Analysis of household food demand and its implications on food security in Kenya: an application of QUAIDS model

This paper evaluates household food security situation in Kenya in terms of access to food, using cross sectional data from the Kenya Integrated Household Budget Surveys (KIHBS). By estimating price and income elasticities, which provides an indication of the sensitivity of households to market shocks and thus the degree of household’s constraint to access food. The empirical approach involve estimation of demand system analysis the QUAIDS model. Existing research focuses on disaggregate food items and other developing countries, but none has specifically used the QUAIDS model for the aggregated food groups to analyse food consumption patterns nationally for the Kenyan context. The empirical results show positive expenditure elasticities while all compensated and uncompensated price elasticities show negative results. While their magnitudes vary; expenditure elasticities for meat and fish, and essential condiments are elastic (sensitive to changes) and are considered as luxuries as their elasticities are greater than one. Whereas cereal and bread, dairy products, fruit and vegetables and other condiments, have both inelastic price and expenditure elasticities; they are considered to be normal goods with values less than one. With respect to low income households, rural households and those highly dependent on the consumption of own produced food (“auto-consumption”), a reverse relationship is exhibited where meat and fish expenditure are inelastic hence perceived as normal foods, which is somewhat unexpected. However, in Kenya this finding may be attributed to the fact that a majority of the households in the survey depend on their own domesticated animals for meat and fish consumption. Hence they are not largely involved in the formal market services and prices. Further analysis shows that household size, regional differences, the ratio of food expenditure to total income and the ratio of auto-consumption are statistically significant, and hence have a great impact on food consumption expenditure. The results are broadly consistent with microeconomic theory, however exceptions indicate an unusual pattern (less sensitivity to income changes) for the rural and low income households’ meat and fish consumption. Interestingly, the low income households in our sample show that the food income elasticity for meat and fish to be less than one. These results should inform the design of policies aimed at improving the nutritional status of the poor, children and other vulnerable individuals.