A new discussion paper by Harounan Kazianga and Zaki Wahhaj, KDPE 1601, January 2016
Non-technical summary
The question as to how resources are allocated within households has long been of interest to economists. Particularly in societies where state support and market institutions are weak, the household remains an important unit of production, and investment in the human capital of children.
The two theories of intra-household allocation that have received the most attention in the academic literature and tested most frequently using household data are the Unitary Model and the Collective Model. The Unitary Model, which postulates that the household behaves as if it were a single individual, has been consistently rejected by empirical evidence.
On the other hand, tests of the Collective Model, which postulates that household members with conflicting preferences are able to achieve Pareto efficient outcomes, have yielded mixed results — not rejected for labour supply decisions in developed countries or consumption decisions in developing countries, but commonly rejected for production in African households.
In this paper we investigate intra-household resource allocation in rural Burkina Faso and show that, within the same geographic, economic and social environment nuclear-family households achieve near Pareto efficiency in allocating productive resources and Pareto efficient allocation of consumption, while extended-family households do not.
We propose a theory where household members with closer familial ties exhibit higher levels of altruism towards each other, which in turn motivate them to make intra-household transfers close to that required for efficiency. The theory predicts that household members who share a nuclear family tie (as opposed to an extended-family tie or no family ties) should contribute higher levels of labour on each other’s individually owned farm plots, and this is confirmed by data on agricultural labour contributions within the household.
We argue that the frequent presence of extended-family members and unrelated individuals within the household reflects a household’s response to the absence of markets for labour exchange and risk-sharing: the additional household members provide extra labour (in exchange for room and board and use of household land for farming) and serve as a means of income diversification. Consistent with this argument, we find that household heads with more inherited land and exposed to greater income volatility (due to local rainfall conditions and the characteristics of their inherited land) are more likely to end up with extended-family households.
The wider implication of the analysis is that as markets develop and agricultural land scarcity increases, extended-family households should give way to nuclear family households. This should result in more efficient allocation of resources for production and consumption because of the ties that bind together members of the nuclear family.