Contents - Previous - Next

This is the old United Nations University website. Visit the new site at

Crop improvement programmes and nutrition in Malawi: Exploring the links

Anne E. Ferguson, Ann V. Millard, and Stanley W. Khaila



A model has been developed to conceptualize the relationships among a series of complex, interrelated factors that account for high levels of malnutrition among children in Malawi. It can be used as a tool for theoretical analysis and for programme evaluation. In this paper, the model is used to examine how interventions in crop-breeding strategies interact with existing rural social structure and influence household nutritional well-being in Malawi. It recognizes three tiers of causation: proximate, intermediate, and inclusive. The proximate tier refers to the immediate biomedical causes of malnutrition; the intermediate causes are those behavioural patterns that increase exposure to the proximate causes; and inclusive causes refer to the broad, social, economic, and cultural processes and structures in which the proximate and intermediate causes of nutritional status are embedded.


Children are considered stunted when their height is less than 97% of that of all children in a well-nourished population of the same age. In Malawi, nearly one-third of children die before their fifth birthday, and more than half of those under age 5 are stunted, the result of chronic, long-term malnutrition [1-3]. A recent study conducted in Malawi's southern region (Zomba South) revealed that approximately two-thirds of children suffer at least moderate malnutrition in the post-harvest period, when food is most abundant, a figure that increases to nearly three-quarters before the next harvest [4].

A series of complex, interrelated factors account for these high levels of malnutrition. A model developed to conceptualize these relationships recognizes three tiers of causation: proximate, intermediate, and inclusive. The model can be used both as a tool for theoretical analysis and for programme evaluation. We used it to examine how interventions in crop-breeding strategies interact with existing rural social structures and influence household nutritional well-being in Malawi.

Causes of malnutrition: A conceptual model

The model is presented in figure 1 (see FIG. 1. Three tiers of causes of high rates of child mortality). The proximate, intermediate, and inclusive tiers designate three categories of processes connected to the patterning of child mortality. The proximate tier refers to the immediate biomedical causes of malnutrition. It has long been recognized that malnutrition is an important contributor to infant and child mortality in developing countries. A synergistic interaction of malnutrition, diarrhoea, and lower respiratory tract infections is common, producing high rates of morbidity and mortality [5-7]. Many medical and public health programmes have addressed this proximate level (for example, through child-feeding programmes, immunization campaigns, and oral rehydration therapy) in an effort to improve health status.

The intermediate causes of malnutrition are behavioural patterns that increase exposure to the proximate causes. Anthropologists have focused on culture-specific patterns of child care that unintentionally predispose children to malnutrition and disease. For example, insufficient feeding of newly weaned children and their frequent exposure to infection may create a mortality crisis early in life. In many societies, including Malawi, children may be weaned to a high-carbohydrate diet that lacks sufficient calories and other nutrients if feeding is not frequent [811]. Also, some traditional ethnomedical treatments of illness in young children can initiate the downward spiral towards malnutrition [12, 13]. Interventions at the intermediate level are common and generally take the form of nutrition education and child-care training programmes.

The inclusive tier refers to the broad social, economic, and cultural processes and structures in which the proximate and intermediate causes of nutritional status are embedded. It is at this tier that we consider how social stratification along class, ethnic, and gender lines, and the power relations that support these divisions, shape the distribution of malnutrition and disease at the household and community levels. Elsewhere, we have termed these processes "ultimate" causes, a term drawn from Greek philosophy to emphasize that these causes are crucial determinants in the broad patterning of child mortality even though they are conceptually abstract and distant from the actual event of death [14, 15]. Analysis of this tier takes into account how developing countries like Malawi are linked into capitalist structures of world trade and international relations; what the implications of these relationships are for class, ethnic, and gender relations within Malawi; and what the health and nutritional outcomes of these social and economic divisions are. Explicit nutrition-related interventions at the inclusive level typically take the form of income-generation projects and formal and informal education programmes. They are directed particularly towards women, who are regarded as the keepers of children's health and as those most vulnerable to the negative consequences of development and change.

More frequent than programmes with explicit nutrition-related impact, however, and more significant in terms of health, are other types of interventions, including agricultural development projects, changes in pricing, and marketing structures or structural adjustment programmes. Indeed, it can be argued that many of these initiatives result in unintended consequences that make nutrition-related interventions necessary. We return to this point when we consider how crop-improvement programmes in Malawi may influence nutritional well-being.

