In the 2016 Global Hunger Index (produced by IFPRI, Concern International, and ), Malawi ranked 88th out of 118 countries, with 20.7 percent of the population suffering from undernourishment and 42.4 percent of children under 5 years of age suffering from stunting. In the lean season, food and nutrition security poses even more of a challenge; according to an assessment by the Malawi Vulnerability Assessment Committee, 2016 lean-season food insecurity (stretching from October 2015 – March 2016) was forecast to affect around 2.8 million people.
As a result of this assessment, the Government of Malawi, along with several development partners, scaled up social assistance programs, including a food transfer program (called MVAC assistance in the paper). A recent article in The Journal of Nutrition examines how such food transfer programs can impact child nutrition and household food security during these lean months. According to the study, such transfer programs may play an important protective role for young children.
The paper is based on a longitudinal, quasi-experimental study based on two survey rounds in the food-insecure Zomba district. The baseline survey was taken in September 2015, during the post-harvest season, while the follow-up survey was conducted in February 2016, during the peak of the lean season. Sixty communities were randomly selected for data collection; within that group, around 20 households were randomly selected from each community for interviews. A total of 1199 households ended up being surveyed; 175 households received the food transfer and were considered the treatment group. Balancing tests were performed to ensure similar characteristics at baseline. The study looked at household expenditure and food consumption, children’s dietary diversity, and children’s nutritional status (based on anthropometric measurements). The authors estimate the impacts of the food transfer program using a combination of propensity score matching and difference-in-difference methods.
The food transfer program provided households with rations on a quarterly basis. Rations included 50 kg of maize, 10 kg of peas, and almost 2 kg of fortified vegetable oil. The value of the transferred food basket depended on current market prices and ranged from 13,000 to 18,000 Malawian kwacha (1 MKW was equal to 25 USD when the paper was written).
The study first presents average household characteristics at baseline. According to the findings, at baseline, all surveyed households spent an average of 78 percent of their total expenditures on food; this spending on food was 13 points higher than in 2010-2011. Sixty percent of the food consumed by households came from their own production. Child stunting and wasting at baseline were similar to the levels seen in the 2015-2016 Demographic Health Survey (around 42 percent and 1 percent, respectively).
By the follow-up period, average food expenditure for the control group (the group that did not receive the food transfer) fell by 2 percent. The authors also found changes in household consumption patterns and nutrient intake between the post-harvest and lean seasons. Throughout the whole study group, protein consumption dropped by 31 percent and iron consumption dropped by 15 percent, while the number of people subsisting on less than 1800 calories per day increased by 10 points.
Children’s dietary diversity scores and food variety scores both dropped by the follow-up survey in control households, by 26 percent and 30 percent, respectively. The prevalence of child stunting (HAZ) and wasting also increased by 4 and 3 points, respectively.
However, things were different for households in the treatment group, particularly for child outcomes. Among these households, food expenditure increased by 19 percent from baseline. During the lean season, children’s average dietary diversity scores and food variety scores declined by only 11 and 13 percent, respectively (compared to the 26 percent and 30 percent declines seen in the control group). Children’s average weight-for-height score (WHZ) increased by 0.09 standard deviations in the treatment group, compared to a decrease of 0.17 SDs in the control group. In addition, household nutrient intake was lower in the control households for all of the nutrients considered (protein, iron, zinc, and vitamin A).
The authors do note that coverage of the food transfer program (15 percent of the surveyed households) was fairly low, and that targeting of the program could be improved to better focus on the poorest households. Analysis revealed that the share of beneficiaries from the top two socioeconomic quintiles was equal to the share of beneficiaries from the lowest two quintiles. The study also found evidence of both inclusion and exclusion errors; 35 percent of households classified as “non-poor” in terms of total expenditure were included in the program, while 85 percent of households classified as “poor” did not receive the food transfer. Households interviewed by the authors attributed these errors to favoritism by village leaders, who have significant say in who receives food transfers and other benefits, as well as to an overall limited amount of transfers available. Thus, the authors conclude that better targeting of food transfers and other social safety net programs could further improve these programs’ effectiveness.
By: Sara Gustafson, IFPRI