There is some controversy in the field of household economics regarding the efficiency of household decisions. We make the point that a flexible specification of spousal preferences and the household production technology precludes the possibility of using revealed preference data on household time allocations to determine the manner in which spouses interact: efficiently or inefficiently. Under strong, but standard, assumptions regarding marriage market equilibria, marital sorting patterns can be used essentially as “out of sample” information that allows us to assess whether household behavior is efficient or not. We develop a new likelihood-based metric to compare marriage market fits under the two alternative behavioral assumptions. We use a sample of households drawn from a recent wave of the Panel Study of Income Dynamics, and find strong evidence supporting the view that household behavior is (constrained) efficient.