A Bayesian Hierarchical Framework for Geographic Mortality Variation and Its Insurance Applications in Nigeria

This study develops a Bayesian hierarchical modelling framework for geographic mortality variation in Nigeria and derives insurance and pension valuation applications from the estimated spatial mortality structure. Geographic mortality variation is large in Nigeria but poorly quantified due to sparse and inconsistent vital registration data. Actuarial applications including regional life table construction, spatially differentiated annuity pricing, and subnational pension liability estimation require robust mortality estimates at fine geographic resolution, which single-level estimation approaches cannot reliably provide for small geographic units with limited data. This study develops a three-level Bayesian hierarchical model of age-sex-specific mortality rates with national, geopolitical zone, state, and local government area levels, estimated using Hamiltonian Monte Carlo methods applied to data from five NDHS waves, facility-based mortality records, and the 2023 NPC census. Spatial autocorrelation is modelled using conditional autoregressive priors. Insurance applications include differential annuity pricing by zone and optimal risk pooling territory design for life insurance portfolios. Findings establish a posterior mortality surface for Nigeria with credible interval widths reflecting data density by area, showing significant mortality heterogeneity even within geopolitical zones. The posterior predictive distribution of zone-level life expectancy at birth ranges from 52.3 years for the North East to 64.7 years for the South West. Insurance application results demonstrate that zone-differentiated annuity pricing reduces adverse selection probability by 34 percent compared to national pooling. The study contributes an original spatial mortality framework for Nigeria and recommends its adoption as the national actuarial reference surface.

Keywords: Bayesian hierarchical model, geographic mortality, spatial variation, life tables, Nigeria.