policy including individual micronutrient policies, and for ensuring that nutrition is a priority on the national agenda. We use the information for fundraising to obtain resources for the implementation of programmes to reduce undernutrition. We use the information for programme development. Normally, the UNICEF programming cycle is five years, and each new country programme is preceded by a situation analysis of children and women, based on the data available in the country. The situation analysis determines the priorities of the new country programme. Midway into the implementation of the five-year pro-gramme, a mid-term review is conducted to assess progress (using available data) and to decide on which mid-course corrections need to be made in the programme. We use the information for social mobilization to identify partners including NGOs who can reach the largest number of children and women in the field. Finally, we use the information for development of programme communication for behaviour change at household and community levels.
What are some of the Interventions?
UNICEF supports, promotes and protects breastfeeding in the following way:
support for training activities and creation of national structures;
advocacy and support for implementation of the International Code of Marketing of Breast Milk Substitutes;
advocacy for the adoption of the protective maternity legislation;
advocacy and support for the implementation of the Baby Friendly Hospital Initiative to ensure that all hospitals and maternities become centres of free or low-cost breast milk substitutes;
support for HIV-positive mothers in making appropriate infant feeding choices;
support for breastfeeding promotion activities at the community level;
promotion of appropriate complementary feeding.
UNICEF supports growth monitoring promotion activities in the following way:
recommendations for individual childrens care, particularly related to illness and feeding but also to cognitive and motor development;
activity plans for the community that aim to make it easier for families to maintain the growth of their children, for example, by addressing problems of food shortages, poor water conditions or collective child care needs that extend beyond a single household;
programme activities that bolster community actions that affect households with special needs, such as HIV/AIDS and income-generating activities;
Some nutrition-relevant actions that UNICEF supports include:
support for systematic deworming;
promotion and distribution of insecticide-treated mosquito nets for the prevention of malaria, the number one killer of children in Sub-Saharan Africa;
promotion of improved personal hygiene, including hand washing;
provision of potable water and sanitary latrines in schools;
promotion of and support for girls education;
provision of antimalaria prophylaxis for pregnant women;
counselling and education on nutrition;
promotion of family planning - care for women and their nutritional status;
prevention of HIV/AIDS.
Emergencies, from wars to natural disasters, can increase malnutrition rates dramatically. UNICEFs work on nutrition in emergency situations is aimed at protecting the childs right, not only to survival, but also to adequate nutrition and freedom from hunger. We play a major role in nutrition surveillance, in providing essential micronutrients, in organizing measles immunization campaigns, in coordinating the supervision and support of therapeutic feeding centres for severely malnourished children, in assuring adequate infant feeding, in feeding orphans, and taking all other actions possible to prevent severe malnutrition. UNICEF and the World Food Programme work together to strengthen and sustain the capacity of households to meet their basic needs for food, care of children and women, health services, water and sanitation, and prevention of HIV/AIDS.
UNICEF is taking action to ensure that micronutrients are available to women and children who need them by working with governments, in both donor and developing countries, to develop innovative pro-grammes to deliver micronutrients in foods or through health services. The support includes vitamin A supplementation through national immunization days and identifica- tion of sustainable community-based channels to ensure adequate vitamin A coverage with two doses per year.
In many countries, UNICEF supports pro-grammes to ensure that iron and folate supplements are available, particularly to pregnant women, and educational efforts to improve awareness of the problem and possible solutions. Much greater effort, however, is needed to address iron deficiency in women and children. The focus is now on development of community-based strategies for the prevention and control of anaemia in children and in pregnant women.
There is growing experience with vitamin A fortification in developing countries following the long-running programme of sugar fortification in Central America. The success of salt iodization in controlling iodine deficiency disorders has provided an impetus for vitamin A fortification in some countries. Fortification of maize with vitamin A is proving successful in Zimbabwe, and the first sugar fortification experience in Sub-Saharan Africa is moving forward in Zambia. Nigeria has just started food fortification with vitamin A and/or iron.
The prevention and control of iodine deficiency has been a tremendous global success story, largely owing to efforts to universally iodize salt, a practice that had been common in industrialized countries for many years. As a result, some 91 million newborns worldwide are protected yearly from a significant loss in learning ability. Approximately 70 percent of households are using iodized salt. This global success story effectively demonstrates the value of public and private sector partnerships in pursuit of a well-defined human rights goal. Thanks to this unique combination of global advocacy, partnership and alliance building, more than 90 countries are committed to universal salt iodization, and many of them are now experiencing the visible impact of their efforts. The battle to prevent iodine deficiency disorders is not yet over. There is a need to accelerate salt iodization activities to increase the number of households using iodized salt, to sustain the progress that has been made and to ensure consistent and continued quality control and monitoring of products and processes. There is a need for continued political commitment and public will.
In conclusion, the data that are collected by UNICEF support programmes are truly action-oriented to improve the quality of interventions to ensure that childrens rights to good nutrition are fulfilled.
I will close with a quick mention of the UNICEF Medium-Term Strategic Plan (MTSP) for the period 2002-2005. This plan provides an exciting opportunity to readdress the past failures in improving the nutritional situation of women and children. The guiding principles of the MTSP incorporate the life cycle concept that is so essential for achieving successful nutritional impacts. If nutrition is adequately dealt with in all three stages of the life cycle, it will contribute greatly to ensure a good start to life, to ensure that children complete a good-quality basic education, and to ensure that adolescents develop fully their individual capacities.
The five priorities of the MTSP are shown below:
promotion of girls education;
early child care for survival and development;
HIV/AIDS prevention and control;
These strategies emphasize the need to change the way in which we have been doing business. These changes should include the integration of activities to ensure that a package of services can reach children and the strengthening of partnerships within the UN system, donors, and other partners and stakeholders. For instance, the presenter is currently on loan to the World Food Programme in Rome to strengthen the collaboration between UNICEF and WFP. UNICEF is willing to collaborate with other agencies in collecting data together if it meets the information needs for moving forward the agenda of fulfilling childrens rights.
