Semi-parametric Small Area Estimation Based Prediction Methods to Track Poverty: Validation and Applications

• Activity Number: 213
• Type: Invited
• Date/Time: Monday, July 30, 2012 : 2:00 PM to 3:50 PM
• Sponsor: Section on Survey Research Methods
• Author(s): Roy van der Weide*+ and Chris Elbers
• Companies: The World Bank and Vrije Universiteit Amsterdam
• Keywords: Small Area Statistics ; Poverty ; Inequality ; Survey Methods ; Semi-Parametric Empirical Best Estimation

Abstract

A small area estimation technique for producing "Poverty Maps", developed by Elbers, Lanjouw and Lanjouw (2002, 2003), has seen application in a large number of countries. Opportunities to formally test small area estimation methods remain rare due to lack of appropriate detailed data. This paper compares a set of predicted welfare estimates based on a generalized ELL methodology against their true values, in a setting where these true values are known. The new approach accommodates Empirical Best estimation but does not make restrictive assumptions about the distribution functions of the errors. We will present a unique empirical application using population census data from Denmark and/or the state of Minas Gerais, Brazil, which include unit record income data.