This paper addresses two questio~s central to the literature on intra-household resource allocations: does the allocation of' of household educational resources in Pakistan favour males over
females? And, what explains the inability of the standard Engel curve approach to detect differential treatment even where discrimination is known to exist? Using uhique individual level
data from the latest household survey from Pakistan, we estimate Engel curves to address these questions. Empirical findings reveal significant pro-male biases in the allocation of educational
expenditure among all school-going age groups in Pakistan. We conclude that one of the main reasons for the failure of studies in identifying biases lie in using highly aggregated expenditure data in conventional estimate.