Improving Population and Poverty Estimates with Citizen Surveys: Evidence from East Africa

The paper sets out to explore the possibility that citizen-led surveys provide a better coverage of populations and specifically of hard-to-reach poorer areas than the international standardized household surveys which are the basis for many of the estimates used in assessing progress toward meeting the MDGs and will be for the SDGs. This hypothesis is based on the argument that, the local volunteer enumerators of citizen-led surveys are likely to be more sensitive to the specificities of local population distribution and (recent) changes than those centrally trained; and may be more effective at reaching hard-to reach groups such as those nomadic groups and those in urban slums. To test the hypothesis, the results of UWEZO (meaning "capability") surveys have been compared at a regional level with those of contemporaneous DHS surveys in Kenya, Tanzania, and Uganda for estimates of access to water and electricity. Overall, at national level, we find that UWEZO estimates for access to clean water were lower at a statistically significant level than those of DHS and these differences were statistically significant at the 0.01% level; in particular, the DHS values were much higher in regions where there are high concentrations of nomads and of urban slums, implying that the UWEZO surveys "catch" more poverty. The suggestion therefore is that citizen surveys such as UWEZO provide better, and more accurate, coverage of the poorest of the poor. Using the lowest estimate of the percentage "missing" in urban slums and extrapolating to all developing countries, there are an estimated 369 million missing from the sampling frames of standardized household surveys worldwide. This has important implications for the "Leave no one behind" appeal of the UN Secretary General and for the UN's "Data Revolution". Some suggestions are made about how to progress improved population estimates. (C) 2017 Elsevier Ltd. All rights reserved.