The first message that emerged very clearly from the discussion is that food insecurity is a complex concept both in its manifestation and in its causes, and therefore, it is useless to try to determine which measurement is most applicable for this concept. No single indicator or measure can or should pretend to embrace food insecurity. Indeed, we were reminded that food insecurity has many distinct dimensions. These include food availability, accessibility and consumption, outcomes such as nutritional status, and living conditions including poverty and its various aspects. Information on food insecurity, in order to be useful, has to address these different dimensions. We heard suggestions to consider a composite index, but it was observed that any such index may not reveal all relevant aspects and also may not detract from the cost of assembling the basic information embedded in its construction. We were advised, nevertheless, to try to move in this direction and to make something that is complicated a bit simpler.
We have heard several times that although much attention has been placed on the con- cept of not enough or energy deficiency owing to food deprivation, the other aspects of malnutrition, ranging from micronutrient deficiency and anaemia to overweight and obesity, are public health problems that are part of the food insecurity phenomenon and therefore should be given all due attention in terms of both information and action.
A second theme from the discussion dealt with the different measurement methods and the various types of food insecurity information they generate. Anthropometry was mentioned as a fundamental source of information in all the examples and cases we heard today. The importance of this is that anthropometry can address any type of geographical level and can provide trends through time and provide information on different groups, families or individuals. It was noted, however, that a large amount of anthropometric information is still child-related, while little is available on adult anthropometry - data that would be very useful. We saw that comparing adult and child anthropometry in the same population, although yielding diverging signals, can provide an important understanding of how things happen. Also, it was noted that for nutritional anthropometry of adolescents, there are still methodological problems to resolve.
Another source of information discussed comes from the various types of household surveys such as expenditure surveys and living conditions surveys, as in the Indian example. While anthropometry provides an outcome but does not clarify the causality behind it, household information is vital in helping to understand the causes and mechanisms at stake, and therefore in leading to appropriate action.
The qualitative or, better named, self-assessment surveys that provide the individuals assessment of their hunger situation were discussed. The term self-assessment is used rather than self-reporting because nearly every survey involves self-reporting on some level. This method is very useful because, first of all, it reflects what really matters, namely how the concerned people themselves perceive hunger and how they suffer or not from the situation they are in, providing deeper insights than what may be possible from other methods. Because of the subjectivity and the risk that respondents may attempt to manipulate their responses, this information at first sight could be regarded as soft, but we were shown that, if properly developed and validated, self-assessment can become hard information and yield results of more value than a simple collection of opinions - it can become scientific, verifiable data.
There were diverging views about how all these methods can be combined. We heard that anthropometric measurements could be added in household surveys at low cost, or that qualitative information could be put into traditional questionnaires, but we also heard that including anthropometry into other surveys means a large increase in costs. Despite the lack of consensus on how to combine meth- ods, the usefulness of these complementary sources of information was made very clear.
The mapping of information was also felt to be of great value. We have seen on many occasions that information needs to be known on different levels and that not only numbers but also trends are dramatically important. Trends indeed may be more reliable than numbers, but the latter remain indispensable for many purposes.
A third major theme was that information collection must be linked to and justified by the use made of the information. We were shown in a number of examples what type of information is used for which type of pro-gramme or action, e.g. for determining access to fair shops or to programmes of targeted subsidized prices. We also had concrete examples of this link from UNICEF, where information on food insecurity is used for designing remedial action. This becomes an important concept when discussing the cost (for the surveyor) and the burden (for the surveyed) of information. When information leads to action and policies, it is necessary to obtain the information even if there is a cost involved. In many cases, the benefit of information will outweigh the costs when its collection is justified by a clear link to actions, policies or programmes that rely on valid and reliable information. Another aspect of the use of information is advocacy. Information for this purpose may not be of the same nature as that reviewed so far, but advocacy is a necessary function and therefore must be considered part of the different uses of information.
Much of what we have heard is linked to subnational-level problems, situations, actions and programmes. At the national level, we have inter alia the FAO measurement, or rather the estimate of the number of people undernourished. We have seen that the usefulness of such information is not the same from country to country. We heard that for India, this kind of global assessment is not really of interest because the global availability of food is secure and is no longer a problem. This was the case also for Colombia but different in Kenya and Mali. From this, it was clear that the FAO assessment of the number of undernourished is considered fairly valid in some countries, far from the mark in others and dubious in other cases. When FAO started publishing its estimates of the number of undernourished at the country level, it was very much with the intention of triggering a process of improvement on the estimate, as if to say: Here are national level data that we have estimated using the FAO method with the information we have received from you. It was expected that some countries would find the assessment valid in their case, while others would challenge it and provide improved data from which a more valid estimate could be based.
Another aspect of the debate that appeared clearly in the example of Kenya, but was also underlined in other cases, is that there is much complexity in the information systems and in the types of information generated and used at national and subna-tional levels, as well as in relation to the international community. This complexity may result in excessive demands for information together with excessive costs and burden for the population, while the results arising from the use of the information may not be optimal. It was interesting to note, in the case of Kenya for instance, that in the resource-poor Early Warning Area, very positive results had been achieved in coordinating information among many partners in a purposeful fashion in order to support badly needed effective action. The next step would be to move towards adopting the same approach and reaching the same degree of progress in other regions, those better endowed in natural resources but with a more complicated information situation. This shows that no case is desperate and that improvement is feasible, although the process remains difficult.
In this respect, we also heard several times that international organizations are not con- sidered to be very helpful in solving the complication problem. Your message was very clear, that international organizations should in all possible ways help countries and governments to have a simpler and more effective approach to information. On this account, there were references to the fact that FIVIMS does not pretend to add information or create new demands but instead attempts to help countries make better use of the information they already have, to eliminate duplication or to fill in gaps. Above all, however, FIVIMS attempts to bring together institutions that are generating and using information on food security to help them make better use of the information.
