A Test of the Household Separation Hypothesis in Rwanda

James E. Anderson, University of San Francisco


How does a farm household in rural Africa react when the government decides crop selection? In developing countries, agricultural households strive to optimize a risk mitigating utility function rather than a traditional agricultural production function. These households are termed “non-separated” as their farming efforts are directed towards family food security rather than maximizing agricultural profits. The lack of integration with labor and commodity markets makes these non-separated households difficult to influence with policy initiatives. Various tests for household separation have been developed.

We use a unique dataset from Rwanda to evaluate these separation tests. The data include households forced into a separated economic model by government policy, as well as, those partially separated and others that remained non-separated. We create a modified test for separation using a measure of the alignment of the production and consumption functions.

Last, we hypothesize that farm households forced into partial separation will alter their crop selection on their other plots to mitigate the perceived risk to family food security. We find a weak association between crop selection on these plots and the consumption profile of the household.