Abstract

Food expenditure elasticity is a key indicator of household consumption behaviour and economic welfare, providing valuable insights into regional disparities and food security. In Nigeria, reliable estimates at a disaggregated geographical level are particularly important for designing targeted socioeconomic policies. However, direct survey-based estimates may be highly unstable and unreliable in areas with small sample sizes, motivating the use of Small Area Estimation (SAE) methods to improve their precision.

This study investigates the estimation of food expenditure elasticity across small areas in Nigeria using the Working–Leser model, which describes the relationship between food budget shares and total household expenditure. Particular attention is devoted to the statistical properties of the resulting elasticity estimates, which are expected to lie between zero and one in the empirical setting considered. This bounded nature raises important methodological challenges for conventional SAE approaches based on Gaussian assumptions, which may produce predictions outside the admissible range and fail to adequately represent the distribution of the parameter of interest.

To address these limitations, we propose a bounded-response SAE framework that accommodates the constrained nature of food expenditure elasticity while borrowing strength across areas. The proposed approach aims to improve estimation precision, particularly in areas with limited survey information, while ensuring theoretically coherent predictions.

The study highlights the importance of combining appropriate expenditure models with distributionally suitable SAE methods to obtain reliable small-area indicators for evidence-based policymaking in Nigeria.