As we have seen, there are several different methods to study food security, all measuring something different. Therefore, it is necessary on some occasions to use two or three methods concomitantly on the same population, to see clearly what kind of information each method provides and how well the various methods might bring a different light on the same situation. One can also evaluate whether a priori hypotheses on convergent or divergent information are borne out, therefore providing information both on the particular situation and on the use of the methods themselves. The Mali case demonstrated how results derived from the use of different methods applied to a particular situation did not necessarily tell the same story. It is indispensable to develop occasions like this purposefully from which one can learn how the various methods work; this also helps those specialized in one method to understand how their approach relates to others.
This issue of the relationship between different methods is linked to a final point raised by our colleagues on several occasions, namely the frequency with which information should be collected and reported. This is, of course, an important issue: as information is usually not needed all that frequently, considerable savings in resources can be made by carrying out surveys at appropriate intervals. For instance, we were given the example of intake surveys that are difficult and costly but provide unique information: they may not need to be done frequently. They can be very useful when conducted at several years intervals to help calibrate and consolidate the information from less precise and less direct observations. A final message was that among the tasks that can be usefully done in the future, a few well-selected cases where different methods are applied in parallel would greatly help clarify what each one brings and how they complement each other, and would help make precious savings in costs in the future while assuring gains in understanding and effectiveness.
Former FIVIMS Coordinator, FAO
I have changed the title of this talk slightly, from The Millennium Development Goals (MDG) Perspective to A Perspective on MDGs. This is partly because I have no formal authority to speak for the MDG process, as I am not really part of that process. However, I am reasonably informed about what the process is. To put our discussion today into context, I will provide some information on the MDGs and then give my views on the relevance of the MDGs for the work of this Symposium.
The Millennium Development Summit held in New York in September 2000 was said to have been the largest-ever gathering of world leaders. Those leaders combined a set of previous goals and targets into those of the Millennium Declaration. In the final stages, this involved principally the UN system and the World Bank. The active role of the Organization for Economic Cooperation and Development (OECD) in promoting the previous International Development Goal process that started in 1995 contributed greatly to the definition of the MDGs. The Declaration contained eight general goals, 18 subsidiary numerical targets and 48 quantifiable indicators to assess progress - the part of the MDGs that has the most to do with this Symposium. To implement the MDG process, there are three elements, one of which is the Millennium Project that I will speak about in limited detail.
The eight MDGs are very broad, multidisciplinary and multisectoraled, and are as follows:
(1) eradicate extreme poverty and hunger;
(2) achieve universal primary education;
(3) promote gender equality and empower women;
(4) reduce child mortality;
(5) improve maternal health;
(6) combat HIV/AIDS, malaria and other diseases;
(7) ensure environmental sustainability; and
(8) develop a global partnership for development.
The most relevant for us today is the first goal, to eradicate extreme poverty and hun- ger. The target associated with the first half of this goal, the eradication of extreme poverty, is to halve between 1990 and 2015 the proportion of people whose income is less than US$1 a day. This is a fairly straightforward target that has a numerical dimension to it and includes three indicators. The first indicator is the percent of population living on less than US$1 a day, measured in purchasing power parity terms. The second indicator for extreme poverty is the poverty gap ratio, which is the incidence multiplied by the depth of poverty. The third is the share of the poorest one-fifth of the population in national consumption. The second target is clearly most relevant to us: to halve between 1990 and 2015 the proportion of people who suffer from hunger. The first indicator of the second target is the prevalence of underweight in children below five years of age. The second indicator is the proportion of people below the minimum level of dietary energy consumption, essentially the FAO indicator.
How will the Millennium Project work? This project is to be a research programme designed to provide the intellectual and analytical foundation for the MDGs. This effort has a budget of over US$10 million over the next several years, and the process is to be directed by Professor Jeffrey Sacks, of Colombia Universitys Earth Institute. He has also been appointed Deputy Secretary General as part of this process, in order to monitor the achievement of these development goals.
Goals (2) to (8) are, in essence, a synthesis of targets and goals from different international meetings and summits from the previous ten years or so. This was a particularly important period because it represented an attempt to link the goals with numerical targets and quantifiable indicators.
The MDG global indicator database is being kept by the UN Statistics Division in the Department of Economic and Social Affairs of the UN Secretariat in New York. The strategy for the Millennium Project was defined by the 18 separate targets. The initial proposal was to have 15 task forces, each with two coaches and 15-18 expert members. Each group would commission papers, and in this way, the project would address the analytical function associated with monitoring of the MDGs. All of this is subject to ongoing negotiations, and apparently a decision is yet to be made on how many task forces will be formed and exactly what the subject matter will be, as well as the relationship between the numbers of subsidiary committees composed of members from the UN agencies and other groups. Reports on the progress of meeting the MDGs will begin with the 2003 UNDP Human Development Report and essentially will provide a preliminary statement on the MDG indicators.
At least two of the undernutrition indicators being discussed in this Symposium are used as MDG indicators: anthropometry and the proportion of people undernourished. The specific language for the latter indicator as used in the Millennium Declaration is the proportion of the population below the minimum level of dietary energy consumption. This is in contrast with the World Food Summit target of halving the number of undernourished people by 2015. I think this difference in language may be subject to some future negotiation. The reporting of these two indicators will arise from existing databases. Obviously, the undernutrition data will come from the FAO Statistics Division, and the underweight data will come from the WHO database in collaboration, in some areas, with UNICEF through their Multiple Indicator Cluster Survey estimates. At the country level, presumably the reporting process will allow for countries and the UN team working with the national government to go beyond the use of these two indicators to better characterize the problem of hunger.
