Determinants of household vulnerability to food insecurity in Malawi

Location: Malawi

What we know: Subsistence farming households are vulnerable to food insecurity.

What this article adds:  A household level study examined food insecurity and largely agricultural determinants in two semi-arid districts in Malawi. Maize was used as a proxy for measuring household vulnerability. The main determinants of household vulnerability to food insecurity were income, household size, land size, access to on-farm employment, climate information and modern agricultural technologies. Female-headed households were more vulnerable. Household income plays a major role in food security. Policy interventions are needed regarding women’s access and control of land; diversification of livelihoods; communication on climate change and production technologies; and adaptive support to low income households.

A recent study examined household vulnerability to food insecurity and its determinants in two semi-arid districts in Malawi. The study was conducted in two Extension Planning Areas (EPAs), Mitole and Manjawira, located in Chikhwawa and Ntcheu districts respectively. The districts are semi-arid, characterised by variable and erratic rainfall patterns, dry spells, floods and droughts.  The main food crop grown in the areas is maize; however, other crops such as millet, sorghum, rice, sweet potatoes, Irish potatoes and beans are also grown. The population in the two areas is composed of smallholder subsistence farmers who mainly depend on rain-fed agriculture for their livelihoods.

A multistage sampling technique was employed to select 100 households from each EPA, making a total of 200 households. Household interviews were conducted to collect cross-sectional data from February to March 2010, using a pretested structured questionnaire. Data collected included maize harvests and purchases, land size, household size and composition, income, access to resources and the characteristics of the household head.

Maize was used as a proxy for measuring household vulnerability to food insecurity because over 90% of the rural population depends on it. It was measured in terms of total maize available (own harvest plus purchases) per adult equivalent. The study considered adults as persons older than the age of 16 years. Children aged 16 years and under were weighted as half an adult. Estimates show that adults require 270 kg of maize per year and children about 135 kg. Maize contributes roughly 55% of total caloric intake. Therefore, per capita consumption requirement is estimated at 126 kg of maize per adult per year. All children in a household were converted to adult equivalents and this factor was used in calculating household vulnerability to food insecurity. The household vulnerability to food insecurity measure was calculated by dividing total maize available (maize harvested plus maize purchased) by the number of adult equivalents multiplied by 126 kg.

A simultaneous equation model was used to assess the determinants of household vulnerability to food insecurity. Most of the explanatory variables are agriculture related. Household vulnerability to food insecurity was conceptualised as the relationship between the amount of maize available (from own farm production and purchases) and household and farm characteristics, income, and on-farm and off-farm employment. Land size cultivated and education were presumed endogenous because they are influenced by other factors such as income and household size, which are also explanatory variables in the model.

The analysis of the study data found that the main determinants of household vulnerability to food insecurity were income, household size, land size, access to on-farm employment, climate information and modern agricultural technologies. The results are consistent with previous findings and the direct relationship between land size and food security is consistent with previous studies. Furthermore, female-headed households were found more likely to be vulnerable to food insecurity than male-headed households because of poor access or control over resources such as land, climate information, technologies and income, which are the main determinants of household vulnerability in the study area.  These findings may result from cultural beliefs, whereby access to resources for food production such as land and inputs is low among women in developing countries. Therefore, policy interventions that enable women to gain control and access over land and those that promote access to climate information and technologies on improving production are likely to have more community benefits than those that focus on men.

The findings also show that household income plays a major role in food availability at the household level. Households that do not produce food because of climate factors can still achieve food security through purchases. The results imply that agricultural policy should consider factors related to food production and household income to achieve household food security.  In case of crop failure because of climate and other factors, households can still achieve food security through purchases if they have sustainable sources of income. Hence, policy intervention should promote diversification of livelihoods. Alternative sources of livelihood, such as small and medium scale businesses, may assist households to become food secure. Furthermore, agricultural policies that support low income households to adapt through the provision of farm inputs and credit facilities are crucial in achieving household food security. Moreover, policy interventions are required to promote the adoption of modern agricultural technologies and to improve the dissemination of climate information among the rural households, especially those that are affected by climate variability. Interventions should also focus on localising the collection and recording of climate factors in rural areas. This can be carried out through capacity building in collecting and recording climate factors using locally available materials to ensure that data on climate are available at local level to improve future prediction of climate and weather forecasts. In line with this, strong social networks among communities may provide a platform for introducing interventions and new policies.  

Footnotes