Anthropological investigations of malnutrition and nutrition programmes frequently employ an approach that addresses only one tier of the model. The results have typically been partial explanations, including some of the victim-blaming variety, or only short-term improvements in health status. We suggest that basic research and health interventions can be improved if a process approach is used, one that takes into account the interactions among factors on the three tiers of the model. It is not crucial that planned intervention or research projects place equal weight on all three, but addressing one tier without acknowledging the potential impact of the others limits theoretical understanding and efficacy in dealing with malnutrition. To illustrate this point, we examine the potential nutritional consequences of different strategies of crop improvement in Malawi.

Crop-improvement strategies: Implications for nutrition

Macro-economic considerations

During the 1970s, Malawi's agricultural sector and economy in general experienced one of the fastest gross domestic product (GDP) growth rates in sub-Saharan Africa. This growth came from the rapid expansion of the estate sector, particularly tobacco produced for the export market [16, 17]. While exports of tobacco, tea, and sugar were increasing, food production was declining, a situation that led the government to increase the prices it paid for maize. The resulting bumper crops of the early 1980s allowed Malawi to export maize to surrounding nations and fuelled the image of the country as one of the economic miracles of Africa. At the same time, however, life expectancy in the country was only 46 years, and agricultural surveys revealed high levels of malnutrition and child mortality. These findings suggested that the economic growth model of the 1970s and early 1980s had benefited only a few, while the majority of small-holder farmers had not seen their circumstances improve.

To deal with this situation and a series of external economic shocks [17, 18], the government has instituted numerous changes in agricultural development strategies over the last five years. In large part, these changes resulted from the conditions attached to structural adjustment loans (SALs) made to the government by the World Bank and the International Monetary Fund (IMF). The SALs address Malawi's deteriorating terms of trade and balance of payment problems by promoting new agricultural pricing policies, state divestiture of agricultural marketing functions, growth of the private sector, removal of fertilizer and other subsidies, and increased production of food and export crops [17, 18].

These initiatives have potential impacts on health and nutrition. For the purposes of this paper, we focus on government and donor efforts to increase food crop production, particularly the production of beans, which are an important source of protein in the diet. Since 1981, researchers with the Bean/Cowpea Collaborative Research Support Program (CRSP) project at Michigan State University have worked with the Malawian National Bean Program at Bunda College to improve bean production in the country. Initial studies centred on understanding the biological and social factors responsible for the great amount of genetic diversity in Malawian beans [19-23]. Beginning in 1988, the focus of research shifted towards making use of this information in the breeding programme. Specifically, the CRSP is attempting to develop an innovative strategy that will maintain a high level of diversity in the bean crop. Our three-tier model can be used to explore the shape of this programme, consider some of the pressures and competing demands faced by its researchers, and examine the potential nutritional ramifications of the crop-improvement strategy.


Proximate and intermediate tier considerations

Malawian farming systems vary by ecological zone, size of land holding, crops grown, and several socioeconomic characteristics, including the type of marriage and descent system and the gender of the household head. Maize is the staple food, and is so important, in fact, that the words "maize" and "food" are used synonymously in some areas of the country [41. Reflecting the fundamental importance of this crop, it provides over 70% of calories in the average diet, and 70% of the land in the small-holder sector is planted in maize each year. Most of it consists of unimproved, local varieties that are often intercropped with beans or other pulses. Although hybrid maize has been promoted extensively, adoption rates have been low, never exceeding 10% and currently 5% [24].

As part of the CRSP-sponsored research, a small-scale longitudinal survey of bean-production practices began in 1986 in Dedza Hills in Malawi's central region. This survey revealed that households planted an average of 13 bean varieties, almost all of them local types. Electrophoretic analysis of bean enzymes revealed that more than three-quarters of the beans grown were large-seeded Andean types. Most of the work involved in bean production is done by women, who are more familiar with bean plant characteristics and nomenclature than men. On many larger farms in the small-holder sector (over l.5 ha), the farmers we talked to preferred to mono-crop beans, although they sometimes planted them in mixed stands. On smaller farms, beans were usually planted with maize and other crops.

Intercropping has numerous benefits, especially for small-scale farmers with under 1 ha of land. From an agronomic standpoint, the genetic and morphological diversity of intercropping makes more efficient use of water, soil, and fertilizer resources, suppresses weeds, and reduces insect and disease infestation. Socioeconomic advantages include more efficient use of land and labour, minimization of risk through distribution of food production and income over the agricultural cycle, and diversification of the diet [25-27]. Intercropping can increase total yield per unit of land by producing higher individual plant yield, higher total plant population densities. and a reduced competition among crop species [28|.

