This dissertation comprises three interconnected essays focusing on the measurement of poverty and mobility.

The first chapter introduces a modification to the standard Alkire-Foster (AF) Method (Alkire and Foster, 2011a,b) for multidimensional poverty measurement. This modification incorporates cardinal information during the identification step when feasible. Following the typical identification-aggregation structure, the adjusted method considers both the number of deprivations and aggregate (or average) achievement for identification and employs the adjusted FGT class of measures for aggregation. The proposed method upholds key axioms, including subgroup decomposability, replication invariance, symmetry, and poverty focus. Empirical results, utilizing data from Indonesia and South Africa, reveal that the modified method is less sensitive to changes in the poverty cutoff and exhibits an enhanced focus on individuals experiencing deeper deprivations.

In the second chapter, we review the three primary approaches for measuring intergenerational mobility and propose an oriented approach capable of separately measuring upward and downward mobility. This pioneering attempt considers multiple dimensions simultaneously, addressing both fully ordinal and fully cardinal data. For fully ordinal data, we advocate using the standard AF identification process, tracking changes in the number of deprivations across generations. For fully cardinal data, we propose utilizing the modified AF identification introduced in the first chapter to construct a comprehensive individual-based index representing overall well-being, tracked over generations. The proposed method adheres to key properties, including decomposability, replication invariance, symmetry, and properties derived from the axiom of expansion. Application of our method to data from the Panel Study of Income Dynamics (PSID) reveals less upward mobility and overall mobility from 2011 to 2021 compared to 2001 to 2011, offering valuable insights for longitudinal comparisons.

The third chapter focuses on chronic poverty measurement. After reviewing the two main approaches in the current literature for both unidimensional and multidimensional cases, we propose a method that incorporates key elements from both approaches: the number of poor periods and ‘permanent income’ or average resources over the entire timeline. The proposed method satisfies crucial properties, including decomposability, time and population symmetry, and time and population replication invariance. Demonstrating the feasibility of our method for both unidimensional and multidimensional cases, we apply our hybrid approach to data drawn from the National Longitudinal Survey (NLS). Results indicate that the hybrid method is less sensitive to changes in poverty cutoff and duration cutoff compared to the traditional counting approach, providing an improved focus on individuals experiencing deeper deprivations.