We formulate a static model of couples’ time-allocation decisions that features intra-household and inter-household heterogeneity. We estimate the model on German time-use data using Bayesian methods. We then use the estimated model as a lab to compare the cross-sectional estimates of individual Marshallian labor supply elasticities to the ones when wages and/or preferences are aggregated. We document a positive aggregation bias, i.e. elasticities are overstated in case of aggregation. The aggregation bias is moderate when both partners of a couple experience a wage increase. If only one of the partners experiences a wage increase, the aggregation bias is substantial both with respect to own- and cross-wage elasticities of the partners.