Highlights
National average maize prices have doubled year-on-year, driving down people’s ability to purchase food. High prices are hampering food access in Southern Malawi.
Poorer households and female-headed households employ more negative coping strategies than wealthier households and maleheaded households. Households in rural areas appear to be more food insecure than those in districts with urban centers (Blantyre,
Lilongwe, Mzimba/Mzuzu).Households in Southern Malawi have more negative perceptions of food security than in other parts of the country. Countrywide, respondents describe relying on purchases of maize at high prices because of poor harvests.
Methodology: SMS Surveys
The methodology has been informed by WFP’s mobile Vulnerability Analysis and Mapping (mVAM) initiative, which involved using mobile surveys for food security monitoring. The project was initiated in 2013 and is now active in 15 countries. December 2015 marked the first round of remote data collection for Malawi. The survey was conducted using text messaging (SMS) on 1 December 2015, on a sample of 1000 household. Participants were randomly selected from a national database of mobile subscribers. Respondents opted-in to the SMS survey and were asked socio-demographic questions, coping behavior questions, and manual labor wages as well as an open-ended question on food security. Because of the limitations of SMS surveys, the questionnaire did not include the food consumption score, an indicator that does not perform well by text message. An airtime credit incentive of US 50 cents (or MWK 300) was provided to respondents who successfully completed the survey to increase completion rates.
It is acknowledged that SMS surveys contain an inherent response bias as well as biases towards better-off and literate households. As of 2013, 54% of households in Malawi had at least one mobile phone (Gallup 2013). Due to biases in our data, an attempt is made not to give precise estimates of food insecurity. Rather, SMS polls are used to capture patterns and trends. This first round of data collection provides the basis of a monitoring system that will track month-to-month changes.
The survey response rate was 27%, which is above the industry standard for SMS surveys. Respondents were reached in 27 out of 28 districts. 44% of respondents came from districts with urban centers (Lilongwe, Blantyre, Mzimba) and 56% came from rural districts. A higher portion of respondents were men (66% men vs. 34% women). Due to small sample sizes in rural areas, neighboring districts with similar characteristics were aggregated (Table 2). Where distributions are highly skewed, medians are reported. Additional information on methodology is available online.