Measuring Changes in Poverty: A Methodological Case Study of Indonesia during an Adjustment Period

Abstract

Analysis of the effects of policy changes on the poor is often hindered by the difficulties inherent in measuring poverty and comparing levels of poverty before and after policy changes. This article outlines two techniques which can overcome many of these measurement problems: stochastic dominance conditions, which can facilitate a robust poverty ranking of distributions of living standards; and a decomposable poverty index, which allows measured changes in aggregate poverty to be disaggregated into their various components, such as the changes among population subgroups, and growth and redistributive components. These techniques can be applied to a wide range of indicators of economic well-being and poverty lines, and to assumptions about the poor. The approaches are illustrated using household survey data from Indonesia before and after external shocks and the subsequent structural adjustment program in the mid-1980s. The study finds that favorable initial conditions and a pro-poor pattern of growth enabled Indonesia to maintain its momentum in poverty alleviation during the period.