There are several key questions about the MDGs that I would like to address today. The first is really more an observation rather than a question. We have an opportunity for a new unified monitoring and evaluation focus within international organizations. For the first time, there is a list of 48 indicators that has been agreed to by virtually all international agencies, including the World Bank (WB) and OECD. Therefore, we have a common framework for monitoring progress over time to the extent that these indicators are good indicators that reflect changes over time. I think the more important question goes beyond simply producing tables of 48 rows and a column for every country to explain the performance over time. If we do observe changes and improvements in income, poverty and hunger, what type of analysis will be part of the process? What are the relationships among the 48 indicators? They cover a wide range of subject matters: for the SOFI reports, we have made several attempts to analyse these international indicators. Experience shows that associations among indicators often can be demonstrated, but causal relationships, e.g. poverty causing hunger, are difficult to show in a convincing way.
Some of the critical explanatory variables that can show why countries are making progress or not are not in the MDGs database, but could be added fairly easily in the future. Examples of these explanatory variables are the presence of conflict or civil war, the amount of official development assistance and the amount of foreign investment and remittances, which are very important for certain countries.
The final question is: How will country case studies be used as illustrations to compare performance across countries? We have seen in our work that it is much easier to illustrate the process of change with a good case study than to generalize across very diverse countries.
The MDGs are a very important step in the development of international analysis and monitoring of progress towards the satisfaction of basic human needs. This initiative presents a number of opportunities for FIVIMS and for those of us concerned with the development of collaborative information system activities, particularly at the country level. The first is the opportunity to make a contribution to the MDG Hunger Task Force and to the broader performance analysis that will take place. The second will be linking the global database to better-quality national-level data and national MDG reports. This will help to maintain an appropriate focus on food security and its role in the poverty reduction process. I think that this structure, as well as having a hunger task force within the MDG process, will allow us the opportunity to put food security back on the radar screens. The final opportunity is that of being able to address the continuing need for better subnational data to inform the UNs Common Country Assessment process at the country level, the Poverty Reduction Strategy process, as well as using the MDG monitoring database for targeting appropriate interventions. To better address these needs, we need more reliable and complete subnational data, and it is here that the FIVIMS initiative can make its greatest contribution.
Washington, DC, USA
The World Bank (WB) has made the achievement of the Millennium Development Goals the focus of its activities by designating them as corporate priorities. The WB is responsible for reporting on the indicators related to hunger, poverty (living on less than US$1 a day, the poverty gap and income distribution) and environmental sustainability. For this Symposium, I will concentrate on the target of eradicating extreme hunger, which is measured through the indicators of underweight or malnutrition in children under five and dietary energy consumption. The WB tracks these indicators internationally and encourages individual countries to track child malnutrition levels as part of the reporting for the Poverty Reduction Strategy Papers. The WB supports countries that track and report on the child malnutrition goals.
Why does the WB track malnutrition? The first reason is the relationship of malnutrition with the multidimensional aspects of both income poverty and non-income poverty such as female literacy, social exclusion and the poor health environment that contributes to malnutrition. From the WBs point of view, tracking trends in nutritional status is useful in monitoring the overall effort to reduce poverty. Second, child malnutrition reflects a number of intermediary processes such as household access to food, access to health services and caring practices. It is, therefore, a very powerful indicator of social well-being and reflects on a number of other developmental goals. The impact of nutrition on poverty is well known; a well-nourished child is more likely to stay in school and become educated, thereby improving lifetime earnings because of increased mental and physical capacity. In the long run, a well-nourished population can enhance the economic development of a country. It has been shown that the higher the levels of malnutrition or clinical indicators such as anaemia, the lower the overall productivity. For example, anaemia has an impact on cognitivability and therefore on productivity. Therefore, good nutrition in children is crucial for achieving the goal of universal primary education as well as gender equity in education.
Female education is very important for nutrition, because improper feeding practices as well as improper caring practices are among the principal causes of childhood malnutrition. Improving female education in general improves child-care practice. Not only do women with a secondary education have fewer children, but their children have lower levels of stunting. Malnutrition, including mi-cronutrient deficiencies, is a principle cause of childhood illness and contributes to a poorer health status into adulthood. According to the WHO, for example, nearly 60 percent of the deaths in children under five are associated with malnutrition. Women with anaemia have a greater risk of dying during childbirth, owing to a lower tolerance to blood loss or to infection. Improving the nutritional status of women will help achieve not only the goal of reducing maternal mortality but also that of improving the status of women, which is the fifth of the eight MDGs. Improving womens health is especially important in developing countries where a disproportionate number of women are affected by malnourishment and health risks, which in turn contribute to malnutrition and illness in their children. It is imperative to interrupt this vicious cycle, illustrated in Figure 1, by intervention during the early years of life.
Because of the close relationship between poverty and malnutrition, it follows that the poor are most affected by malnutrition, as illustrated in Figure 2. The previous speaker, David Wilcock, indicated that if you work with poverty and if you focus on poverty, you eradicate malnutrition. At the WB, we found that at both national and household levels, improved incomes do have a steady but slow impact on malnutrition: as average incomes increase, extreme poverty declines and children become better nourished. We have found, however, that it is quicker to close the nutrition gap than to close the income gap. Focusing on nutrition also has a fall-out effect in terms of affecting the other determinants that cause poverty and malnutrition. In this instance, we are trying to take advantage of the synergy by using the approach that yields the greatest impact.
FIGURE 1. RELATION OF NUTRITION TO MATERNAL HEALTH
Source: Administrative Committee on Coordination/Standing Committee on Nutrition (ACC/SCN). 2002.
Fourth report of the world nutrition situation. Geneva, United Nations.
FIGURE 2. CHILD MALNUTRITION: POOR PEOPLE ARE THE MOST AFFECTED
Source: Demographic and Health Survey data
The FAO indicator of undernourishment tracks changes in food supply over time across countries and determines the percentage of the population who do not meet minimum energy requirements. This type of data is useful to have and, when it is possible to identify the causes for variation in food supplies over time, contributes to effective interventions. Knowing only that changes in food supply are taking place but not knowing where and why this occurs does not add very much and, from the WB point of view, is not regarded as a policy indicator.
