Food and nutrition security:
why food production matters

FOOD SECURITY AND NUTRITION IN THE LAST 50 YEARS

The years 1945 to 1952 saw Europe struggling to restore pre-war food production, consumption and security. Even in the mid-1960s, former levels still had not been reattained in Asia - rightly seen as the highest-risk area, where chronic undernourishment left people extremely vulnerable. Between 23 million and 30 million lives were lost in the Chinese famine of 1960-1962. In 1965 and 1966, famine was barely avoided in South Asia. More than 75 percent of Asians (and probably 90 percent of undernourishment victims) depended on food production for their income.

Then, in much of the world came rapid growth, green revolutions, land reform and poverty reduction. The proportion of chronically underfed persons fell in developing countries from 36 percent in 1970 to 20 percent in 1990. The proportion of underweight children under five years of age fell globally from 42 percent in 1975 to about 32 percent in the late 1990s. The reduction in undernutrition was rapid in East Asia, substantial in South Asia and Latin America but very slight in Africa, and it recently reversed in the republics of the former USSR. In these last two regions, related death rates are rising. Moreover, globally, progress in combating poverty and undernutrition was slower between 1987 and 2000 than between 1970 and 1985, corresponding to slowdowns in the yield growth of food staples, in land redistribution and, hence, in rural employment. Despite past progress, during the 1990s one in five people in developing countries ate less than the caloric minima for metabolic, work and other functions. Worldwide, there are still more than 150 million children under five who are underweight; more than 200 million - more than one in four - are stunted. These conditions appear to be implicated in about half of the 12 million deaths annually of children under five and, for some of the more damaged survivors, in physical and even mental retardation.

Notwithstanding the slowdown in the progress against poverty and undernutrition, governments were increasingly given credit for such a large fall in food insecurity. At the national level, overt colonialism retreated massively twice, between 1947 and 1965 and in the early 1990s, leaving the state formally accountable to its own nationals. In both periods, many countries (not only former colonies) moved towards democracy. Moreover, national populations were becoming more organized in effective "civil societies" with more literacy, information, communication and power to put pressure on their governments for adequate food access. At the international level, food security was advanced by parallel changes in institutions and public awareness. Institutionally, the process began when the 1943 Hot Springs Conference (in part responding to President Roosevelt's 1941 designation of "freedom from want" as a human right) wrote the blueprint for FAO. For the rest of the century, international agencies obtained money and support, from taxpayers of almost all nations, to underpin major components of food and nutrition security.

At the 1943 Hot Springs Conference, which set the blueprint for FAO, countries accepted responsibility for ensuring food and nutrition security.

IFAD, with a mandate to emphasize nutrition and food production for the poorest, was the harbinger of a growing focus on poverty as the main cause of food insecurity. In 1973, the World Bank began to express its increasing concern about the impact of poverty; but in the 1980s the focus shifted to country lending strategies. These initially emphasized stabilization but, increasingly in the 1990s, poverty impact also. The current World Bank President, James Wolfensohn, asks that the Bank be judged by its impact on poverty. Despite small (but significant) lending for nutrition as such, the Bank's approach implies that poverty reduction, together with increased agricultural output, is the main route to food and nutrition security. Yet the share of Bank lending (and of total aid) going to agriculture has fallen since the early 1980s, as have yield growth in food staples and the speed of poverty reduction and improvement in household food security.

A series of international conferences, most notably the World Food Conference of 1974, clarified the issues but led to sarcasm about "resolutions without resolution". Yet, in 1996, the Copenhagen World Summit for Social Development (the Social Summit) and the Rome World Food Summit culminated in targets to halve world poverty and undernutrition by 2015, and the Organisation for Economic Co-operation and Development (OECD) has supported moves to make this target country-specific, monitored and backed by aid.

The targets for reducing poverty and food insecurity can be met. In East Asia the number of food-insecure fell faster between 1970 and 1985, owing in part to fertility transitions and bringing sharp falls in child-adult ratios as well as a rising ratio of workers to dependants (see the section Demographic transitions and food security, p. 203). Even larger shifts are now under way in Africa and South Asia. Can these heartlands of poverty and food insecurity expel them through this demographic "window of opportunity"?

Increased food production is necessary but not sufficient to improve food security - entitlements must also grow.

Past and current trends and issues: implications for the early 2000s

Postwar experience suggests that, for rapid falls in undernutrition to continue in the low-income African and Asian countries where it remains worst, yield growth in main food staples will need to recover to the 3 percent level achieved in the 1970s - in the 1990s, as in 1950-65, it was barely 1 percent per year in developing countries. Growth in yields will also need to spread to some of the neglected staple crops and poorly watered lands, especially in Africa, that were bypassed by the green revolution. However, in real terms, funding for public agricultural research has been stagnant in international centres since the mid-1980s and has fallen in Africa and Latin America. Threatened water scarcities and the overfarming of marginal lands, although sometimes seen as reasons for not aiming at higher yields of food staples, are partly due to inadequate funds for research.