Many of these benefits also occur when several varieties of a single crop are planted in the same field. This practice stabilizes production by reducing the risk of disease epidemics and weather-related losses. The women in Dedza Hills were aware of the advantages diversity confers and gave two reasons for keeping large numbers of bean varieties. First, they reported that they needed many varieties because if some failed during the growing season, others would survive to feed their families. In fact, women planted beans that matured at different times to guarantee a steady supply over the growing season. Second, they held that each variety had its own characteristics and fulfilled different household needs. For example, some varieties produced superior pods or leaves, important items in the diet immediately preceding harvest, when food supplies were often low. Other varieties were grown because they were fast-cooking, stored well, or were easy to sell on the local market.

Recognition of the importance of crop diversity for family well-being is reflected in women's answers when they were asked their reasons for selecting the four bean varieties they planted in the greatest quantities during the 1987 growing season. In descending order of importance, they identified yield, taste considerations, cooking quality, marketability, date of maturity, health-related issues, insect and disease resistance, and ability to withstand environmental stresses. Although yield ranked first, it is significant that three-quarters of the reasons for selecting varieties had to do with other factors. Chief among these were taste considerations, cooking quality, and health concerns, which accounted for nearly half the responses.

We are currently exploring the relationships among size of land holding, labour constraints, and the amount and type of crop diversity. While intercropping is more common on farms smaller than 1 ha, these farmers have less diversity in the varieties of beans that are grown. For example, in surveys carried out in Dedza Hills and the Lilongwe Agricultural Development District, where holding sizes are smaller, the average numbers of varieties grown were 13 and 6 respectively. Similar results indicating a decline in diversity on small farms were also found in a survey of 300 farm families in Matapwata Extension Planning Area in the Blantyre/Shire Highlands Rural Development Project in the southern region.

The nutritional implications of different amounts and types of genetic diversity of crops have not been examined systematically in Malawi, so we have little to report at the proximate level of analysis at this time. There is reason to believe, however, that, in the absence of agricultural inputs that stabilize production, declines in crop diversity may result in greater nutritional risk, at least for certain strata of farmers. In our Lilongwe survey, many households with less than I ha of land did not produce enough maize and beans to meet consumption needs during the year. In fact, these households often consumed most or all of their planting stock and had to work for other farmers to obtain food and seed. Thus. opportunities to select seed types were restricted-farmers had to plant assortments of the available varieties. We hypothesize that women in these households find their families' precarious food supplies further jeopardized because they cannot obtain sufficient seed, especially of early-maturing varieties that provide food during the lean months.


Inclusive tier considerations

Although beans and maize are basically considered to be one crop by Malawian small farmers, for the purposes of agricultural research they are treated as separate commodities and assigned to different national research institutions. Bunda College of Agriculture, as mentioned earlier, has the mandate for beans, whereas most research on maize is carried out by the Ministry of Agriculture in collaboration with various donor agencies. Until recently, interactions between researchers in the two programmes have been relatively infrequent.

This division of the agricultural research establishment into commodity groups is only one element in the inclusive tier of analysis that we consider briefly. We are also interested in examining the pressures and competing demands placed on the national bean programme in its efforts to design and implement a plant-breeding strategy. Farmers themselves are one source of these pressures. We noted some of the advantages that women felt came from maintaining a diverse planting stock. Many agricultural researchers are aware of these benefits, and acknowledge that intercropping and mixed variety planting are rational strategies for farmers, especially in the absence of ready access to credit, fertilizers, improved seed, and other inputs.

Nevertheless, small-scale farmers' interests are only one element in the equation determining the type of plant-improvement strategy adopted. Researchers who want to improve indigenous cropping systems and to maintain and use indigenous germplasm in breeding programmes face strong pressures from other sources to moderate their views. Ultimately, these pressures derive from powerful economic concerns, including seed companies and other large distributors of agricultural inputs, who promote input-dependent styles of agricultural development. Bolstering these business interests is an ideology of scientific research in agriculture often promoted by international agricultural research centres and other international research organizations. In this view, problems in agriculture are defined as technical and production-oriented rather than social or political, and, not surprisingly, technological solutions are sought [29, 30].

