摘要：Information on the spatial distribution of poverty is particularly useful in designing geographically targeted programs to address regional disparities—a matter of high priority for many countries. Until recently, most developing countries were forced to design these programs based on rough indicators of poverty or on the results of household budget surveys, which typically generate poverty estimates for a limited number of regions. In the late 1990s, a new approach was developed combining census data and household budget survey results to generate poverty estimates for small areas such as districts, allowing the construction of “poverty maps.” A1998 IFPRI study in Vietnam was one of the first to experiment with this approach. A similar method was concurrently developed and subsequently refined by researchers at the World Bank. Poverty mapping studies have now been carried out in more than a dozen countries. This is an example of how an international public good can evolve from a set of specific country studies by the broader research community. This report uses data from the 1998 Vietnam Living Standards Survey and the 1999 Population Census to estimate various measures of poverty and inequality for 614 districts in Vietnam. The results confirm conventional wisdom regarding high rates of poverty in the upland areas, low rates in urban areas, and intermediate rates in the irrigated lowland areas. But the results also offer some surprises, such as the fact that most poor people in Vietnam do not live in the poorest areas. The study goes a step beyond standard poverty-mapping analysis by investigating the geographic determinants of district-level poverty rates. Distance to cities, soil type, and topography are significant predictors of local poverty rates. Surprisingly, proximity to small district centers is a stronger predictor of poverty than distance to large cities or distance to roads. Thus, rural-urban linkages at a local level matter a great deal for poverty reduction. At the national level, this report provides valuable information for more precise geographic targeting of poverty assistance in Vietnam, and also offers insights into the geographic factors that contribute to rural poverty. Statistical authorities in Vietnam are now experimenting with variants of this method for use in poverty monitoring. The study demonstrates the rewards that come from combining survey data, census data, and geographic data to focus on the challenge of reducing rural poverty. IFPRI continues to pursue this line of research, having recently carried out poverty-mapping analyses for Mozambique, Malawi, and Zambia.