In the 1940s and 1950s, the founders of FAO saw such needs clearly. So, in the 1960s, did the creators of the green revolution. Such production-oriented approaches are not sufficient to improve food security and health environments, but they are necessary.

Since 1945, the following global trends affecting food security and nutrition have emerged:

• increasing accountability on the part of nation states and the international community, not just for famine deaths, but for food and nutrition security;
• from about 1980, a shift from state action to reliance on markets;
• a shift from concern about national food security, measured by staples self-sufficiency, to potential household food security, measured by dietary energy supply (DES), and then to actual household food security, measured by the absence of protein-energy malnutrition (PEM);
• a shift towards consideration of food entitlements, rather than food availability, as the main guarantor of food security;
• major, although patchy, improvements in household food security and measures against undernourishment.

Consequently, attention is increasingly focused on the food-health-nutrient-activity balance and safety, involving micronutrients as well as calories; health (including bioabsorption and biodiversion), work and child care as well as food intake; and overnutrition as well as undernutrition.

Linked to the above, staples production for local employment and consumption has been downgraded since the early 1980s. This is also due to expanding European and Asian staples output and yields; falling real prices of staples; environmental concerns; failure to appreciate that many low-income people require extra staples production to obtain employment-based entitlements to food staples; and reliance on comparative advantages and trade options to permit a "switch" to cutting undernutrition mainly by non-food employment. Growth in such employment opportunities has seldom been substantial or nationwide until late in the development stage, even in East and Southeast Asia.

Staples yield growth, which was 3 percent per year in the developing world in the 1970s, fell to just above 1 percent in the 1990s. The need to raise yields, and thereby staples-related employment and entitlements, in poor areas will increase as workforces grow by 2 to 2.5 percent per year in much of Asia and Africa up to 2025.

The yield potential for tropical and subtropical staples surged for maize in the 1950s, and for rice and wheat in the 1960s, but has subsequently slowed down. In most countries with widespread undernutrition, contrary to general belief, rapid progress in food staples production implies renewal of growth in yield potential. The shift from states and international public action towards markets must be modified in at least one key area, biotechnology, if research is to be reoriented to the still pressing requirements for reduced undernutrition.

Three themes underly the trends described above:

i) The end of the dichotomy between "productionist" and "distributivist" approaches to individual food security - between failures of "food availability" and "food entitlements" as causes of famine (and chronic hunger). For most of the undernourished, extra employment income from local staples production has been the key to enhanced food entitlements in the period from 1950 to 2000, as discussed below. This will remain so until 2025, given the continued rapid growth of workforces and the need to restrain and stabilize local staples prices. Such employment is generated as a result of higher yields of food staples and better access to land, credit and institutions. The slowdown in food staples yield growth since the 1970s, and in its employment impact, offers new challenges. The experience of the green revolution shows how to address them in ways that enhance food entitlements and household food security. So far, however, the promising new tools of biotechnology have not been applied for the purpose of raising staples yields for poor smallholders.

ii) Increased policy focus on nutrition security (see the section From energy adequacy to nutrition security, p. 231) in response to new knowledge and new problems and in recognition of the fact that nutrition encompasses aspects such as nutritional quality and balance, food safety and physical activities that prevent obesity.

iii) The need for people and organizations wishing to improve food and nutrition security to respond correctly to changes in the role of the state vis-à-vis markets. This theme includes the joint action of these groups towards, and possible biases against, weaker or less politically active groups at special nutritional risk (e.g. rural or remote groups, children, females, minorities and refugees). The role of states and markets is crucial in determining whether such groups gain income, access to food and information to permit nutrition security. Many opponents of "free" or globalized markets fear that, as a result of perverse interactions between states and monopoly-like corporations, such markets do not help, and may harm, some of these vulnerable groups in their progress towards nutrition security. Even though competition can make some poor groups poorer still, there is much evidence to suggest that countries that liberalize their markets are generally those that are most likely to raise incomes, and this tends to reduce poverty faster.

FOOD SECURITY AS ENERGY ADEQUACY

For most of the world's poor, "food security" is knowing that future meals will provide enough dietary energy to meet their requirements. Box 18 reviews the concepts of food adequacy and security that are used in the text.

THE FRAMING CONDITIONS: PEOPLE, FOOD AND ENTITLEMENTS

Demographic transitions and food security

The past 50 years have seen an astonishing demographic transformation. Population growth accelerated between 1940 and 1960 in Asia and Latin America, and ten years later in Africa. The really massive improvement was among children under five, which raised the ratio of children to adults. Some 10 to 20 years later, the age groups that had been saved from child death increasingly entered the workforce, and parents, now more confident that their children would survive, began to reduce their fertility rates. As population growth slowed down, the ratio of adults - savers and workers - to children began to climb steeply.

