Poverty mapping with aggregate census data: what is the loss in precision?

Minot, Nicholas; Baulch, Bob. 2003. Poverty mapping with aggregate census data: what is the loss in precision? In Proceedings of WIDER Conference on Spatial Inequality in Asia.

Spatially disaggregated maps of the incidence of poverty can be constructed by combining household survey data and census data. In some countries (notably China and India), however, national statistics agencies are reluctant, for reasons of confidentiality, to release household-level census data to outside researchers. This paper examines the loss in precision associated with using the type of aggregated census data, such as village- or district-level means of the data, that is usually publicly available. We show analytically that using aggregated census data will result in poverty rates that are biased downward (upward) if the rate is below (above) 50 percent and that the bias approaches zero as the poverty rate approaches zero, 50 percent, and 100 percent. Using data from Vietnam, we find that the average absolute error in estimating provincial poverty rates is about 2 percentage points if the data are aggregated to the enumeration-area level and around 3-4 percentage points if they are aggregated to the provincial level. Even census data aggregated to the provincial level perform reasonably well in ranking the 61 provinces by the incidence of poverty: the average absolute error in ranking is 0.92. -- Authors' Abstract