Decomposing Multidimensional Poverty Dynamics
MetadataShow full item record
A growing interest in multidimensional poverty measures among academics and policymakers has been patent in recent years. Yet the literature has focused on cross-sectional evidence. This paper proposes a novel decomposition of changes in multidimensional poverty, as measured by the basic members of the Alkire-Foster family of measures. The method works for any type of dataset; and, in the case of panel datasets, it is useful for relating changes in these Alkire-Foster measures to transitions into and out of multidimensional poverty. The decomposition techniques are illustrated with the Young Lives panel dataset comprising cohorts of children from Ethiopia, Andhra Pradesh, Peru and Vietnam. Changes in the adjusted headcount ratio are decomposed into changes in the multidimensional headcount and changes in the average number of deprivations among poor people. Each of the latter is further decomposed into changes in relevant statistics including the probabilities of moving into and out of multidimensional poverty. This paper is one initial attempt to build a bridge between the literatures of poverty dynamics and multidimensional poverty measures. These have developed substantially but separately, for a long time. The underlying motivation is a question whether it is possible to analyse multidimensional poverty dynamics in a way that is conceptually meaningful, empirically informative, and useful for policy decisions. While we believe that much more work needs to be done in this direction, we hope that this paper provides some ideas and examples of what could be accomplished.