Intrahousehold resource allocation: A principal-agent analysis with empirical evidence.

Abstract

This study examines intrahousehold resource allocation in farm households in the Gambia River Basin. I use a multiperiod principal-agent game to explore the contractual arrangements associated with resource allocation. The household head remunerates household members for effort supplied for household food production. Household members individually contract with the household head and receive personal land on which they can grow income generating crops and a subsistence allotment of food. The model generates, as a solution, a sequence of payment schedules where remuneration is a function of effort on household plots. As men and women age, the sequence of payment schedules increases. The solution to the contracting problem also predicts that the payment schedule is greater for men than for women in any period. This payment differential is a function of the institutional structure of the household--women who leave the household may not return. The irreversibility of a woman's decision to leave the household in any period leaves her strategically weak while she maintains household residence. Consequently, household heads can extract a surplus from women's remuneration. I test the implications of the model using the Gambia River Basin Studies project data, collected in 1983-1984. A payment equation is estimated using two model specifications with different measures of the dependent variable which represents payment for effort on household plots. Three econometric approaches are used to estimate the models. The explanatory power of the model is explored, and selectivity and endogeneity biases are considered. The empirical results are sensitive to model specification. There is some evidence to support the hypotheses that household members are remunerated for labor supplied for communal food production and that payment increases over time. Strong evidence consistent with the hypothesis that there is a payment differential for men and women that is a function of the household's endowment with public goods did not emerge. The results suggest an agenda for further research, including a refinement of the model specification and the development of a more comprehensive measure of household public goods.