Robustness of distribution-sensitive multidimensional poverty and the Shapley decomposition

Multidimensional poverty measurement is still developing. Recently, Datt proposed an updated version of the widely-used method from Alkire and Foster, which adds a sensitivity to the severity of deprivation. Additionally, Shorrocks proposed a new method for decomposing the multidimensional poverty index, using the Shapley value. In this paper, I apply both Datt’s and Shorrocks’s method and test for robustness. Datt’s method is theoretically better than the Alkire-Foster method, but does not seem to deliver significantly different results. Shorrocks’s method has potential, but the results are strongly dependent on the choice for a weighting scheme, on which there is still little reliable data or scientific research available. At this point, decomposition does therefore not offer any substantial benefit.