Many commodity-research programmes in Malawi have strong ties to the centres. In the case of beans, the link is with the International Center for Tropical Agriculture (CIAT). This organization recently established a southern African programme incorporating Malawian scientists, and it anticipates posting a plant breeder to Malawi in the near future. As a large-scale international research organization, CIAT brings with it, when it enters a country, its vision of the nature of constraints on bean production and promotes solutions based on this internationalized perspective. In Malawi, disease and insect infestations, particularly bean fly, have been identified as the major constraints on increasing bean yields. The solution proposed for these problems is to introduce germplasm with pest resistance. This usually involves screening material from various international nurseries (bean fly nurseries, rust nurseries, nurseries for anthracnose, angular leaf spot, halo blight, and other prevalent diseases) and releasing the varieties showing resistance, or attempting to introduce the desired trait into local material.

In Dedza Hills most of the beans farmers grow and prefer are the large-seeded varieties of Andean origin. Our surveys show that farmers throughout the country share these preferences [21, 31]. The CIAT material, in contrast, is often small-seeded and Meso-american in origin. Furthermore, many of the sources of resistance found in these nurseries are in the small and black-seeded types disliked by most Malawians. While these desired resistance traits can be introduced into more accepted Malawian seed types, there appear to be lethality barriers between gene pools, and such crossings take time to accomplish. The net effect of these factors and of the proliferation of nurseries is that Malawian researchers' time may be increasingly spent screening germplasm that is inappropriate for Malawian farmers, while efforts to screen and identify sources of resistance in their indigenous germplasm fall by the wayside.

These pressures against serious evaluation of local germplasm are coupled with other more generalized pressures that undermine the maintenance of genetic diversity on small farms. One of the most significant and pervasive of these is the focus on yield. The objective of most crop-improvement programmes is to develop higher-yielding disease- and insect-resistant varieties. Typically, these varieties are promoted as replacements for. rather than additions to, existing crop inventories maintained by farmers. Reflecting this orientation, the national seed companies in Malawi and Tanzania multiply and sell only one or two varieties of beans at a time.

This yield-focused and sequential approach to crop development contrasts with existing bean-production practices and preferences. Although farmers want high-yield varieties, by themselves these varieties will not meet the full range of household needs, and if promoted exclusively, they could have detrimental nutritional effects under current farming conditions.


The competing demands on the Malawian national bean-improvement programme are likely to remain a continuing source of debate in the years to come. Researchers in the programme must attempt to strike a balance between small-scale farmers' interests, which are expressed in the idiom of crop diversity, and the pressures from donor and Ministry of Agriculture officials to pursue commodity-oriented research, emphasizing use of exotic germplasm and focusing narrowly on increasing yields.

The three-tier causal model of malnutrition provides a useful framework for considering the complex series of biological, behavioural, and socio-economic interactions that give rise to different patterns of malnutrition. In the Malawian case, considerable stratification exists within the small-holder sector. Size of land holding, types of crops grown, gender of the household head, and other variables differentiate groups of farmers from one another. The inclusive tier decisions in the agricultural arena that shape plant improvement strategies are likely to have different nutritional consequences for various groups of Malawian small holders. Malawian agricultural scientists are not alone in facing such choices; similar pressures exist throughout eastern and southern Africa. The important questions are whose interests will be served by the plant-breeding and improvement strategies that are instituted, and whether these strategies will increase or decrease differentiation and poverty.


