AGRICULTURAL & RESOURCE ECONOMICS
UNIVERSITY OF CALIFORNIA AT BERKELEY
|
THE DEVELOPMENT FIELD AGRICULTURAL & RESOURCE ECONOMICS UNIVERSITY OF CALIFORNIA AT BERKELEY |
|
 |
The research program on "Poverty Maps" refers to a recently developed technique for greatly increasing the detail available in poverty profiles. The method combines survey and census data to estimate consumption-based welfare indicators for small geographic areas such as districts and sub-districts. Apart from its obvious use in targeting expenditures for poverty reduction, there are a large number policy relevant questions that can be answered using these newly-available poverty and inequality figures in each country -- concerning the interaction between such distributional outcomes and health, crime, the environment, and social funds. There is also scope for linking "poverty maps" to studies of the distributional impact of macro policies so as to ascertain the spatial incidence of such impacts. At present there are poverty maps completed or in process in a large set of countries: Ecuador, Bolivia, Panama, Nicaragua, Guatemala, Brazil, Mexico, Zambia, Tanzania, Kenya, Malawi, Mozambique, South Africa, Uganda, Albania, Bulgaria, Kyrgyzstan, Sri Lanka, Nepal, Cambodia, Vietnam, Thailand, Laos, Indonesia, Papua New Guinea, Philippines, and China. |
There remain a number of research questions associated with the poverty mapping methodology, generally aimed at improving the precision of estimates and their reliability. The method can also be extended. First, it could also be used to combine two or more sample surveys. For example, one could combine LSMS-style household surveys with the data collected in Demographic and Health Surveys (DHS), and thus combine the welfare indicators of the LSMS with the detailed health and education indicators of the DHS. Relatedly, the method can be applied to explore important questions of comparability across non-identical sample surveys, by imputing comparable definitions of welfare into otherwise non-comparable surveys. Second, the method could be used to provide detailed breakdowns of poverty measures along dimensions other than geographic. For example, one could construct inequality and poverty estimates across criteria such as age, education, ethnicity, occupation, as well as locality, and also combinations of these. This would allow much more detailed poverty profiles than have been possible in the past. Third, where there panel survey data are available or appropriate ancillary data sets can be identified the method can be extended to produce up-dates over time.