This paper examines the combined eects of changes in prices, income and demographic composition on adult and young, male and female members of households. The recently developed Quadratic Almost Ideal Demand System (QUAIDS) is used since a demand system provides a unified framework for analysing the combined effects in a systematic fashion. The `outlay equivalent method', which was used with single demand equations in previous studies, is married to the demand system literature. Underlying preference structures for classifying goods into different groups is also examined by conducting alternative tests of separability in preferences. Panel/longitudinal data are used helping to control for household level heterogeneity. The empirical results show that Ethiopian rural households respond to price, income and demographic changes in a more complicated manner than usually assumed; demographic groups absorbing most of the impact differ for different types of changes. Changes in household income affect male members of households (men and boys) more than female members (women and girls). On the other hand, changes in price affect women and boys more than men and girls. In addition, adjustments in household expenditure due to demographic changes imply that boys are favoured relative to girls. But the overall position of boys and girls in the household depends not only on the `outlay equivalent ratios' but also on the effects of changes in household income and prices as determined by budget and price elasticities. These findings show that households distribute risks among different demographic groups rather than only one group absorbing all shocks. The findings indicate that studies that only looked at the `outlay equivalent ratios' tell only part of the story.