Without data on individual consumption, inequality is invariably inferred by applying adult equivalence scales to household-level consumption data. To assess the e effectiveness of these household-based measures of inequality, we exploit a rare opportunity in which individual food consumption data for each and all household members are available in China. We find that standard adult-equivalent measures understate cross-sectional individual inequality by 40%. The discrepancy is driven by the dispersion of "vices" consumption among adults -alcohol, tobacco, coffee and tea- and food among young children, which doubles that of adults. Our results suggest caution in the use of adult-equivalent scales to measure inequality.