Estimates of Frisch labor-supply elasticities are biased in the presence of borrowing constraints. We show that this estimation bias is less pronounced for secondary than for primary earners. The reason is that, in households with two earners and joint borrowing constraints, wage-rate fluctuations of the secondary earner are less important for the couples’ willingness to borrow than wage-rate changes of the primary earner. We illustrate the differential estimation bias in the framework of an incomplete-markets model with two-earner households and provide empirical support using PSID data. We show that Frisch elasticities can be estimated more consistently in samples of secondary earners. Our empirical results show that Frisch elasticities are larger than often reported in microeconometric studies. Further, we show that differences in labor-supply elasticities of men and women are overestimated when borrowing constraints are ignored.