Just as the early phase of demographic transition, with its rapid growth in child numbers, had harmed economic growth, income distribution and, hence, the reduction of poverty and PEM, so the later stage - with children reaching working age, and fertility falling - helped all these factors. The effects on food security are considerable. Directly, smaller families are less likely to be poor and, given their poverty level, less likely to suffer PEM. Indirectly, about one third of East Asia's growth in real income per person from 1965 to 1992 can be accounted for by rising adult-child ratios; across more than 50 developing and transitional countries for which adequate poverty data are available, the poverty-reducing impact of lower fertility rates is about as strong via better income distribution as via faster economic growth.¹

The "demographic gift" of an increased worker-dependant ratio may help sub-Saharan Africa and South Asia to reduce food insecurity, as occurred earlier in East Asia.

The heartlands of world poverty, South Asia and sub-Saharan Africa, have entered their fertility transition and will face a very rapid rise in their ratios of workers and savers to dependants in the next two decades. For example, in Kenya, on the UN "medium variant" projection, the ratio of prime-age adults - the main workers and savers - to children under 15 years of age will rise from only 1:24 in 2000 to 1:87 in 2020. Will this "demographic gift" sharply cut poverty and, hence, PEM in Africa and South Asia - as it did in East Asia? This depends on whether extra workplaces offering rising incomes become attractive for the growing workforce and their employers - as was the case in East Asia - and whether extra investment opportunities, with a high impact on growth and poverty, attract the new savers. In East Asia this was achieved through a first phase of growth in staples yields and employment as a result of the green revolution between the mid-1960s and late 1980s; those countries that successfully continued the struggle to reduce poverty and malnutrition then moved into a second phase of rising non-farm employment, both urban and rural.

Sharp rises in adult-child ratios and thus in worker-dependant ratios and in savings, which were experienced in South Asia and Africa, are a "window of opportunity" for these regions to slash remaining poverty and PEM, as was done in East Asia. It has been proved that farm intensification through methods suited to local conditions can achieve similar results on a sustainable basis in large parts of South Asia and Africa. But the window is now curtained by threats to the growth in yields of, and employment-based entitlements to, main food staples and, in some cases, by severe land inequality.

Rural poverty in Honduras Poverty does not completely explain undernourishment - FAO/20729/A. Proto

Poverty and entitlements to food

Hunger, not just famine, mainly afflicts people who have insufficient "food entitlements".² As Sen emphasized, food production still matters: the poor rely on it for most of their food entitlements, whether based on small-scale farming or employment.

However, one would expect postwar improvement in food security to be associated with a substantial retreat of poverty (used here in the limited sense of "low real private consumption per equivalent adult"). Broadly speaking, this is correct. Poverty declined (and nourishment and DES improved) little in sub-Saharan Africa, in Latin America mainly in 1965-78, and in Asia mainly in 1975-90.³ Recent resurgences of poverty and undernourishment in the transitional countries are clearly linked.⁴

Yet, many poor households do not suffer undernourishment, owing to "positive deviance" in child care, low energy requirements or other behavioural adaptations. Many non-poor households, for opposite reasons, do suffer undernourishment. And, although the poorest show much higher responsiveness of calorie intake to extra income than others, the short-term responsiveness of the average household, even in quite poor communities, is often rather slight; poverty does not completely explain current household caloric inadequacy or PEM.

Nevertheless, poverty - and its main proximate cause, inadequacy of income from paid or self-employment and, hence, inadequacy of food entitlements - substantially "explains" long-term group undernourishment risk. Among and within countries, differences in the decline in poverty help to predict the improvements in both caloric underfeeding and anthropometric undernutrition. For several reasons, the relationship between poverty reduction and increased DES is imperfect:

• Anthropometric undernourishment measurements - stunting in children of normal weight for height, PEM at ages 12 to 24 months and micronutrient deficiency - may be a signal for some very poor victims of undernourishment that they should have other, perhaps higher, priorities for using extra income than simply acquiring more "empty" calories.
• A hard-worked poor person whose income rises because of a windfall or promotion to less heavy labour may well use and consume fewer calories. This person is more likely to use income increases to buy extra dietary energy when extra work is required, or in peak seasons.
• Estimates of calorie-income elasticity are usually based on the average sample member of the community. The poorest show much higher responsiveness.
• Different extents and speeds of response of household food security and DES to poverty reduction are also caused by variations in: health risk and care; work requirements; child care; intrahousehold gender discrimination; relative prices of foods and other needs; seasonality; and genetic factors affecting metabolism, weight, height and, therefore, food requirements.

However, these and other qualifications do not negate the strong causation between poverty and DES shortfall, food insecurity and PEM.

Self-sufficiency in staples is not an indication of national food security.

