Estimating Engel Curve for Food in Ghana (1991-2013)

Abstract

The study examines the nature of food Engel curve using total expenditure as a proxy for income. The Working-leser and the quadratic specifications of Engel’s model are employed and modified to include certain demographic characteristics of households. The study uses secondary data from the last four rounds of the Ghana Living Standard Survey. Results from the study show a significant relationship between budget share of food and total expenditure. The budget share of food also varies across the country. The estimates for the last four rounds of the Ghana Living Standard Survey shows that households living in rural areas have larger budget share of food than those in urban areas. Using income as instrument in correcting the problem of endogeneity, in addition to other instruments, the estimates from the augmented regressions suggest that income is significant in examining the budget share of food with respect to total expenditure. Using the goodness of fit, the quadratic specification has a better fit for the last four rounds of the Ghana Living Standard Survey data than the Working-leser specification. The Working-Leser model estimates give larger elasticities of food with respect to total expenditure than the quadratic model. The estimates for elasticity of food with respect to total expenditure for all four rounds were less than unity for both the Working-Leser and quadratic models.