  1. Center for Social Research, University of Malawi. The Malawi maternal and child nutrition study (MMCN): interim findings from the baseline census. Presented at the Household Food Security and Nutrition Workshop, Chancellor College, Zomba, Malawi, 29-31 Aug 1988.
  2. Center for Social Research, University of Malawi. The characteristics of nutritionally vulnerable sub-groups within the smallholder sector of Malawi: a report from the 1980/81 NSSA. Zomba. Malawi: Chancellor College, 1988.
  3. Government of Malawi and UNICEF. The situation of children and women in Malawi. Lilongwe, Malawi, 1987.
  4. Peters PE, Herrera MG. Cash cropping. food security and nutrition: the effects of agricultural commercialization among smallholders in Malawi. Final report to US Agency for International Development Project PDC-0082-G-SS-6213-00. Cambridge, Mass, USA: Harvard Institute for International Development, 1989.
  5. Hull TH, Rhode JE. Prospects for rapid decline of mortality rates in Java: a study of the causes of death and the feasibility of policy interventions for mortality control. Working paper series, no. 16. Jogjakarta, Indonesia: Population Institute, Gadjah Mada University, 1978.
  6. Mata LJ. The nutrition-infection complex and its environment factors. Proc Nutr Soc 1979;38:29-40.
  7. Scrimshaw NS, Taylor CE, Gorden JE. Interactions of nutrition and infection. Geneva: World Health Organization, 1968:11-142.
  8. Cassidy CM . Benign neglect and toddler malnutrition. In: Green LS, Johnston FE, eds. Social and biological predictors of nutritional status, physical growth and neurological development. New York: Academic Press, 1980: 109-36.
  9. Cassidy CM. Protein-energy malnutrition as a culture bound syndrome. Culture, Medicine Psychiatry 1982; 6(4):325-5.
  10. Scrimshaw SCM. Infant mortality and behavior in the regulation of family size. Pop Dev Rev 1978; 4:383-403.
  11. Quinn V, Chiligo M, Gittinger JP. Household food and nutritional security in Malawi. Paper presented at the symposium on Agricultural Policies for Growth and Development, Mangochi, Malawi, 31 Oct-3 Nov 1988.
  12. Scheper-Hughes N. Maternal detachment and infant survival in a Brazilian shantytown-a critique of bonding theory. Soc Sci Med 1984;19:535-46.
  13. Nations MK, Rebhun LA. Angels with wet wings won't fly: maternal sentiment in Brazil and the image of neglect. Culture, Medicine Psychiatry 1988;12:141-200.
  14. Ferguson AK, Millard AV, Khaila SW. Agriculture and nutrition in Malawi. Paper presented at the American Anthropological Association Meeting, Washington, DC, 15-19 Nov 1989.
  15. Millard AV. A causal model of child mortality. Paper presented at the American Anthropological Association Meeting, Chicago, 111, USA, 18-22 Nov 1987.
  16. Kydd J, Christiansen RE. Structural change in Malawi since independence: consequences of a development strategy based on large-scale agriculture. World Dev 1982;10(5):355-75.
  17. Christiansen RE, Kydd J. The political economy of agricultural policy in Malawai. Washington, DC: World Bank working paper, in press.
  18. Lele U. Structural adjustment, agricultural development and the poor: some lessons from the Malawian experience. Paper presented at the symposium on Agricultural Policies for Growth and Development, Mangochi, Malawi, 31 Oct-4 Nov 1988.
  19. Ferguson AK. Socioeconomic factors influencing genetic diversity in Malawian bean landraces, central region. Malawi/Michigan State University Project, Socioeconomic Annual Report. East Lansing, Mich, USA: Michigan State University, Bean/Cowpea CRSP Women in Development Program, 1987.
  20. Ferguson AK, Sprecher S. Women and plant genetic diversity: the case of beans in the central region of Malawi. Paper presented at the American Anthropological Association Meeting, Phoenix, Ariz, USA, 16-20 Nov 1987.
  21. Barnes-McConnell P. Culture and agriculture in northern Malawi: social contributions to the maintenance of genetic diversity in beans. East Lansing, Mich, USA: Bean/Cowpea CRSP Management Office, Michigan State University, 1989.
  22. Martin G, Adams MW. Landraces of Phaseolus vulgaris (Fabaceae) in northern Malawi. I. Regional variation. Econ Botany 1987;41(2):190-203.
  23. Martin G, Adams MW. Landraces of Phaseolus vulgaris (Fabaceae) in northern Malawi. II. Generation and maintenance of variability. Econ Botany 1987;41(2): 204-15.
  24. Carr S. Modification and extension of the national rural development program. Paper presented at the symposium on Agricultural Policies for Growth and Development, Mangochi, Malawi, 31 Oct-4 Nov 1988.
  25. Francis CA. Development of plant genotypes for multiple cropping systems. In: Frey KJ, ed. Plant breeding II. Ames, Iowa, USA: Iowa State University Press, 1981:179-232.
  26. Francis CA, Flor CA, Temple SR. Adapting varieties for intercropped systems in the tropics. In: Multiple cropping. ASA Special Publication no. 27. 1976:235-54.
  27. Hamblin J, Zimmerman MJO. Breeding common bean for yield in mixtures. Plant Breeding Rev 1986;4:245-72.
  28. Andrews DJ, Kassam AH. The importance of multiple cropping in increasing the world food supplies. In: Multiple cropping. ASA Special Publication no. 27. 1976:110.
  29. DeWalt BR. Halfway there: social science in agricultural development and the social science of agricultural development. In: McCorkle CM, ed. The social sciences in international agricultural research: lessons from the CRSPs. Boulder, Colo, USA: Lynne Rienner, 1989:3961.
  30. Jennings BH. Foundations of international agricultural research: science and politics in Mexican agriculture. Boulder, Colo, USA: Westview Press, 1988.
  31. Ayeh E. Evidence of yield stability in selected landraces of bean (Phaseolus vulgaris L). Exp Agriculture 1988;24: 1 -7.

Contents - Previous - Next