A comment on multidimensional poverty indices

Abstract

Listen

Translate article

The three papers on multidimensional poverty published in the 2011 June issue of the Journal of Economic Inequality [1–3] present strong cases in favor of their respective methodologies to capture and measure well-being and poverty. Ultimately the choice of preferred indicators, whether unidimensional or multidimensional (and, in the latter case, which dimensions to incorporate in such indicators), depends ultimately on the purposes they should serve. Money-metric (income or consumption based) poverty measures have been useful to the World Bank in estimating approximately the magnitude of global poverty as well as individual regions’ and countries’ progress towards the reduction of poverty. The Human Development Index has provided the development community with a broader measure of well-being—adding health and educational dimensions to income—and demonstrating that the correlation between the monetary and non-monetary dimensions of poverty within and across countries was far from perfect. Finally, multidimensional indicators such as the Alkire-Foster (A-F) measure respond to an effective demand for scalar estimates of poverty that could be used by policymakers in the allocation of funds to reduce poverty in an efficient and equitable way. Hence, notwithstanding their shortcomings, each of these indicators fulfills somewhat different yet important functions. A strong feature of the A-F methodology is that it is anchored on the joint distribution of the various dimensions of well-being and the resulting deprivations that poor people experience. It is therefore essential that all data come from the same household survey. It should be noted that this means that any given household information is available not only on possible deprivations in some dimensions but also on possible surpluses over and above the deprivation thresholds in other dimensions. This immediately raises the issue of possible trade-offs among dimensions.