This article introduces a method for estimating structural labor supply models in the presence of unobservable wages and deviations of households' marginal revenue product of self-employed labor from their shadow wage. This method is therefore robust to a wide range of assumptions about labor allocation decisions in the presence of uncertainty, market frictions, locational preferences, etc. We illustrate the method using data from rice producers in Côte d'Ivoire. These data, like previous studies, reveal significant systematic differences between shadow wages and the marginal revenue product of family farm labor. We demonstrate how one can exploit systematic deviations, in the present case related to household characteristics such as the land/labor endowment ratio, to control for both unobservable wages and prospective allocative inefficiency in labor allocation in structural household labor supply estimation.