Household deprivation in Ghana: A latent class analysis of the 2021 Population and Housing Census
Conventional multidimensional poverty measures, including the Alkire–Foster (AF) methodology, rely on predetermined thresholds and normative aggregation weights that assume households differ only in degree along a single continuum of deprivation. This study tests that assumption directly by using the 10% public-use sample of Ghana's 2021 Population and Housing Census (N = 835,697 households). Latent Class Analysis (LCA) was applied to ten deprivation indicators spanning housing quality, water and sanitation, energy and utilities, assets and household space. The class solution was selected using multiple fit and stability criteria and assessed against local-independence diagnostics. A simultaneous one-step latent class regression was then used to model class membership on socio-demographic covariates, avoiding the attenuation bias of deterministic class assignment. Three structurally distinct classes emerged: a low-deprivation class (33.0%); a large service-deprived class (46.6%) whose members occupy structurally sound dwellings yet overwhelmingly lack improved sanitation, clean cooking fuel, and refrigerator ownership; and a severely deprived class (20.4%) facing compounding deficits across all domains. Region, rurality, education, tenure arrangement, and household headship characteristics significantly predicted class membership, with rural, northern, and less-educated households at markedly elevated risk of severe deprivation. Comparison with the AF classification showed only moderate agreement (Cohen's κ = 0.560): 28.3% of the households the AF index classified as non-poor belonged to the service-deprived class, indicating that the two approaches are complementary rather than redundant. The findings support differentiated, class-specific policy targeting and more comprehensive data collection to support structural profiling alongside existing headcount indices.
Authors
- Seth Antwi Ofori (ORCID: https://orcid.org/0009-0004-7137-4568)
- Bernard Afriyie Owusu (ORCID: https://orcid.org/0000-0002-6216-4006)
- Erika R. Manu (ORCID: https://orcid.org/0009-0006-8808-0018)
- Sarah Asaah Owusu-Kwankye (ORCID: https://orcid.org/0009-0000-5974-0853)
Institutions
- University of Ghana (GH)
Publication Details
- Journal
- Social Sciences & Humanities Open
- Published
- 2026-11-18
- DOI
- https://doi.org/10.1016/j.ssaho.2026.103986
- Primary Topic
- Income, Poverty, and Inequality
- Type
- article
- Field-Weighted Citation Impact
- 0.00