Food expenditure patterns in South Africa: Evidence from the NIDS

Abstract

This study evaluates food expenditure patterns in South Africa using the Engel framework which states that proportions spent on food fall with income. Non-parametric methods are used to estimate Engel curves, and regression analysis to evaluate the effects of several variables on shares of total expenditure on food using the Working-Leser model. Pooled OLS is used to compare the exposure and sensitivity to changing expenditure capacity between waves. We find that households were spending proportionally less on food in 2008 compared to 2010 and 2012 and that food is the most important item of expenditure in most households by looking at budget shares. The sensitivity of the share of total expenditure dedicated to food varies with expenditure capacity. The effect of food price inflation on all households in South Africa is conjectured to contribute markedly to this trend, though we cannot confirm that hypothesis with NIDS data alone. The implications for food security is fertile ground for further research on this issue.