Identifying and Characterizing the Poorest Urban Population Using National Household Surveys in 38 Cities in Sub-Saharan Africa

Identifying and classifying poor and rich groups in cities depends on several factors. Using data from available nationally representative surveys from 38 sub-Saharan African countries, we aimed to identify, through different poverty classifications, the best classification in urban and large city contexts. Additionally, we characterized the poor and rich groups in terms of living standards and schooling. We relied on absolute and relative measures in the identification process. For absolute ones, we selected people living below the poverty line, socioeconomic deprivation status and the UN-Habitat slum definition. We used different cut-off points for relative measures based on wealth distribution: 30%, 40%, 50%, and 60%. We analyzed all these measures according to the absence of electricity, improved drinking water and sanitation facilities, the proportion of children out-of-school, and any household member aged 10 or more with less than 6 years of education. We used the sample size, the gap between the poorest and richest groups, and the observed agreement between absolute and relative measures to identify the best measure. The best classification was based on 40% of the wealth since it has good discriminatory power between groups and median observed agreement higher than 60% in all selected cities. Using this measure, the median prevalence of absence of improved sanitation facilities was 82% among the poorer, and this indicator presented the highest inequalities. Educational indicators presented the lower prevalence and inequalities. Luanda, Ouagadougou, and N'Djaména were considered the worst performers, while Lagos, Douala, and Nairobi were the best performers. The higher the human development index, the lower the observed inequalities. When analyzing cities using nationally representative surveys, we recommend using the relative measure of 40% of wealth to characterize the poorest group. This classification presented large gaps in the selected outcomes and good agreement with absolute measures.