On measuring the sources of changes in poverty using the Shapley method. An application to Europe

For decomposable poverty measures in incidence, intensity and inequality among the poor, poverty changes between two periods can be expressed in terms of the three poverty components in the two periods. However, most of the poverty decompositions cannot be written in a linear form of the terms. We apply the Shapley decomposition approach in order to decompose the overall poverty change as the sum of the contributions of the three poverty components' changes. We provide a method to compute the contributions for any decomposable poverty index, and specifically, the contributions formulas for the Sen index and the Foster, Greer and Thorbecke index for α=2 are shown. Using EU-SILC data for 2008 and 2015 for 28 European Countries, we analyze the change over time in the FGT2 poverty index and the value of the marginal contributions of the three components.