In conclusion, the WB focuses on nutrition outcomes in the work that it carries out in countries and in its own work programmes.
We will continue to use anthropometric indicators of children and adults to track malnutrition and will continue to identify its underlying determinants. Our programmes will focus on improving breastfeeding and caring practices, and strategies will be undertaken to increase micronutrient intake, to improve womens status with expected benefits on food distribution within the household and to improve the health environment in which the people live.
Department for International Development
The UK Department for International Development (DFID) welcomes this Symposium and the opportunity to present our perspective on the measurement and assessment of food security. DFID strongly believes that good statistics play a crucial role in the reduction of poverty, both to inform policy and to monitor progress. We are very keen to work with developing countries to build the capacity for producing accurate statistics that are relevant to their needs as well as those internationally, and also to help improve their capacity for using the statistics, an area that is often neglected even in the developed world. Like the WB, our work at DFID is largely guided by the Millennium Development Goals. Therefore, we need good data for the MDG indicators to monitor progress towards the goals, for advocacy and political purposes and to inform the allocation of our resources.
Turning now to the particular issue of measuring food insecurity, hunger and nutrition, I agree that we are not looking for a single indicator. We need to look at a range of different indicators in different situations and to use them in a complementary way. A crucial point that DFID would like to highlight is that the indicators we need depend on the questions we are trying to answer - we need to be sure that the statistics that are produced are relevant to the issues we are concerned about and to the users actual needs. The choice of indicators also depends on the level for which we need the information, whether it is global, regional, country or within the country. We need to use indicators in combination for a variety of different reasons. We might want to use two indicators to validate one another or to combine information for improving the estimates. For example, we might want to look at different stages in the process such as food availability, food access and the associated outcomes. I would like to support the approach, suggested earlier in the Symposium, to move towards the concept of a suite of indicators. As we saw in the presentation by David Wilcock, the idea of a suite of indicators is reflected to some extent in the MDGs. This means that beyond simply reporting one or more food-related indicators for a country or region, we want to encourage analysis of what the different trends show us by compar- ing or combining several indicators. We also want to encourage the process of examining the relationships of the nutritional indicators to those of the other MDG targets such as education, gender equality and maternal health: this was mentioned by the speaker, S. Patel, from the WB.
One focus of this Symposium has been the FAO undernourishment indicator, which is used as one of the MDG indicators. I think it deserves a few comments from a DFID perspective. One of our main principles is that indicators should be fit for a purpose. We need to be clear about what they are being used for. We can spend many years and much money improving indicators, but we need to look at the costs and the benefits when we do this and ensure that the indicators are fit for the intended purpose. If we want to measure food availability at the global level, we have only the FAO indicator of undernourishment. But having said that, DFID does see some significant weaknesses in the method, and many of these have been discussed over the last couple of days. I will highlight a few points. A major concern is that the distribution information, that is the coefficient of variation, is not updated over time. Therefore, as has been said several times before, the indicator is simply measuring the supply information in the food balance sheets. The implications of this from a policy perspective are that we appear to be saying that the only way to deal with hunger is through an increase in production, and I do not think that everyone working in this field would necessarily feel comfortable with that approach. In Dr Naikens paper, he did put forward the proposition that the results are not that sensitive to changes in the coefficient of variation as long as certain assumptions hold. What I think is not quite clear is the extent to which those assumptions do hold and what effects updating the coefficient of variation would actually have. Indeed, in the work presented by Lisa Smith and her colleague in the contributed papers session, there was evidence put forward suggesting that it did make a considerable dif- ference when information on the distribution was updated. So my query, when looking at this, is can we do anything about this, and if so, can we do it at a reasonable cost?
I think the evidence over the last two days of the Symposium has suggested that the information from household expenditure surveys, which already exists in many cases, can be used at least to check the assumptions that have been made and possibly can be incorporated into the FAO calculations for the improvement of the method. Dr Naikens paper also suggests that household expenditure survey data could be reconciled with the food balance sheet data to try and improve these estimates as well. Using household expenditure survey data from the countries would also have another benefit: it would link the indicator more closely with countries own statistical outputs. I get the impression that the FAO indicator is often seen as being calculated here in Rome with little contact with outside countries. More consultation with the countries to which the figures relate and more use of country-owned and country-produced data would strengthen FAOs dialogue and relationship with those countries.
There has been some question over the reliability of using the household surveys for distributional analysis, and we should not ignore this. I do not think it rules out using this information to improve the FAO method - there are going to be weaknesses in whichever approach we take - but we do need to be aware of the issues. Instead, I think it provides an opportunity for working in collaboration with other analysts who use household surveys for income distributions and poverty indicators towards improving this aspect of the survey methodology. We could also work with international organizations who put together the guidelines on how these surveys should be carried out.
What I have said so far is very much at the global level. Perhaps an area of greater interest is helping developing countries build their capacity to measure and monitor hunger at the national and subnational levels, because this is the type of information that is most useful to inform their policies and to make a real difference in reducing poverty and hunger. At this level, we need much more detailed information on who is hungry, where and why. DFID would also advocate doing more work on looking at vulnerability and who is at risk of being hungry in the future. Many of the methods put forward over the last few days can shed light on these aspects, and, as with the global indicators, we need to use them together in a complementary way.
As I said at the beginning, we need to be clear about what questions we are trying to answer. This is going to depend on the particular country and situation that is being looked at. How can we ensure that we produce statistics that are relevant to policy-makers and other users? And how, in an environment of scarce resources, can we develop a coordinated approach to make the best use of the money we have to spend on monitoring and measurement?
