The model proposed here for obtaining the labor supply functions of moonlighters uses a double self-selection system to explore the husband's decision to moonlight together with his wife's decision to work. Subsequently, the labor functions are classified under two regimes depending on whether the wife works. The model is estimated based on a cross-section of 4, 448 married couples from the Survey of Income and Program Participation, Wave 2. I find that the household production time of husbands and wives are substitutes and that specific human capital deters moonlighting. Copyright 1990 by MIT Press.

Much of the literature on consumption smoothing and on risk sharing has focused on the ability of the household as a unit to protect its consumption. Little is known about the ability of individual members of the household to keep consumption smooth over time or relative to other members of the household. We use data on adult nutrition in Ethiopia to investigate whether individuals are able to smooth their consumption over time and within the household. We find that poorer households are not able to do so. Furthermore, poor southern households do not engage in complete risk sharing; women in these households bear the brunt of adverse shocks. This result implies that the collective model of household organization, which imposes Pareto efficiency on allocations, is rejected for these households. Finally, we obtain estimates of the relative Pareto weights in household allocation. We find that a wife’s relative position is better if customary laws on settlements at divorce are favorable or if she comes from a relatively wealthy background and that poor southern women have lower Pareto weights in allocation.