<![CDATA[This study estimates child multidimensional poverty across Nigeria’s 774 Local Government Areas (LGAs) by integrating the 2021 Multiple Indicator Cluster Survey (MICS) with WorldPop 2020 high-resolution population density data. Using the Alkire–Foster framework, the analysis produced a national weighted Multidimensional Poverty Index (MPI) of 0.292 and a raw MPI of 0.23632, based on a poverty incidence value of H = 0.56, which implies approximately 55.7 million children in multidimensional poverty, and an average deprivation intensity of A = 0.422. To generate LGA-level estimates, this study applied the Fay–Herriot small area estimation (SAE) model by regressing MPI on the logarithm of population density, conflict indicators, and infrastructure measures. The model explained more than 80% of the variance, improving the goodness of fit from R² = 0.695. Non-linear specifications were tested but were not retained based on the Akaike Information Criterion (AIC). A logit transformation was applied to bound predictions within a plausible range of 0–0.569, thereby eliminating negative estimates. Uncertainty estimation incorporated MICS sampling variance through bootstrapping, with coefficients of variation below 15% for 90% of LGAs. The findings reveal substantial regional disparities in child multidimensional poverty, with the North West recording a zonal MPI of 0.447 compared with 0.090 in the South East. Although constrained by data limitations, the study demonstrates the utility of SAE for producing granular poverty estimates and contributes to policy-oriented poverty measurement by strengthening evidence for geographically targeted child poverty reduction in Nigeria.]]>
Copyrights © 2026