ENERGY ADEQUACY AND INADEQUACY: LEVELS AND TRENDS

Self-sufficiency in food staples and national food security

National food security (see Box 19) may be sought by increasing national staples self-sufficiency (NSSS), capacity to pay for staples imports, or stocks. A country's NSSS trends do not reveal trends in food security, although they help our understanding of its causes:

• For five of the eight staples of key interest to the food-insecure, trade affects a small and fluctuating proportion of total output; for at-risk countries where these are the main staples, trade dependence (the main indicator of NSSS trends) tells us little about food security.
• A country's degree of NSSS depends on global markets for wheat, rice and maize (which overwhelmingly dominate the staples trade), not just on the country's food security and policy towards it. Between 1948-50 and 1995-97, Western Europe's farm subsidies helped to cut its share in the world import volume of wheat, rice and maize from 62.6 to 16.1 percent; its export share rose from 5 to 17.9 percent. With such huge net supply rises, countries outside Western Europe are almost bound to move away from staples
  self-sufficiency.
• While policies aimed at NSSS, such as irrigation expansion and support for agricultural research in South Asia, have often been well conceived, in other cases the quest for NSSS readily turns into a policy of extracting rural produce cheaply to feed the cities, thereby creating perverse incentives that discourage farmers from growing more food, especially in Africa.⁵
• Above all, a greater degree of NSSS may mean better or worse HFS and PEM levels (Box 19).

A second indicator of NFS is the ratio of food imports to total exports. Given sufficient political will and functioning market and transport systems, countries with low ratios can respond adequately to poor harvests or sharp rises in food import prices. In 1988-90, 11 countries in Africa (including Ethiopia, Egypt and Mozambique) had food imports-total exports ratios of above 55 percent - levels that were seen elsewhere in the developing world only in Haiti, Samoa and Yemen. Such countries tend to rely heavily on food aid. Table 10 tracks an even more relevant ratio, that of imports minus exports of food staples to non-staples exports, for some large countries.

This indicator gives a rather better guide to NFS than self-sufficiency. It does suggest the precarious food security situation in China in the early 1960s, in India when landless labourers' employment income was slashed by the successive disastrous monsoons of 1965 and 1966, and in the Sudan during episodes of civil war. However, the ratio must be used with care. Net staples imports can fall, not only because domestic supply is rising, but because people are becoming poorer and less able to afford sufficient staples. A rise in non-staples exports, as with petroleum in Nigeria, the former USSR and Indonesia, may do surprisingly little to improve staples availability via imports. If such export income accrues largely to owners of capital and/or governments who receive royalties but have other priorities than food security, little employment income is created for the poor and the income will not be used to buy food staples.

The ratio of stocks in public hands to normal consumption is also an indication of a country's NFS. In years of dearth, states can undertake large releases of staples. This helps the poor by suppressing the rise in prices. It also encourages the timely release of traders' hoards when prices start to rise; public capacity to act in this way in Bangladesh in 1984, but not in 1974, probably made the difference between famine and non-famine.⁶

Underfeeding: potential and actual DES

For the First and Second World Food Surveys,⁷ FAO had measured potential underfeeding by setting national and regional mean daily "dietary energy supplies" (DES) per person against average requirements. Since the Third and, especially, the Fourth World Food Surveys, estimates of distribution of supplies were also taken into account. Supply estimates often depend on questionable output data, and requirements are controversial. However, sharp trends and major turning points in DES are usually significant, relating strongly to PEM outcomes and, hence, to HFS. Table 11, p. 212, provides summary DES data for a longer period.

Before 1939, "in areas containing over half the world's population, food supplies at the retail level [provided] less than 2 250 calories per caput daily.... Average total calorie supplies were around 2 000 or less in many large countries...".⁸ Europe recovered rapidly from wartime shortages but the proportion of the world's people in countries with DES below 2 200 kcal per day rose from about 40 percent just before the Second World War to about 60 percent in the late 1940s. "Over most of the Far East, where nearly one half of the world's population is concentrated, the declines [in DES] were about 10 percent." Average DES soon after the Second World War was 24 percent below requirements in India, 21 percent in French North Africa and 18 percent in Mexico. ⁹ Work effort, and to some extent basal metabolic rate and body size, "adapted" to such low intakes or requirements, fell but such adaptation was often harmful, resulting in widespread mortality, illness and mental and physical underperformance.

Table 11 suggests that DES in China, India and Kenya had, even by 1976-78, no more than recovered to the inadequate 1934-38 level, although DES in Latin America had improved sharply. The set of three-year averages confirms that 1976-78 was a turning point. India, China and some other Asian countries shifted from no change to rapid improvements in DES. In Central and East Africa, DES shifted from stagnation to steady decline. Interestingly, DES in West Africa showed no trend between 1961-63 (2 090 kcal) and 1982-84 (1 990 kcal), but then rose steadily to 2 400 kcal in 1995-97.

