An investigation into food-away-from-home consumption in South Africa

Abstract

The food-away-from-home (FAFH) sector in South Africa has continued to increase in popularity. This is illustrated by the increased presence of FAFH in the diets of the country’s citizens. However, the sector in South Africa remains un-researched with regard to understanding household preferences and the composition of consumer expenditure. This study analyses the effects of income and socio-demographic variables on FAFH expenditure for South Africa. These results will be useful to the foodservice sector and policy makers in order to identify potential customers, respond to current customers’ changing demands and develop marketing and operational strategies, and address important nutrition and health consequences, respectively. Data from Income and Expenditure Surveys (IESs) of 2005/2006 and 2010/2011 of StatsSA (Statistics South Africa) were used to estimate the responsiveness of household FAFH expenditure in South Africa to income and a number of socio-demographic variables. The IESs contain a large number of households with zero FAFH expenditure observations which means that the use of ordinary least squares (OLS) would result in biased and inconsistent results. Furthermore, omitting households with zero FAFH expenditure, and applying OLS reduces the sample size and consequently the efficiency of estimation. Previous studies made use of the univariate and multivariate an adjustment factor and a two-stage process where the second stage is a Generalised Method of Moments (GMM) Within-Group estimator. The majority of studies suggest that double-hurdle models are appropriate for applications where zero expenditure observations are due to abstention or economic factors. The double-hurdle model is more flexible than the tobit model because it allows for the possibility that zero and positive values are generated by different mechanisms. The model used assumes independence between the two hurdles. The first hurdle determines the probability of purchasing FAFH, while the second hurdle determines the amount spent on FAFH. The double-hurdle models estimated for the IESs of 2005/2006 and 2010/2011 illustrate that households headed by younger White females with a small household size and living in an urban settlement are most likely to purchase FAFH. However, households headed by younger White males with a small household size and living in an urban formal settlement are likely to have the highest expenditure on FAFH. An increase in income positively affects the decision to buy FAFH and the amount spent by participating households. The APE (average partial effect) was calculated for the income variable. The APE determines the probability of purchasing FAFH and the income elasticities (conditional and unconditional) of expenditure on FAFH by households. The estimated conditional income elasticity of expenditure is 0,27 and the unconditional income elasticity of expenditure is 0,611 for the IES of 2005/2006. While the estimated conditional income elasticity is 0,171 and the unconditional income elasticity is 0,472 for the IES of 2010/2011. The probability of purchasing FAFH is 0,0905 and 0,0568 for the IESs of 2005/2006 and 2010/2011 respectively. The income elasticity of expenditure on FAFH is inelastic and FAFH is a normal good for the average South African household. The small size of the participation elasticities mean that growth in the FAFH sector will be driven by households with existing expenditure. Future studies should focus on per capita FAFH expenditure, the effect of the lifestage of the individual, rather than age, on FAFH expenditure, FAFH expenditure for different meals (breakfast, lunch and dinner) and facility types (quick- and full-service restaurants) and the effect of income and socio-demographic factors on FAFH expenditure on different food types (for example beef, chicken, lamb, potatoes and salads).