From Unidimensional to Multidimensional Measurement of Poverty in Trinidad and Tobago: The Latent Class Analysis of Poverty Measurement as an Alternative to the Financial Deprivation Model

Abstract

Nonmonetary indicators of poverty are increasingly being used by themselves or in conjunction with monetary indicators to measure poverty levels as previous constructs based solely on monetary indicators are being deconstructed. This is predicated on the notion that low income by itself is insufficient to measure poverty. Despite this general move away from using financial indices as a sole measure of poverty, Trinidad and Tobago continues to use this approach to distinguish the poor from the non‐poor. The current effort examines the unidimensional measurement of poverty in Trinidad and Tobago and proposes an alternative measure, the latent class analysis (LCA) model, to measure poverty on the island. The authors propose the usage of a latent class model (a multidimensional construct) to analyze the nonmonetary indicators associated with two dimensions of poverty—the housing domain and the consumption domain.