DES per caput in Asia and West Africa is now about 20 percent higher than in the mid-1970s, and in Latin America (from a much higher base) about 7 percent higher. In the same period, in East and Central Africa, DES appears to have fallen slightly by about 2 to 3 percent from already low levels. However, these DES estimates show that the national potential to reduce underfeeding has risen fast and far for the vast majority of inhabitants of at-risk countries - dramatically so in East Asia and Latin America. The fact that potential improvements in underfeeding (increases in average daily DES based on food balance sheets) do translate into actual improvements in the proportions of underfed people can be shown in two ways:

i) A few, long series of disaggregated, fairly reliable survey data measure year-round daily calorie intake directly per "consumer unit". In India the average intake rose from 2 061 in 1967 to 2 283 in 1989, but it rose faster among farm labourers (the poorest group) and children under five (those at greatest risk from a low DES).¹⁰ Income distribution in India hardly changed over the period; clearly the poor were spending higher proportions of income on extra calories, thus translating a higher mean DES into
better HFS.

ii) Box 20 reports regional FAO results, which estimate how trends in potential undernutrition (measured in terms of average daily DES) between 1969-71 and 1990-92 translated into changing actual undernutrition: the numbers (proportions) of individuals with a basal metabolic rate (BMR) below 1.54 fell from 920 million (35 percent) to 840 million (20 percent) in 1990-92. The improvement was most dramatic in East and Southeast Asia (from 41 to 16 percent). It was substantial in other developing regions, except for sub-Saharan Africa where there was an increase from 38 to 43 percent, although improvements in the West African mean DES suggest that this may have been partly remedied since 1992 (Table 11, p. 212). Proportionate energy shortfalls for the underfed also fell, except in Africa. Thus, the regional pattern of increases in average DES or potential underfeeding is close to that of falls in actual underfeeding - proportion of populations with daily calorie intakes below 1.54 BMR - as estimated in the Sixth World Food Survey. These patterns are also close to those of trends in PEM. Both are fastest in East and Southeast Asia; moderate in the Near East and North Africa, in Latin America in the 1970s and in South Asia in the 1980s; and nil in sub-Saharan Africa. There have been huge improvements for some populations: in China DES rose by more than 60 percent from 1961-63 to 1994-96, and in the Republic of Korea at an even faster rate, so that obesity is more of a threat than PEM. However, although changes in PEM and DES are roughly consistent, continental levels show an HFS discrepancy. The results (see Table 11) suggest much worse underfeeding in sub-Saharan Africa than in South Asia, yet the proportions of stunted children - which are relatively reliable figures - are much lower (Box 20).

Higher food availability translates into reductions in undernourishment.

The recognized weakness of national food production estimates for most of Africa may suggest that estimates of mean DES in Africa are much too low - and, hence, the estimates of underfeeding, and implied lack of HFS, are much too high. But there is no evidence of a continentwide bias in that direction. Another explanation of the HFS discrepancy, which is borne out by mortality, famine prevalence and other evidence, is that smallness of body size, except at extremes, is less harmful in South Asia than in Africa. If this is so, South Asia's much more widespread smallness gives a less reliable guide to IFS than does Africa's much more widespread DES shortfall. In any case, countries and regions tend to show similar trends in stunting and DES (Box 20).

The huge improvement in DES, and also in PEM, is not a cause globally for easy optimism about food security. More than 800 million people are still seriously underfed in DES terms. Africa has not seen a secular improvement; and in the transitional economies, real DES may be declining. Above all, despite favourable signs in the heartlands of DES shortfall and of PEM (e.g. the spread of the fertility transition to Africa and the acceleration of growth in South Asia) the production and employment conditions for continuing rises in food entitlements may be fading, together with growth of employment and yields in food staples.

Household food security against PEM

The incidence of child PEM, measured as weight for height (wasting), height for age (stunting) or weight for age (underweight) is still shockingly high, even though it is falling. About 36 percent of children under five in the developing world - 197 million - were stunted in 1995, 29 percent were underweight and 9 percent were wasted, a condition that usually leads to hospitalization in rich countries. PEM is associated with the deaths of some 6 million children each year, and with persistence into adulthood of incomplete mental development, reduced physical work capacity and damaged immune response for many millions more.

Food insecurity as a result of natural disasters A malnourished child is dwarfed by a ruined cereal crop in the Navrongo district of the Gambia, where two thirds of the crops were destroyed by drought - FAO/11231/Y. MULLER

In terms of measurement, it is less problematic to assess PEM trends than levels. Table 12 shows a slow but consistent improvement in PEM indicators and, hence, in individual food adequacy outside sub-Saharan Africa. The pattern is confirmed by national surveys of children under five for stunting for 1980-1995¹¹ and low weight for age for 1976-1995.¹²

If we could compare all the 50-year periods in human history, 1950-2000 would almost certainly win first prize for speed, scale and spread of nutritional improvement. The falls in underfeeding came first in Europe; indeed, underfeeding (and probably PEM) rose in much of the developing world in 1945-60. The fall in PEM was most dramatic in East Asia, but it was also large in Latin America and South Asia. Only sub-Saharan Africa saw no improvement in DES and no fall in PEM. In 1985-2000, the decline in PEM slowed globally and reversed in transitional countries.

