The commonly observed difference between household income and expenditure appears increasingly important in the poverty reduction debate. Previous studies that explored poverty reduction using household survey data have struggled with significant discrepancies between household income and expenditure but they have left them at most only arbitrarily reconciled. However, because this difference is more significant for the poorest households, ignoring it has placed a significant strain on the credibility of the poverty results without fully understanding the differences between income and expenditure of the poorest households. To address the problem of the difference between household income and expenditure, this paper proposes explaining this difference as household saving determined within the household demand system. It justifies this inclusion of savings in households demand structures based on empirical and theoretical grounds. Following the work of Hertel et al. (2004), this paper performs a multilateral trade liberalization in an adjusted GTAP model to explore the implications of such a treatment of savings on the poverty results. The paper finds that household saving greatly affects the very poorest households that rely on selling their own assets as a coping strategy. Even though their ability to dissave, in general, acts positively on the poorest households that are able to afford the basic consumption, a falling price of capital goods greatly reduces their wellbeing through placing more strain on their asset reduction. The paper also finds that the price of capital goods has opposing effects on the poor and the rich, thus potentially increasing the size of the poverty gap when rising, and decreasing it when falling.