Binary data, hierarchy of attributes, and multidimensional deprivation

Abstract

Empirical estimation of multidimensional deprivation measures has gained momentum in the last few years. Several existing measures assume that deprivation dimensions are cardinally measurable, when, in many instances, such data is not always available. In this paper, we propose a class of deprivation measures when the only information available is whether an individual is deprived in an attribute or not. The framework is then extended to a setting in which the multiple dimensions are

Assessing Deprivation with Ordinal Variables: Depth Sensitivity and Poverty Aversion

“…Yet one important concern has been how to conduct meaningful poverty assessment when the underlying dimensions are ordinal. This challenge has not been overlooked and various multidimensional poverty measures motivated by the counting approach (Atkinson, 2003) have been proposed (see, for instance, Chakravarty and D'Ambrosio, 2006;Alkire and Foster, 2011;Aaberge and Peluso, 2012;Bossert et al, 2013;Alkire and Foster, 2016;Dhongde et al, 2016). We have already mentioned that our approach to measure poverty with ordinal variables is motivationally and conceptually different from the counting framework, and yet their functional forms are very similar.…”

“…Even though additively decomposable measures are common in cardinal poverty measurement literature, our measures are reminiscent of the additively decomposable poverty measures within the multidimensional counting approach (Chakravarty and D'Ambrosio, 2006;Alkire and Foster, 2011;Bossert et al, 2013;Alkire and Foster, 2016;Dhongde et al, 2016), albeit with two subtle, yet crucial, differences. The first difference between our framework and the counting approach is conceptual.…”

“…See http://www.un.org/sustainabledevelopment/poverty/ (accessed in April 2017).3 We refer to the unidimensional context here. The issue of ordinality has certainly been examined thoroughly in the context of multidimensional poverty measurement(Alkire and Foster, 2011;Bossert et al, 2013;Dhongde et al, 2016;Bosmans et al, 2017). However, even in the multidimensional context, ordinal variables are dichotomised in practice, thereby ignoring the depth of deprivations within…”

Assessing Deprivation with an Ordinal Variable: Theory and Application to Sanitation Deprivation in Bangladesh

“…Although grounded on prioritarianism, our new form of poverty aversion encompasses, as limiting cases, both previous attempts at sensitising ordinal poverty indices to the depth of deprivations (e.g., Bennett and Hatzimasoura, 2011;Yalonetzky, 2012) as well as current burgeoning approaches to distributional sensitivity in ordinal frameworks based on Hammond transfers (Hammond, 1976;Gravel et al, 2015). We define a range of properties based on this new form of degree of poverty aversion and characterise the corresponding subclasses in the context of multidimensional poverty measurement (Alkire and Foster, 2011;Bossert et al, 2013;Dhongde et al, 2016;Bosmans et al, 2017). However, even in the multidimensional context, ordinal variables are often dichotomised in empirical applications (see, Alkire and Foster, 2011;Bossert et al, 2013;Dhongde et al, 2016), thereby ignoring the depth of deprivations within indicators.…”

“…We define a range of properties based on this new form of degree of poverty aversion and characterise the corresponding subclasses in the context of multidimensional poverty measurement (Alkire and Foster, 2011;Bossert et al, 2013;Dhongde et al, 2016;Bosmans et al, 2017). However, even in the multidimensional context, ordinal variables are often dichotomised in empirical applications (see, Alkire and Foster, 2011;Bossert et al, 2013;Dhongde et al, 2016), thereby ignoring the depth of deprivations within indicators.…”