SECURITY AND INSECURITY: SPATIAL AND TEMPORAL VARIATIONS

Seasons, years and famines

Apart from inadequate "normal" energy supplies, vulnerable people are faced with the risk of sharp downturns - poor seasons and poor years. Improved average DES reduces food insecurity linked to poor seasons and poor years. So do the greater integration of food markets and public action to reduce nutritional and food price instability. Therefore, famines have been much rarer since 1945; since 1963 they have shifted almost wholly from Asia to Africa, and since the 1970s they have been largely confined to cases of war, civil war or the breakdown of a state. One reason for the retreat and relocation of famines has been the spread of political pluralism and openness; with rapid global communication technologies, including television, it is almost impossible for famine (as opposed to chronic, even fatal, undernourishment) to persist in an open and democratic society.¹³ The risks of large famines have also been reduced by better early warning and response systems.

Famines causing more than 500 000 deaths in Asia have been virtually absent - in sharp contrast to the historical record - since the disaster in China in 1959-61. The worst subsequent Asian experience, in Bangladesh in 1974-75, cost fewer than 500 000 lives. Most deaths have been in sub-Saharan Africa, but only in Ethiopia (1984-85) did they approach 1 million.¹⁴ Globally, there have been big swings in deaths resulting from famine or drought; periods such as 1982-86 saw "clusters" of war or civil breakdown in several drought-affected countries (Table 13). Seriously affecting 800 million people and causing 6 million deaths per year among children under five alone, since the mid-1960s chronic PEM has damaged far more people than have famines.

Seasonal undernutrition has enduring consequences for physical and economic health.

The incidence of severe seasonal worsening of nutrition has also almost certainly lessened, but there is much more evidence of the great harm that it does. In areas with fluctuations, people adapt to reduce stress, both behaviourally (storage, borrowing) and biologically (tending to deposit and release fat rather than lean).¹⁵ However, some bad seasons are worse than normal, or follow previous bad harvests or economic downturns. And poor people, especially if they suffer low average DES and hence body size, face greater problems from seasonal stress and greater difficulties in coping with it. Human action makes a big difference: around 1994, some 370 million rural Asians and "only" 33 million Africans lived in areas that put them at nutritional risk from the effects of severe agricultural seasonality. However, the fact that irrigation and food entitlements are more widespread in Asia probably meant that more Africans than Asians were harmed by seasonal nutrition stress.

Seasonality especially harms food adequacy for three groups. Stillbirth rates increase if the hungry season occurs in the second trimester of pregnancy, and there is a higher mortality rate if it occurs at ages 6 to 12 months (as passive immunity is lost before it is fully replaced by active immunity).¹⁶ Furthermore, children born in the hungry season face a sharply raised risk of age-specific death in adulthood. In the Gambia, the survival chances of more than 3 000 babies, born in different seasons, have been documented for more than 50 years, revealing "very marked ... raised mortality among adults born during the hungry season.... Infectious deaths [and their secondary effects] were the largest single [cause, suggesting] that early life events have caused permanent damage to the immune system".¹⁷

GROUP DISTRIBUTION AND FOOD SECURITY

Food and nutrition security is harmed for the rural, the remote, females, refugees and displaced persons, as well as for ethnic minorities and children,¹⁸ by reinforcing disadvantages in three ways:

• Disadvantages overlap in food-insecure regions: disadvantages existing between gender and among ethnic groups, etc., overlap with those among illiterates, those in rural and remote areas, etc.¹⁹
• Disadvantages cumulate to reduce prospects of eating the right amount of the right foods, whether the main risk is undernutrition or overnutrition.
• Disadvantages harm many things at once, for example food consumption, access to production-based food entitlements (work, land and other assets) and mobility towards better prospects for production, consumption and, hence, nutrition.

Each of the following is normally more prevalent with than without each of the others: remoteness, rurality, discrimination against females, inadequate schooling, bad sanitation, absent or distant health care, poverty, heavy and seasonally peaked work even in pregnancy, and low calorie consumption per person.²⁰ In most developing countries, those in the "worst" quintile for all of these characteristics are likely to have 50 to 65 percent of the national average life expectancy and years of education, and double the age-specific death rate and periods of illness.

Rurality

Average energy consumption is usually somewhat higher in rural areas, but not among the poorest. In rural India, the poorest decile received only 1 212 kcal/person/day in 1972-73 (despite using 82 percent of consumption for food), as against a still shocking 1 316 (79 percent) in urban areas. Especially for employee households, work tends to be harder and infection more frequent and there are more pregnancies per woman in rural areas, so more food is needed. Confirming this, at the same level of income per equivalent adult, an increase leads to a larger rise in calorie consumption in rural than in urban areas. A greater food shortfall relative to requirements means that in rural areas stunting and wasting are typically at least 1.5 times more common than in urban areas.²¹