At this stage, I would like to echo the words of the dinner speaker, Peter Matlon, who identified some principles for how we could make progress in setting up these processes. Some of the things that he stressed were that we need a user focus and a country focus. We need to ensure that the demand is there within the country for this information because that is the only way to obtain a sustainable system to continue producing the information in the future. We also need to bring together the various interests within the country and outside - the policy-makers working within different government ministries, the various agencies and also other users. We would like to see a process that allows countries to work with donors, agencies and others to assess the needs for information on food security, hunger and nutrition, and to put in place a plan that meets a wide range of these needs in a coordinated fashion. Existing data collections could be used in new ways, and surveys at the planning stage could be adapted to make them more friendly to food security analysis. As we heard yesterday, an- thropometric measures can be added to some of these surveys without adding too much to the cost, and qualitative modules might be added as well. I think the principle is that we need to work in a joint way to make the best use of our resources.
The national FIVIMS initiative has the potential to incorporate many of these ideas. We are particularly pleased that a review of FIVIMS work has been initiated and should lead to a strategic planning process. We hope that the plan will incorporate some of these ideas. Another important consideration is that national FIVIMS processes or other similar mechanisms be integrated fully with the Poverty Reduction Strategy processes currently under way in many countries. To a certain extent, this vision we have for national FIVIMS mirrors the work that DFID is trying to do on a wider statistical scale with regard to poverty information systems and indicators. We hope to work with developing countries, other donors and other agencies to develop a coherent plan for information systems agreed upon by both policy-makers, users and data producers. This should make it easier for donors to work together and coordinate funding.
We feel it is very important that the development of a countrys statistical infrastructure is country-owned and country-driven. There are very often monitoring subgroups within countries as part of the Poverty Reduction Strategy processes that look at the information needs for these strategies. We are keen to try and use this structure as our entry point to encourage a collaborative and integrated approach. Our vision would be that the data from these country-owned systems would then feed the international systems, so that we have a coherent flow of data feeding through to the global monitoring of the MDGs.
There are plenty of opportunities for better cooperation and coordination between all of us to try to meet our data needs. DFID is looking forward to working with this agenda and hopes that the FIVIMS initiative, FAO and other agencies will be involved as well.
Altrena G. Mukuria
ORC Macro International
Calverton, MD, USA
At ORC Macro International, we conduct the Demographic and Health Surveys that are funded by the United States Agency for International Development (USAID). USAIDs support for national population-based fertility surveys began in the early 1970s with the World Fertility Survey, which was designed to collect information on fertility levels and mortality. During that survey period, there were very few data on nutrition; the only aspect of nutrition studied was breastfeeding as an influence on fertility. In the mid-1970s and early 1980s, there was a set of surveys called the Contraceptive Prevalence Surveys that included questions on the current status of breastfeeding and duration - again looking at the influence of breastfeeding on fertility levels and contraceptive use. Next came the series of national population-based surveys, called the Demographic and Health Surveys, initiated in the mid-1980s. The focus was expanded to include more on child health, and anthropometry measurements were added in order to report on the nutritional status of children. In addition, the questionnaire was expanded to include questions on breastfeed- ing practices and timing of the introduction of complementary food into infant diets. At this time, there was a growing concern about nutrition. The current set of surveys, called the Measure DHS+, started in 1998 and will be completed in 2004. In the current round, there is a more complete focus on nutrition and anthropometric measurements not only of children but of women as well. We also have an expanded 24-hour dietary recall of food given to children under three years and a seven-day food frequency questionnaire.
The DHS is a population-based, nationally representative household survey that uses a two-staged, cluster sampling design. The sample used for the DHS is aimed to provide reliable estimates of important indicators at the national level, for rural and urban area comparisons and for subnational, regional or provincial areas, but not the district level. Since 1985, with funding from USAID, the World Bank, UNICEF and others, approximately 150 surveys have been conducted in over 60 countries around the world with over 40 countries having had more than one survey. In addition to the main survey, which will be discussed in more detail later, 74 surveys have been conducted on men, and 20 specialized surveys have been conducted on other topics.
The DHS survey is conducted as a series of modules that a country can include or not, as they choose. There are two main core questionnaires. The first is the household questionnaire, which lists the usual members and visitors in the household and provides background information on those members of the household together with their age, sex, education and relationship to the head of the household. Background characteristics of the household dwelling are collected, including water availability and sanitation facilities, the number of rooms and the building materials. The second core questionnaire that all countries must use is the womens questionnaire for collecting information on women between the ages of 15 and 49. Background information is collected on their child-bearing experience, marriage and sexual activity in addition to nutrition, family planning and health practices. The nutrition content of the Measure DHS+ surveys includes anthropom-etry, infant feeding and micronutrients.
With regard to anthropometric measures of women, we look at both BMI and height. An analysis of recent data shows that that there is a growing problem of overnutri-tion and overweight in countries around the world. For example, in a recent survey in Egypt, we found the mean BMI for women to be 29, already indicating overweight, with over 40 percent of all Egyptian women having a BMI of 30 or more, indicating obesity. We also collect anthropometry on children that allows us to calculate underweight, wasting and stunting but also overweight.
The surveys also provide background information on practices that influence nutritional status. An area of concern is infant feeding and breastfeeding practices. We collect data on all children in the household under six years of age whose mother is present. Information on the timing and initiation of breastfeeding and the use of colostrum is collected. We ask whether the child was given any prelacteal feeds. Exclusive breastfeeding is derived from the 24-hour recall, and information on frequency of breastfeeding in the previous 24 hours, the duration of breastfeeding and the continuation of breastfeeding as well as the use of bottles is also collected. Complementary feeding practices are also explored. We collect data using a 24-hour dietary recall on the last born child less than three years of age for each mother in the household. We also have a seven-day food frequency module that is somewhat expanded from previous surveys in which we ask about the frequency of solids and information on a number of meals a child has eaten in the previous 24 hours.
One exciting new area for us is micronutri-ents. We collect data on iodine and vitamin A supplementation of mothers and children, we ask mothers of their experience with night blindness in pregnancy, we test household salt for iodization, and in one country, we collected serum retinol levels. Hopefully, we will be able to do so in more countries in the future. Iron is another area that we collect data on. We carry out anaemia testing of women and children, and we collect information on iron supplementation in pregnancy. A few countries have collected information on iron supplementation of children, but that is not standard.
