Why Is Food Consumption Inequality Underestimated? A Story of Vices and Children

Abstract

Without data on individual consumption, inequality across individuals is almost invariably inferred by applying adult equivalence scales to household-level consumption data. To assess whether these household-based measures of inequality are effective, we exploit a rare opportunity in which individual food consumption data for each and all household members are available. We use the China Health and Nutrition Survey that covers roughly 6,800 households and 25,000 individuals interviewed repeatedly from 1993 to 2011. We find that standard adult-equivalent measures understate cross-sectional individual consumption inequality by 40%. The discrepancy is largely driven by the inability of these measures to account for (a) the dispersion of the consumption of "vices" among adults (i.e. alcohol, tobacco, coffee and tea) and (b) the dispersion of the consumption of food among young children, which doubles that of adults. Our results suggest caution in the use of adult equivalent scales to measure inequality, whose effectiveness depends on the inclusion of "vices" in the consumption basket and the presence of young children.

Keywords: Food, Inequality, Individual Data, Vices, Children

JEL Classification: D12, E21