Estimating Poverty in the Absence of Consumption Data : The Case of Liberia

Dabalen, Andrew ; Graham, Errol ; Himelein, Kristen ; Mungai, Rose

World Bank Group, Washington, DC

关键词: ALLOCATION OF RESOURCES;    CALCULATION;    CELL PHONE;    CELL PHONES;    CELLPHONE;   

DOI  :  10.1596/1813-9450-7024
RP-ID  :  WPS7024

学科分类：社会科学、人文和艺术（综合）

来源: World Bank Open Knowledge Repository

【 摘 要 】

In much of the developing world, thedemand for high frequency quality household data for povertymonitoring and program design far outstrips the capacity ofthe statistics bureau to provide such data. In theseenvironments, all available data sources must be leveraged.Most surveys, however, do not collect the detailedconsumption data necessary to construct aggregates andpoverty lines to measure poverty directly. This paperbenefits from a shared listing exercise for two large-scalenational household surveys conducted in Liberia in 2007 toexplore alternative methodologies to estimate povertyindirectly. The first is an asset-based model that iscommonly used in Demographic and Health Surveys. The secondis a survey-to-survey imputation that makes use of smallarea estimation techniques. In addition to a standard basemodel, separate models are estimated for urban and ruralareas and an expanded model that includes climaticvariables. Special attention is paid to the inclusion ofcell phones, with implications for other assets whose costand availability may be changing rapidly. The resultsdemonstrate substantial limitations with asset-basedindexes, but also leave questions as to the accuracy andstability of imputation models.