Regional perspectives to the multidimensional poverty index

The Multidimensional Poverty Index (MPI) is not the first attempt to examine poverty along multiple definitions. Vis-a-vis the existing work and other presently available measures, this method has greater advantage in terms of international comparability and reporting. However, the methodology of the Multidimensional Poverty Index (MPI) has been under strong scrutiny since its inception. One of the reasons for these critiques lies in the variation in the MPI country ranks and scores based on different indicators and a different weighing scheme. This paper analyzes the consequences of a different weighting scheme within the MPI, using a more data driven approach rather than a normative or equal weighting scheme. It attempts to assess this alternative weighting via its impact on the scores and relative ranking of various countries. Moreover, it attempts to resolve the differences in the definition of poverty that might emerge upon changing indicators, and thereby evaluate how this affects the construction of the MPI. An analysis covering 22 countries, using the Demographic and Health Survey data, is carried out to quantitatively evaluate the weights assigned to each of the indicators, using the technique of Principal Component Analysis (PCA) and Multiple Correspondence Analysis (MCA). A more detailed country-level analysis is carried out for India, wherein additional indicators based on the data are made available and therefore an alternative multidimensional construct is possible. The analysis shows that equal weighting of the three dimensions cannot be statistically justified and that in trying to capture a more multidimensional view of poverty and well-being, there might actually not be so much multidimensionality in the MPI.