Poverty Mapping in Selected Developing Member Countries
(September 2002 – March 2004)
Issues
The implementation of ADB's poverty reduction strategy requires ex-ante and
ex-post assessments of the impact of projects and programs on poverty. Information
requirements to conduct such assessments vary across types of projects and programs,
as does the availability and suitability of using existing data, and the feasibility
of collecting new information. ADB experience suggests that, while nearly all
poverty assessment exercises refer to existing statistics, often these are used
to describe poverty at a very general level using published secondary data.
A major constraint is the high level of geographic aggregation of existing data
on poverty, and the non-availability of data at project levels.
Many household surveys and censuses have been conducted in DMCs at significant
expense. In most cases however, the use of these data for operational purposes
remains limited. Restricted accessibility of micro-level data to secondary analysts,
limited research capacity in DMCs, and insufficient survey sample sizes have
often prevented available statistical data from being used for project design
and monitoring. ADB as well as governments and other donors face the problem
of lack of relevant data for proper targeting and monitoring of poverty reduction
interventions.
National surveys provide statistically valid data to measure poverty at the
national level and at limited sub-national levels. Unfortunately, such survey
data are rarely statistically representative at local level, which creates difficulties
in their use in poverty assessment for projects that are usually location-specific.
Nevertheless, a number of techniques are available to generate data valid at
local level. Poverty mapping is a methodology that enables poverty data obtained
from household sample surveys to be extrapolated to the entire population by
combining them with population censuses data (using multiple regression techniques).
An immediate practical application of poverty mapping is to spatially locate
the poor in a given country and use it to target direct anti-poverty interventions.
The insights gained from a well-constructed poverty map regarding the spatial
distribution of poverty will be a valuable input into the formulation of country
strategies and poverty partnership agreements. Such information will be useful
in assessment of the likely poverty impact of proposed project/program and in
defining funding priorities, by ADB, DMCs policy makers, and other development
stakeholders.
A significant improvement in designing, targeting and monitoring of poverty
reduction interventions could be achieved by implementing recently developed
poverty mapping techniques in selected DMCs. Poverty mapping is, in essence,
a collection of tools to characterize the spatial variability of poverty at
the local level. In its various forms, it involves techniques that permit disaggregation
of poverty and inequality measures to local administrative levels, or small
geographical units, based on existing data. The RETA proposes to implement this
technique to two DMCs on a pilot basis (Pakistan and another country to be identified).
Purpose and Outputs
The goal of the RETA is to generate reliable estimates of poverty and inequality
at disaggregated levels (geographic, or along other dimensions). Based on these
estimates, poverty maps will be produced that will help achieve the following
purposes: (i) more accurate and cost-effective targeting and monitoring of poverty-reduction
projects and programs; (ii) improve ex-ante impact assessment of proposed projects
and policies; (iii) improve poverty analysis and statistical capacity in the
participating DMCs; and (iv) foster good governance by increasing the transparency
of government resource allocation and disseminating information about the geographic
distribution of poverty to stakeholders.
Methodology and Key Activities
The RETA will apply a technique that combines income/expenditure survey data
and population census data to generate poverty and inequality profiles at low
levels of aggregation (a technique successfully applied in some Latin American
and African countries). The main components and outputs will be the following:
(i) production of training material, and training of statisticians in the selected
DMCs; (ii) for the selected countries, development of regression models of the
distribution of household expenditure based on variables (predictors) available
both in the population census and the income/expenditure survey dataset; and
(iii) implementation of poverty mapping in the selected DMCs, and preparation
of the outputs: maps, databases, and technical/analytical papers.
Contact
For more information, you may contact the TA officer:
Mr. Olivier Dupriez, Poverty Statistician
odupriez@adb.org
Link
This link will lead you outside the ADB
website. Press the BACK button to return to this page.
The World Bank, Geographic
Aspects of Inequality and Poverty
Economics and Statistics 
Top of page