Through DHS data, we provide nutrition information for problem identification and advocacy. We provide data on the prevalence of nutritional status and micronutrient status as well as the determinants and correlates of nutritional status. Our data have been used for nutrition and health policy development and also indicator development. In programme planning, the DHS data are used to establish coverage levels for food supplementation, immunization and a variety of other measures such as vitamin A and iron supplementation. They are also used to establish targets globally, internationally, regionally, nationally and subnationally. For instance, with the integrated management of childhood illness programme, our data have been used to monitor the progress of a country towards achieving those goals. The MDGs utilize our data on undernutrition, and UNICEF has integrated our data into their database for evaluating their mid-term strategies. Data are also used for monitoring the prevalence of undernutrition within a country and for determining how effective nutrition policies or programmes are in reducing the prevalence of malnutrition or micronutrient deficiencies, as well as for determining coverage and success of iron supplementation programmes for pregnant women. The data also supply information on changes in behaviours and practices, for example breastfeeding practices, complementary feeding and child care practices.
Our users are many and include the World Bank, WHO, UNICEF, USAID, FAO, national governments and programme managers. Non-governmental organizations (NGOs), private voluntary organizations and donor agencies utilize our data in their planning and programme development. Now that pro-grammes are being held more accountable by their governments for reaching targets, an emphasis is on numbers, which are used for programme evaluation. Since DHS data are in the public domain, they are used by researchers and technical experts around the globe. We have a collaborative relationship with the experts who use our data to decide upon the best indicators to include in the surveys. For example, the need was expressed for the collection of serum retinol levels for vitamin A deficiency in addition to questions on night blindness in pregnancy, so we have included these indicators in our survey. As we produce information on the extent of a particular problem, questions may be raised by scientists or policy-makers, leading to requests for additional information. Their needs are fed back into the survey, and the survey adopts new questions or protocols accordingly. We will be completing this round of surveys in the next two years, but if we receive continued funding for the next round, we will update the questionnaire to include new indicators or other types of information on nutrition, demographics and health, as indicated by the users.
Washington, DC, USA
USAID uses the indicators discussed in this Symposium for four levels of its work: (1) international policy; (2) agency policy-making; (3) strategic performance measurement; and (4) food and nutrition field programming.
International policy level
The United States government sent a high-level delegation to the World Food Summit: five years later, and there it reconfirmed its commitment to the World Food Summit target to reduce the number of hungry by half by 2015. We have also endorsed the MDG to reduce the proportion of hungry by half by 2015. These international policies imply heavy reliance on two indicators discussed in this meeting for measuring food deprivation and undernutrition: the FAO food balance measure and the anthropometric measurement of underweight of under-five-year-olds. The FAO measure, in spite of its weaknesses, helped to put the question of hunger on the screen of the leadership and policy-makers within USAID.
Measures of food insecurity are of particular relevance to the USAID Food Aid and Food Security policy, which was adopted in 1995. This policy focuses on household nutrition and agriculture productivity as the two critical sectors of food security. The policy incorporates a definition of food security with three components: availability of food, access to food and proper utilization of food by the body. For estimating food availability, we use the FAO measure, for measuring access we have found household expenditure surveys to be relevant, and for food utilization we often use anthropometry. In fact, improved child anthropometic measures only occur after households have been able to satisfy all three aspects of food security.
Three emerging USAID policies were prefigured in the briefing papers prepared for the World Food Summit: agriculture productivity, famine prevention and nutrition. Agricultural sector growth is the principal measure under the agricultural policy. Under the famine prevention concept, reduction of child wasting is considered to be central to famine mitigation, while the reduction of high levels of chronic malnutrition is seen to be central to prevention. In the nutrition concept paper, reductions in the incidence of child underweight, low birth weight of children and mi-cronutrient deficiencies are central objectives. I do not know the actual targets and goals that may be established, but we have begun to think about them.
Monitoring strategy performance
The use of indicators discussed in this Symposium took a quantum jump at USAID, following the passage of the 1996 Government Information Performance and Results Act by the United States Congress. The law placed all United States Government agencies on a results-management track. Every agency, and the operating units within it, was required to develop strategic plans with strategic objectives, and associated measurable indicators and targets were designed to monitor the performance of their strategies. For example, the USAID health strategy has used underweight of under-five-year-olds as an indicator related to reduced infant mortality. In monitoring the humanitarian assistance goal, USAID has recently begun to measure its progress in emergency response, using wasting and acute malnutrition of under five year olds as well as crude mortality rate. For our economic growth strategies, we make use of household expenditure survey data. Country missions also have strategic objectives and associated indicators in their country development programmes that must be monitored. Anthropometry is among the many health indicators used by missions in countries with high levels of food insecurity and undernutrition. The DHS surveys play a major role in providing the data for monitoring the performance of some of these mission strategies.
In food insecure regions, field programmes involve the types of measures discussed at this Symposium. Perhaps one of the most operationally significant uses of these measure is for deciding how food aid is to be targeted. USAID development food aid amounts to approximately US$500 million per year in resources targeted to Low Income Food Deficit Countries, as determined by FAO using food balance sheets. NGOs implement 80 of these development food aid programmes in 25-30 countries. Anthropometry is further used to identify the areas that are most food insecure and to target the food aid within the recipient countries.
About one-half of these food programmes are maternal, child health and nutrition pro-grammes in which NGOs are required to use anthropometry to set baselines and to monitor progress over the five-year duration of the programmes. Some have begun to experiment with frequency questionnaires such as those presented in the contributed papers session by Patrick Webb from Tufts University and Edward Frongillo from Cornell University. NGOs are also working with FANTA (USAID-sponsored Food and Nutrition Technical Assistance) and Eileen Kennedy to try to develop qualitative hunger measurements. This would reduce the burdensome and expensive efforts to measure anthropometry in every food security project.
