Measurements of child poverty are commonly carried out using household expenditure data in per capita terms or adjusted by some standard parameter of economies of size. In this paper, we use adult equivalence scales and economies of scale coefficients estimated from the data to assess child poverty in Vietnam. By doing so, we show child poverty in Vietnam to be overestimated by conventional techniques. The commonly used technique for the estimation of adult equivalence scales contains an implicit household income distribution. Therefore, we used these techniques to estimate child poverty and patterns of resource allocation for different household groups. Better-off families (which are largely urban, more educated, female-headed and from the Khin ethnic majority) spend a lower share of their income on children than other household groups, though their children are still better off in absolute terms. A comparison of the data between 1992-93 and 1997-98 also reveals that children from better-off family groups have experienced a more rapid increase in welfare levels than children from other sectors of Vietnamese society during this period. The presence of gender discrimination in child treatment is also investigated. Rural and male headed households are those that more prominently appear as discriminating against female children. We found no clear relationship between gender discrimination and expenditure or between gender discrimination and level of education.