Advancing small area estimation

The poverty mapping methodology for estimating welfare rankings from small areas has proven to be useful in guiding allocation of government funds, regional planning, and general policy formulation. Nevertheless, poverty mapping also suffers from a series of by now well recognized shortcomings. We apply an approach based on first order dominance (FOD) to small area estimation. Five advantages to the FOD approach are highlighted. First, it can serve as a complement to, substitute for, and/or extension of the poverty mapping methodology. Second, it directly uses census data with a minimum of assumptions imposed. Third, the methodology is straightforward to implement and the concepts are intuitive. Fourth, the FOD approach is multi-dimensional allowing for a broader conception of poverty. Finally, FOD indicators can be chosen that relate directly to public expenditure priorities We apply the approach to census data from Mozambique for 1997 and 2007 and compare results with the poverty mapping methodology. We conclude that the FOD approach is well suited to small area estimation.