Trinity College Dublin

Skip to main content.

Top Level TCD Links

A Food Demand System Estimation for Uganda

Ole Boysen

IIIS Discussion Paper No. 396


This article estimates a household demand system for Uganda from cross-sectional household survey data. More specifically, a 13 item two-stage demand system model is estimated for rural and urban households separately where the main second-stage is represented by a Quadratic Almost Ideal Demand System which accounts for socio-demographic household characteristics and censoring and focuses on food items. Elasticities are calculated for three household expenditure groups as well as for the aggregate. We find that food expenditures tend to be more elastic for poorer households than for richer ones. All foods are generally price inelastic and price elasticities tend to decrease with rising expenditure level. A number of substitutional and complementary relationships between food items are identified. Simulating and analyzing the consequences of different policies in developing countries using formal models requires parameter and elasticity estimates of demand. However, little research has been published that provides such estimates, in particular for Sub-Saharan Africa. The present paper adds to the stock of such estimates and enables demand response analyses for Uganda. Furthermore, it describes in detail the process of dealing with several developing country-specific data and estimation issues which frequently occur in studies relying on cross-sectional household survey data.

Last updated 28 August 2014 by IIIS (Email).