‘New variant famine’ revisited: chronic vulnerability in rural Africa
by Lisa Arrehag, freelance consultant, Alex de Waal, Global Equity Initiative, and Alan Whiteside, University of KwaZulu-Natal April 2006

The ‘new variant famine’ (NVF) hypothesis was first published in The Lancet in 2003:

Our hypothesis is that the generalised HIV/AIDS epidemic in Southern Africa, first, helps to explain why many households are facing food shortage, and second, explains the grim trajectory of limited recovery. Four factors are new:

1. Household-level labour shortages due to adult morbidity and mortality, and the related increase in numbers of dependants.

2. Loss of assets and skills due to adult mortality.

3. The burden of care for sick adults and children orphaned by AIDS.

4. The vicious interactions between malnutrition and HIV.

The NVF hypothesis did not discount existing contributors to food crisis – drought, poverty, macro-economic disparities, poor agricultural policies and economic mismanagement – but noted that AIDS made these factors more severe and intractable. Hunger, we suggested, could become a structural feature of life for many people unless effective interventions were developed. This would have substantial implications for aid policy and programming, including a shift towards long-term social welfare for communities heavily impacted by AIDS, and the integration of AIDS and humanitarian activities.

Recent literature

New literature has allowed for a refinement and elaboration of elements of the NVF hypothesis. Recent work has shown the effect of adult illness and death on farm production among smallholder cotton farming households in Zambia, and new evidence demonstrating the impact of AIDS on the agricultural sector was presented in April 2005 at a conference organised by the International Food Policy Research Institute (IFPRI). In particular, the intersection of a high prevalence of HIV/AIDS and other concurrent shocks has been shown to be the cause of a serious production crisis in Makete district, Tanzania, while the absence of such additional stresses allowed Kagera Region to withstand such crisis. This paper draws on published research and two recent studies in Malawi and Swaziland to reassess the new variant famine hypothesis.

The evidence from Malawi and Swaziland Both Malawi and Swaziland are predominantly agricultural, poor and vulnerable, and are suffering protracted high-prevalence epidemics of AIDS. Swaziland is a small country with a population of 1.1 million people, two-thirds of them rural. In 2003, GDP per capita was $1,350 per annum. Classified as a ‘lower middle income country’, 69% of the population live below the poverty line. Agriculture and agro-industry form the basis of the economy, and smallholder agriculture employs about two-thirds of the population. Malawi has a population of 12.5 million, and a 2003 GDP per capita of $160 per annum. More than three-quarters of the economically active population are engaged in smallholder agriculture. In both countries, farming is labour-intensive, reliant primarily on hoe-cultivated maize in a single farming season.

Food crisis, including chronic malnutrition and recurrent famine, is common to many parts of Malawi and Swaziland. In both countries, 2005 was unusually bad. In June, it was reported that drought in Malawi meant that more than 4.2 million people (34% of the population) were unable to meet their food requirements. The following August, the Food and Agriculture Organisation (FAO) stated that the country was facing its worst food crisis in more than a decade. In Swaziland, the Ministry of Agriculture reported that nearly one-third of the population, about 330,000 people, needed food aid.

Both countries also face a severe HIV/AIDS epidemic. Adult prevalence among ante-natal clinic (ANC) attendees in Swaziland – 42.6% – is the highest in the world. According to the most recent government estimates, in Malawi prevalence among ANC attendees is 14.4%. Mortality in Swaziland has almost tripled over the past ten years, from about eight deaths per 1,000 in 1994 to about 23/1,000 in 2004 and adult mortality has also risen sharply in Malawi. In both countries, these increases are almost entirely due to AIDS.

HIV/AIDS and loss of household labour, assets and skills

Results of household studies in both countries during the 2002–2003 food crisis clearly show that HIV/AIDS was one cause of declining agricultural production. Evidence from Swaziland found a reduction in maize production of 54.2% in households with an AIDS-related death. Household data collected during emergency food security assessments in Malawi, Zambia and Zimbabwe in the second half of 2002 also show a severe impact on both crop production and income. In Malawi, households without an ‘active adult’ suffered a 26% drop in tuber production, a 53% fall in cereal production and a 51% reduction in cash crop income compared to households with at least one ‘active adult’.

These data do not allow us to evaluate whether affected households can recover from these setbacks, and whether the impact of AIDS is equally harsh in the absence of concurrent drought. However, a panel survey between 1990 and 2002 in five countries (Kenya, Malawi, Mozambique, Rwanda and Zambia) addresses this. It shows that, in AIDS-affected households, the mean crop income is lower following a prime-age adult death than in non-affected households across four of the investigated countries, including Malawi. In Malawi, the difference is even more pronounced in households experiencing the death of a household head; here, income was reduced by as much as 40% (from about $170 to $280).

There is also an important gender dimension to this impact. In Malawi, in households with a recent adult male death, the area planted is 32% lower than in households with a recent adult female death. This gender effect is especially critical in tobacco- or sugar-growing family enterprises, perhaps because specialised knowledge has been lost with the death of a male adult.

