Robust Pro-Poorest Poverty Reduction with Counting Measures: The Anonymous Case

Abstract

When measuring poverty with counting measures, it is worth inquiring into the conditions prompting poverty reduction which not only reduce the average poverty score further but also decrease deprivation inequality among the poor, thereby emphasizing improvements among the poorest of the poor. For comparisons of cross-sectional datasets of the same society in different periods of time (i.e. an anonymous assessment), the literature offers a second-order dominance condition based on reverse generalized Lorenz curves, whose fulfillment ensures that multidimensional poverty decreases along with a reduction in deprivation inequality for a broad family of inequality-sensitive counting poverty measures. However, the condition holds for a predetermined vector of weights for the poverty dimensions. In this paper we refine this second-order condition in order to obtain necessary and sufficient conditions whose fulfillment ensures that multidimensional poverty reduction is robust to a broad array of weighting vectors and inequality-sensitive poverty measures. We illustrate these methods with an application to multidimensional poverty in Peru. JEL classification: I32

On the measurement of multidimensional poverty as a policy tool

“…First order conditions are scarcely found in the literature to prove stability of rankings and to support normative decisions. Recently, second order dominance conditions have been found useful for the case of the AHR (Gallegos et al, 2015), and therefore they are still not widely used. Hasse Diagram Techniques are applied for other composite indicators but to the knowledge of the author they have not been applied to the AHR yet.…”

“…As the same individual (or household) is not followed over time only performance on the average trends of a country (or region) can be deducted (anonymous case). As a consequence, the time analysis presented in this book is not based on the analysis of transitions nor on the differentiation of chronic and transitory poverty, for which the works of Nicholas et al (2013), Bossert et al (2013), Apablaza and Yalonetzky (2013), , and Gallegos et al (2015) show interesting results.…”

“…The concept of 'pro-poor growth' has been traditionally attached to monetary poverty. Nonetheless, recently the concept of 'pro-poor' has also been extended to non-monetary measures of well-being as in the case of Gallegos et al (2015). By using a counting social poverty index (Equation 5.4) that is sensitive to the distribution of the poorest as the AHR of Alkire and Foster, Gallegos, Yalonetzky and Azpitarte derivate a pro-poor condition.…”

“…If a curve strongly dominates another, then one can safely conclude that the results are robust to different poverty cut-off as the level of H depends on the number of indicators in which a person is considered multidimensionally deprived. According to Gallegos et al (2015), reductions in poverty were pro-poor whenever the area under the curve is clearly smaller in t = 2. In the case of Ecuador there is a clear improvement in the multidimensional poverty for the three population groups (See Appendix 5.6.1).…”

“…The multidimensional index used in this empirical application is presented in Chapter 3 where population characteristics, country basic information and indicators were described for the national average. Two different techniques will be performed: a stringent and less stringent derivation of first and second order dominance conditions following Yalonetzky (2014) and Gallegos et al (2015), and a graphic analysis of partial ordering of rankings using Hasse Diagrams Techniques (HDT) following Bruggemann and Carlsen (2006) and Bruggemann and Patil (2012, Ch. 2).…”