Two weeks ago in Rome, at a meeting at the United Nations Food and Agriculture (FAO), Mozambican Agriculture and Rural Development Minister Celso Correia claimed that 90 per cent of Mozambicans enjoyed food security.
“That is, they manage to eat three meals a day”, Correia told FAO’s Director-General, Qu Donavu, adding that there was no risk of famine or food insecurity in the country, with just under 10 per cent of the population in a situation of food insecurity.
Correia’s words that 90 per cent of the population eat three meals a day caused an uproar, with opposition parties and civil society finding his claim outrageous.
Leading the charge was the Graça Machel’s Foundation for Community Development (FDC), which said in a press release that Correia’s claim contrasted “with the reality experienced by millions of citizens in Mozambique. People are dying of hunger in our country”.
FDC thought that the “three meals a day” also did not seem to dovetail with other official statistics, namely the World Food Programme, the United Nations Development Programme, and Mozambique’s own National Statistics Institute (INE).
But today, Minister Correia doubled down and he had evidence to buttress his earlier claim: he launched a report on food security undertaken by the ministry’s Technical Secretariat for Food Security and Nutrition (SETSAN).
According to the report, between October 2022 and March 2023, SETSAN surveyed 12,890 families in 94 per cent of the country’s 161 districts, including municipal districts.
But a study covering 151 districts would still be unprecedented and extraordinary. It also does not much sense that with a country of 30 million people, the survey only covered an extraordinary low sample size: because the average family unit comprises five people, the survey covered 64,450 people, which is slightly more than a miserly 0.2 per cent of the country’s population.
Surely, a reasonable sample size would be ten per cent? However, Correia argued that IPC is not based on large sample sizes adding that SETSAN uses the Integrated Food Security Phase Classification (IPC) methodology, “an innovative multi-partner initiative for improving food security and nutrition analysis and decision-making,“ according to the IPC website.
It also reads that “[b]y using the IPC classification and analytical approach, Governments, UN Agencies, NGOs, civil society and other relevant actors, work together to determine the severity and magnitude of acute and chronic food insecurity, and acute malnutrition situations in a country, according to internationally-recognised scientific standards”. However, Mozambican NGOs claim that they were never consulted or asked to join in the survey to lend credibility to the outcomes of the report.
Correia’s claim that the IPC methodology is not based on large samples is baffling. Zambia IPC process sampled about half of its 13.4 million rural population between July and September 2022.
@2025, Mozambique Insights. All Rights Reserved
Great initiative and interesting content. However, three things I need to point out: 1) sampling that many districts is not unprecedented, the ministry of agriculture has been doing annual surveys covering a similar number of districts (only the districts that are in insurgency-struck areas of Northern Cabo Delgado were not sampled), the IOF (household budgetary survey) has been sampling all districts of the country for quite some years. 1)the sample size does make sense if it was randomly selected using enumeration areas as clusters. Sample size is not necessarily just a figure of % of total population, it needs to be able to capture the different settings that make up the population and the division of enumeration areas is the most scientific way to do that. Demographic Health survey, IOF and other nationwide studies will cover a range of 12000-14000 households. 3) There is some confusion about the IPC, as in some cases (even in Moz sometimes) they collect census data of areas affected by disasters and shocks that will affect food security and I am assuming that they will also carry out a sampled data collection exercise.