USAID also has a wide set of child survival, micronutrient and infant feeding programmes that use many of the indicators from the DHS survey presented today by Altrena Murkuria. USAID emergency programmes often rely on wasting measurements to assess the seriousness of disasters and to measure the impact of food and dollar emergency assistance in particular emergency situations.
In summary, although the FAO measure has very limited operational significance for us, we have adopted it implicitly at the international policy level because USAID has accepted the goal of the World Food Summit. We should retain it despite its shortcomings. Certainly, anthropometry has become a foundation for many of our field programmes and our policy performance monitoring. Household expenditure surveys are increasingly important to much of our work, while food frequency measures are used in a limited way in NGO field programmes. Qualitative surveys of hunger are an emerging area that may hold some promise for low-cost food security monitoring.
The first main result of the discussion this afternoon is how important the MDGs have become. After they were presented by Mr Wilcock, we heard from international organizations as well as donor countries that they are all using the MDGs as their guiding thread for objectives and targets at all levels. We can conclude that the purpose of the MDGs seems to be very well fulfilled, as they now serve as the benchmark for international development assistance, and among their top objectives, we see poverty and hunger eradication. Another observation about MDGs is that they clearly encompass all the dimensions that have been so consistently affirmed as essential for comprehensive understanding of food insecurity, ranging from income poverty to education and health in its different dimensions, and of course hunger and malnutrition. We also noted that they helped to trigger momentum in assembling and enriching the information.
Let me pause briefly on a comment that while reducing hunger is part of both the MDGs and the WFS goals, the expression of these two targets leads to very different objectives. In reality, there is a considerable quantitative difference between the two - between proportions and numbers. The targets of the MDGs and the WFS differ by hundred of millions of people, or by decades, and therefore it is appropriate that these differences be discussed.
As I move from the question of national perspective to international perspective, the topic of this afternoon, let me comment en passant about the issue of international versus national demands for data. Data are not collected for the purpose of feeding a worldwide database or for monitoring a worldwide target. Rather, information collection is always directed at solving particular problems in a country. However, there is a need for comparability, standardization and harmonization of the data - we heard the need for this expressed repeatedly. This need does create a conflict at the country level between what has been done traditionally and what countries are now being asked to generate and produce. However, rather than countries being asked to generate data they do not need, it is quite often just a question of which particular method they should use to collect the data. This is not unreasonable, because if we collect information in different ways on the same problem in different places, within one country or among several countries, in the end we may not be able to use the information in the most efficient way.
The second main observation is that even though this was a session on international perspective, very quickly we were called back to recognize that what is important is not the worldwide objectives, but rather what happens at the country level. While on the international level information is useful for advocacy, monitoring or resource allocation, the country perspective is most important; it is here that real action can be achieved. From this observation, two conclusions emerge: information has to be country-demanded, and capacity building at the country level for generating and utilizing information is crucial. I would like to underline what was said by the panellists about capacity building, because quite often this critical link is badly missing. It is not easy to create the necessary capacity to generate and utilize information, but it is decisive in changing the demand for, and hence the quality of, generated information - not so much more information but more useful information - to use for improving policy and action.
This discussion has brought us back again to the relevance of the national FIVIMS - the effort at rationalization and optimal utilization of information systems. One new point I noted was related to the Poverty Reduction Strategy Papers, a very important channel for linking information to action at the national level that should be part of this streamlining effort to improve assistance to policy-makers.
Anthropometry was once again highlighted in several occasions as basic information, and more broadly, nutrition-related information was considered to be extremely important. Indeed, nutrition is important not just as one dimension of poverty but in itself, because improvements on the nutritional front have a direct beneficial impact not only on the welfare of the people but also on economic ability at micro- and macro levels, as hunger and poverty mutually reinforce each other.
Vulnerability as a distinct dimension may not be able to be measured with the type of methods we are using now. I understand that much work on vulnerability is still required and that it is a work in progress.
The necessity to justify information indicators by their use for policy decision and preparation for action was reiterated, and this led to our Chairmans suggestion of this morning to think in terms of a matrix. The matrix column heads would show who does what, the row heads would list the level at which action is taken, and the information needs would be specified at the row-column intersections. I find it interesting in a meeting called a scientific symposium, where we would expect to hear mostly about technicalities of nutrition, socio-economy and surveys, that the strongest conclusions concerned the importance of ascertaining that what we are doing with information will be useful in an action-oriented perspective.
We were reminded of the importance of proper terminology and not to use words indiscriminately like hunger, undernourishment or undernutrition, food energy deficiency and so on. While this is well understood, it is also true that as one moves from the technical field to that of politicians and public opinion, it is difficult to adhere to strict terminology and, even more, to avoid using public terminology. The case of hunger is very clear. We all know that when the word hunger is used, it generally has a very loose, unscientific meaning, but everybody understands it, and therefore it tends to invade a broader field than what might be proper. This is not easy to avoid and thus is a problem.
There were a few more words about the FAO method at the end of the discussion. We heard comments that more country-based information should be used in this method and that household surveys should be injected into it as well. This, in fact, is the case: the method is based on data provided by countries, and many surveys are reviewed and used when possible. But I must warn that country data and household surveys have very serious problems of gaps, credibility and consistency. Another person asked why the coefficient of variation (CV) should be kept constant through time. Personally, I fully agree with this comment, and the only reason I can see for keeping the CV constant is that in the past, very seldom were usable values of the CV available in a country, so when you had one, you did not even dream it could change over time! The situation is different now, and while I am not an authority on this issue, I do know the situation reasonably well, and I do not see that changing the established approach should create any difficulty other than using a countrys coefficient when it becomes available and changing the value in the formula for the relevant period when it is updated.
I believe, Mr Chairman, that these are the main points in the discussion. Let me add simply that we also heard many colleagues express that they found this meeting quite a useful event, and we all thank you for that.