Since about 1980, owing to a reclassification of rural areas and to migration, the proportions of urban people suffering from PEM (and usually absolute urban numbers affected) increased. However, the already larger rural incidence of undernourishment usually diverged further from the urban incidence. The ratio of rural to urban incidence of stunting and wasting - as of poverty and early death - tended to rise, except in Africa, where the disparities had been greatest initially. The 60 percent of the world's population who live in Asia have seen big falls in undernourishment, child mortality and poverty risk, but mostly alongside some rise in rural-urban and regional divergence within nations.²²

An aspect of the "reinforcing disadvantage" is that, even if both rural women and rural men fall behind their urban counterparts by the same proportion in nutrition indicators, rural women suffer more. Compared with urban women, their risk from low food consumption is worsened by less chance of trained attendance at childbirth and by a greater lag in education. To compound the damage in Asia and Africa, where undernourishment is worst, there is a larger proportion of rural women than rural men.

Region

In China, much of the rest of Asia and Latin America, remote, marginal, poorly watered or mountainous regions are the most liable to caloric undernutrition and to the other disadvantages that have an impact on death rates and on the development of those who survive: undersupply of health care, schools and roads, as well as a high proportion of residents suffering language or other discrimination, for example ethnic minorities. In Brazil, poverty explains only part of the much worse nutrition status in the north and northeast of the country than elsewhere; much of the rest is probably due to less access to health services.²³

Assets

In some places (rural South Africa, northeastern Brazil) income is so unequal that even rapid growth has little impact on PEM or poverty. This is usually due in substantial part to extreme inequality of assets, especially land and education. Infant and child mortality, strongly linked to PEM, are often higher among landless labourers than among small farmers. Distributive land reforms have been linked with sharp falls in undernutrition and poverty.

There is ample evidence that households with an educated adult - especially a woman - have a higher standard of living and, even at a given level, have less PEM. Educated women marry later and have lower marital fertility. Their households are better nourished owing to more knowledge of food and farming and to higher worker-dependant ratios and less sibling competition.²⁴ Less widely recognized is the implication that, where education is especially unequally distributed (by region, gender or income group), PEM is higher than elsewhere - particularly among the most vulnerable (small children) owing to high fertility in uneducated households.

Gender

What is the role of gender in reducing PEM and/or overnutrition? In some countries, tests showed that stunting or low weight was much more prevalent than predicted from average DES. This is partly because of adaptation, but also partly because the gender distribution of DES and health care (and hence treatment of infections, which alters the efficiency of DES use) is especially unequal in some countries. This contributes to the unfavourable stunting and wasting status of India - North India, like Bangladesh and Pakistan, shows strong food discrimination against girls aged two to five years, with known harmful outcomes.²⁵ The same is probably true of Mauritania. The nutritional ill-effects of gender discrimination may rise even where nutrition is generally improving; in India, the incidence of adult body mass index (BMI) below 16 (third degree chronic energy deficiency) was 11.4 percent of men in 1975-79 but 8.8 percent in 1988-90, while for women the incidence only fell from 12.7 percent to 11.3 percent.²⁶

Girls aged two to four years suffer serious disadvantages compared with boys in DES (relative to need) and/or access to health care - and as a result worse PEM - in many parts of Asia, but not in sub-Saharan Africa or Latin America.²⁷ Furthermore, girls almost everywhere have fewer chances of education than boys, so that women have less access to skilled work. Women also suffer discrimination in pay even for the same job or task, and more severely in access to land, legacies and credit. Although the evidence (especially in Asia) does not suggest a much greater poverty risk for women, their control over income is certainly less.

Apart from directly damaging the food security of women, these gender disparities harm children's food security, in two ways. First, income - including extra income - is more likely to be spent on improving the nutritional status of children under five if it accrues to women. Second, discrimination against small girls in feeding or health care not only cuts female life expectancy to the levels of males - elsewhere it is three to six years more - but is transmitted to later generations. Difficult pregnancies due to small womb size induce low birthweight, imperilling infant life and development. This helps explain why, for example, several South Asian regions suffer more frequent stunting and wasting than many African countries with lower calorie intakes and similar interhousehold distribution.²⁸

International refugees and internally displaced persons

International refugees were estimated to total 1.8 million in 1960 and remained around 1.5 million to 2.5 million until 1976. They rose to a peak of 18 million to 19 million in 1991-92, then fell to about 12 million at the end of 1998. In sub-Saharan Africa, refugees and internally displaced persons (IDPs) fell from the 1995 peak of 16 million to 12 million in 1997; in Asia, the number of refugees fell from 5.8 million at end-1993 to 4.5 million at the end of 1995, plus more than 1.7 million IDPs.

However, IDPs often suffer extreme hunger or worse, especially when displacement is: sudden; results from violence, drought or natural disaster; is prolonged; separates victims from land or other means of production; or concentrates on women, children or the elderly. These victims, often without male heads of family (who are away fighting), face multiple threats to food entitlements, health and hence nutrition, especially among children. Well-funded refugee camps can substantially reduce PEM, as was the case for Afghan refugees to Pakistan and Kosovar refugees to nearby areas, or can achieve less (as in Angola and the Sudan) owing to underfunding, exacerbated by repeated bouts of civil violence, inducing inflows of refugees who are underfed even in normal times, disrupting food access, or obstructing food flows.

