Links between household agricultural production and nutrition

Summary of research¹

Location: Nigeria, Uganda, Ethiopia, Zambia, Mozambique, Tanzania and Nepal

What we know: It is generally accepted that agricultural development will improve household food and nutrition security but direct evidence of pathways is lacking.

What this article adds: A special issue of The Journal of Development Studies, featuring eight studies from Africa and South Asia, supports the hypothesis that household agricultural production has direct and important linkages with dietary patterns and nutrition. The analysis reflects challenges in establishing a link, e.g. which variables, data and methods to use. The main pathways of influence include income from agriculture; consumption of own production (or some combination of these); and factors linked to gender (women’s social status and empowerment in agriculture). Commercialisation of agriculture may not impact on or may negatively affect child nutrition. Livestock production emerged as particularly linked to nutrition. Effects will be determined by local conditions.

Global, national and local policies and programmes for agricultural development are recurrently justified based on their alleged role in improving food and nutrition security. However, strikingly little evidence is available to prove that a direct, household-level link between agricultural production and improved nutrition exists. A recent special issue of The Journal of Development Studies systematically and empirically tests whether a relationship between household agricultural production and nutrition can be found. Eight featured studies examine the relationship between agricultural production (crops or livestock), household dietary diversity, and children’s (and in some cases maternal) diet and anthropometric outcomes across countries in sub-Saharan Africa and south Asia (Nigeria, Uganda. Ethiopia, Zambia, Mozambique, Tanzania and Nepal). The introductory article, featuring highlights of each study and overall conclusions, is summarised here.

The main pathways through which expanded agricultural production can influence nutrition at farm-level include income from agriculture; consumption of own production (or some combination of these); and factors linked to gender. Increased household income from any activity, including agriculture, can alter the amount, composition and quality of food consumed and facilitate the purchase of health and nutrition-related goods and services. However, on the basis of previous evidence, the authors assert that the commercialisation of agriculture and the resulting shift away from staples to cash crops have not necessarily resulted in improvements in children’s nutritional status and can have negative nutritional consequences. Recent research, looking specifically at child nutrition outcomes, mirrors this view and shows that, while income is important, the Millennium Development Goal of halving the prevalence of underweight children is unlikely to be met through income growth alone. Where there are market imperfections, agricultural production by a household can influence household consumption, depending on what is produced and seasonal factors. The authors suggest that the mechanisms by which these effects occur are mediated by gender relations within the household, including women’s social status and empowerment in agriculture. Policies and programmes that seek to use agriculture as a direct means to improve food security and nutrition of agricultural households implicitly assume that these household-level effects exist and that their magnitude is economically meaningful.

The authors of this article summarise challenges in establishing a link between agriculture and nutrition at farm level. Firstly, there are challenges in identifying the right set of variables to analyse. Studies may focus on production diversity or types of agricultural products, and on nutrition inputs (e.g. dietary diversity or dietary intake) and/or nutrition outcomes (e.g. anthropometric measures). The selection of variables may be limited if using data not collected specifically to analyse the agriculture-nutrition link. There are also challenges associated with the use of appropriate data and methods to establish an agriculture-nutrition link. For example, national data sets provide broader validity, but tend to allow less detailed examination of local issues compared to case studies. The benefit of experimental data is that it is easier to establish a causal relationship between agriculture and nutrition; however experiments have well-known limitations. Observational data play an important role in trying to understand the farm-nutrition relationship, but identifying a causal impact between agriculture and nutrition outcomes can be difficult. Notwithstanding the substantial variation in measures, methods and approaches, eight papers in this special issue all trace some associations between agricultural production and diets or nutrition outcomes. The authors synthesise the results of each study in turn (see Table 1 for summary findings).

One study looks at the impact and intensity of participation in a biofortification (B-carotene-rich orange sweet potato) programme on vitamin A density, micronutrient adequacy and diversity of children’s diet in Mozambique (De Brauw, Fozenou & Moursi, 2015). In this study, the authors are able to trace positive impacts of programme participation on all outcomes, particularly on vitamin A density; there is increased impact with greater intensity of participation in the programme.

Two papers study the impact of engagement in agriculture and crop production diversity on dietary diversity, anthropometry outcomes and consumption of specific food groups in Nigeria (Dillon, McGee & Oseni, 2015) and Zambia (Kumar, Harris & Rawat, 2015). Both studies find positive associations between crop and dietary diversity; the Zambia study also shows a positive association between crop diversity and height-for-age Z-scores (HAZ) in children aged 24 months and older.

Three papers look at the role of livestock ownership in improving diets by increasing consumption of animal-source foods. One study from Ethiopia shows a strong effect of household cow ownership on increased frequency of milk and dairy consumption among young children (Hoddinott, Headey & Dereje, 2015). Data from Uganda show a positive association between livestock ownership and the consumption of a range of animal source foods (Azzarri, Zezza, Haile et al, 2015). Data from Tanzania show that ownership of large livestock holdings is associated with lower odds of stunting among pre-schoolers and that children aged five to nine from pastoralist households are better nourished (Slavchevska, 2015).

Two papers that use data from Nepal, one from a nationally-representative survey (Shively & Sununtnasuk, 2015) and one from the baseline survey for an impact evaluation of a multi-sectoral nutrition programme (Malapit, Kadiyala, Ouisumbina et al, 2015), find associations between production diversity and dietary diversity, and between production diversity and nutrition outcomes (HAZ in the Shively study and weight-for-height Z-score in the Malapit study). While identification issues prevent these two studies from making casual claims, these studies go beyond a general link between agriculture and nutrition to look more specifically into agricultural commercialisation (Shively et al, 2015) and the mediating role of women’s empowerment (Malapit et al, 2015). Shively et al use data from Nepalese agricultural households to detect a positive, if small, association between market orientation and HAZ for children under two. In this case, increased income generated by agricultural sales more than offsets possible adverse impacts associated with less food being available for own consumption. Authors of the second Nepal study find that women’s engagement in the community, control over income, reduced workload and the overall empowerment score are positively associated with better maternal nutrition. Control over income is also associated with better child HAZ and a lower gender parity gap improves children’s diets and HAZ.

Taken together, the authors of the introduction to this special issue support the hypothesis that household agricultural production has direct and important linkages with household dietary patterns and the nutrition of individual members. The magnitude of the impacts varies, probably as a result of differences between the studies with regard to several key factors, such as commodities, contexts and location and the intensity of programme participation. While links to crop production and diversity of production are found to matter in certain contexts, livestock production seems also to emerge as particularly important and positively linked to nutrition. The results suggest that support to agriculture can play a direct role in promoting nutrition, but that the effects might not be as dramatic as anticipated and depend on local conditions. The papers also highlight the limits of available data in precisely measuring farm-level causal relationships between agricultural production and nutrition. The authors conclude that there is a need to devote more efforts to filling the data gaps that make such analyses so difficult in the developing world.

Footnotes