HIV/AIDS and rural livelihood coping strategies

The loss of adult labour gives rise to hard livelihood decisions. The nature of this decision varies, depending on the circumstances. Rural Malawians have often resorted to low-labour crops. The panel survey referred to above found that three-quarters of investigated households in Malawi changed their usual crop mix towards less labour-intensive crops in response to labour shortages and lack of resources. However, land is rarely left unutilised, reflecting its scarcity. In Swaziland, by contrast, decreasing the area under cultivation was a common response. According to one study, the area under cultivation has decreased by an average of 51% in households with an AIDS-related death, compared with 15.8% for households with a death that was not AIDS related.

Selling or liquidating assets is another coping response. In Malawi, one household study found that some 40% of those affected by chronic illness sold a portion of their assets to buy food or to pay medical or funeral expenses. This way, households are able to mitigate the short-term effects of adult mortality and other shocks. Over time, however, this can impoverish the household, increase its vulnerability to income shocks and decrease its use of cash inputs in crop cultivation, resulting in lowered production. Previous research has suggested that the AIDS epidemic may lead to a concentration in the ownership of cattle, as afflicted households sell off productive assets such as livestock to those with resources to accumulate them. Findings from Malawi show that vulnerable groups (defined as households with a chronically ill member) own fewer cattle than the general sample population. In Swaziland, households with an AIDS death experienced a 29.6% reduction in the number of cattle they owned.

HIV/AIDS and changing dependency patterns

In the worst-affected countries – such as Malawi and Swaziland – there is plentiful evidence of increasing numbers of AIDS orphans. In 2003, UNAIDS estimated that about 500,000 children (about 50% of the total number of orphans) in Malawi and 65,000 children (65% of the total number of orphans) in Swaziland below 17 years of age had lost one or both parents to AIDS. Caring for an increasingly large number of orphans is placing a tremendous burden on extended families and community networks. At the same time, however, kinship networks have proved resilient in providing at least a minimum level of care and socialisation for children orphaned by AIDS. Part of the reason for this is that African societies had very high levels of fostering before AIDS, and fostering ‘capacity’ had been underestimated.

HIV/AIDS and malnutrition

There is solid evidence that HIV/AIDS has changed the profile of child malnutrition in Southern Africa. Overviews of nutritional surveys during the 2002–2003 drought found clear signs that double orphans have a much higher prevalence of malnutrition compared with children with one or both parents living. It also found that, although child malnutrition rates were higher in rural areas (which tended to have lower HIV prevalence), the decline in nutritional status was most marked closer to towns – areas which have traditionally enjoyed better food security, but which are now affected by higher prevalences of HIV/AIDS. If this finding is correct and generalisable, it points to the emergence of a new category of the vulnerable: children in high-prevalence urban and peri-urban communities. Finally, the survey reported that, while drought and the presence of an HIV/AIDS epidemic were each independently associated with a decline in the nutritional status of children, this decline increased where both factors were present. There is also preliminary evidence that the rebound in nutritional status after the end of the drought in 2003 was less robust than expected.

Other aspects of the relationship between malnutrition and the HIV/AIDS epidemic remain speculative and under-researched. Little is known about the indirect impacts of the HIV/AIDS epidemic on the spread of childhood infectious diseases, and studies of adult nutrition and HIV infectivity and virulence are complex, contradictory and/or inconclusive.


Recent research supports our view that AIDS is challenging rural livelihoods, undermining resilience to other shocks and stresses and creating new patterns of malnutrition. Humanitarian and development agencies will need to consider this scenario carefully as they contemplate future activities in heavily AIDS-affected areas. Responses to losses of adult labour and the availability of additional resources are emerging as critical factors in households’ resilience in the face of the epidemic. Specific data from the 2002–2003 crisis in Southern Africa indicate vicious interactions between HIV/AIDS and drought. Preliminary evidence from the current crises in Malawi and Zimbabwe tell a depressingly similar story.

However, the major probable impacts of the epidemic on food security remain in the future. AIDS is a long-wave event, and its secondary impacts unfold over an even longer timescale, often masked by the clearer effects of climatic, economic and political volatility. But in the hardest-hit areas, where HIV/AIDS coincides with drought, agricultural crisis and the decay of basic governance structures, the change is unmistakeable.

In many areas of Southern and Eastern Africa, each turn of the cultivating seasons is seeing a small, significant and usually negative change in rural livelihoods. While communities are resourceful and inventive in responding to the stresses they face, a significant proportion of the rural population is being ground down into chronic destitution. This is preventable, but it is not being stopped. Until it is, we face the prospect of major, ongoing interventions to support social welfare in affected communities.

Lisa Arrehag (lisa.arrehag@bredband.net) is a freelance consultant. Alex de Waal (dewaal@fas.harvard.edu) is a Fellow of the Global Equity Initiative, Harvard University, and Program Director, Social Science Research Council. Alan Whiteside (whitesid@ukzn.ac.za) is Director of the Health Economics and HIV/AIDS Research Division and a professor at the University of KwaZulu-Natal. For an earlier discussion of ‘new variant famine’, see Alex de Waal, ‘New Variant Famine: Hypothesis, Evidence and Implications’, Humanitarian Exchange, no. 23, March 2003, http://www.odihpn.org/report.asp?ID=2512.