STAPLES PRODUCTION AND ACCESS TO FOOD

Changing perceptions of availability, entitlements and production, 1945-2000

This section shows that the conventional wisdom of postwar years - that very poor, mainly farming populations first escape poverty mainly by increasing their output of food staples - was quite wise after all, despite later swings of "development fashions". Between 1945 and 1965, emerging nations, which were accountable for food security but seeking to conserve foreign exchange for industrialization, emphasized self-sufficiency in food staples. Few analysts questioned tropical countries' comparative advantage in domestic staples production. This was therefore seen during this period as the key to availability, which was expected to ensure HFA/HFS. However, as land shortages began to be felt, price and allocation bias against agriculture, as well as distortions of farm markets, made it harder to attain food production goals.

Three changes occurred after the mid-1960s:

• The green revolution reached many Asian and Latin American smallholders.
• Land transfer to the poor - much more of it than conventional wisdom suggests - raised the rates of employment per hectare and (more modestly) yields per hectare, and resulted in greater emphasis on food in production.²⁹
• From the mid-1970s, disincentives created by the anti-farm price bias weakened.

The resulting upsurge in smallholders' employment and staples production provided a local source of food, together with employment-based income entitlements to it, thereby slashing poverty and PEM in much of Asia and Latin America between 1965 and 1988. Since the early 1980s, the following factors have weakened the emphasis on expanding staples production in areas where PEM remains:

• Globally, population and effective demand have been far outpaced by staples output, stimulated in Europe by subsidies, and in many developing countries by the green revolution. For 40 years, staples prices have fallen by some 0.4 percent per year relative to manufactures prices; despite droughts in Africa there has been no credible "global food scare" since 1972-74. From this it was concluded that extra staples production may be undesirable.
• The global incidence of PEM (while still high) has fallen sharply. Remaining PEM is seen to be due mainly to inadequate food entitlements - usually determined by poverty. From this it was concluded that more food production, even in places with PEM, will do little to reduce it; and that attention should shift from PEM to other nutrition issues, i.e. micronutrients, infections, food safety and problems of increasingly old, obese populations.
• The progressive decrease in sustainable returns to water, land and agrochemicals have led to environmental stress. From this it was concluded that one must beware both of intensifying staples production, especially on a narrow genetic base, in lead green revolution areas - irrigated, well-fertilized plains - and (more justifiably) of expanding production into marginal lands.

However, expanded local food production must still be seen as part of the cure for nutrition problems. PEM remains the world's main cause of death and misery. The great majority of its 800 million victims are afflicted mainly because they (or their working parents or children) lack employment-based entitlements to food, i.e. they cannot produce, or earn enough to afford, sufficient food. Extra food entitlements to reduce PEM and to accommodate the expected 50 percent increase in working-age populations in Africa and South Asia by 2020 to 2025 will continue to depend on rising employment or self-employment in farming, mainly food production.

Food staples production, yields and income from employment

In the developing world, falls in poverty, caloric underfeeding and PEM have generally accompanied rapidly increasing food staples output. Table 14 shows why. People in these areas still depend mainly on farming and farm labour for employment income. For the poor, such income is overwhelmingly the main source of food entitlements. For the most food-insecure - the poorest, the rural, the remote - farm dependence is still
higher. Increased staples production nearby, generating
more employment and self-employment income to
ensure entitlements to reliable supplies, is usually the key to food security, until enhanced farm growth followed by
successful diversification has reduced employment
dependence on farming towards current East Asian and Latin American levels.

Those who are most food-insecure rely on nearby staple food production until they can diversify their income sources.

The years 1945 to 1959 saw some acceleration of staples growth, based mainly on land area expansion and (in Asia) irrigation. From the late 1950s, with population growth and urban expansion, more and more prime agricultural regions ran out of quality spare arable land, while the number of people seeking work grew faster than ever, and industrialization proved either slower, or much less job-rich, than had been expected by the planners. Fortunately, 1965 to 1985 were golden years for yield growth of main staples in Asia and Latin America, as the green revolution increasingly provided varieties that were friendly to employment-intensive small farmers. Table 15 summarizes the outcomes since 1961.

As is shown in Table 15, yield growth in Asia and Africa shows a downward turning point around the mid-1980s, although timing varies among crops and regions. Moreover, although staples yield growth still generates much more employment per unit of extra GDP than, say, grazing, industry or construction (and than most urban activity), the employment effect declined: a 20 percent rise in wheat or rice yield in Asia led to about 8 percent more employment in the mid-1970s, but only 3 to 5 percent by the late 1980s. It is not surprising that the regional and temporal distribution of staples yield growth matches that of reduced poverty, underfeeding and PEM.