Gender Roles and Non-Cooperative Behavior: Field Experiment on the Effect of Asymmetric Information on Intra-Household Allocation in Ghana

This study found that unanticipated cash and in-kind transfers had differential impacts on household spending patterns, depending on the role of the spouse in the household, and whether other household members knew about these funds:

• Both men and women spent public prizes on shared household needs: men spent it on home goods, equipment and health, while women invested in farming. This was particularly true for public livestock prizes -- men used these winnings to barter for in-kind health payments or breeding opportunities, while women’s livestock winnings were utilized on the household farm.
• Prizes awarded privately to one household member were more likely to be hidden from the partner, and subsequently spent in less visible ways. While men were more inclined to spend private prize money on alcohol and entertainment, women showed a higher propensity to save the income.
• Receiving private prizes also increased the gifts or loans that men and women provided to others. Compared to those who did not win, wives were more likely to give loans or in-kind transfers to friends and relatives, while men tended to increase their amount of gifts to others. Both of these actions could be to build support networks, but the authors suggest that women may use loaning money as a way to keep the prize money away from their husbands.

In sum, this study makes a contribution to existing research on intra-household resource allocation by showing that unanticipated household income or assets might have differential effects on household spending, depending on how visible the additional funds are to the spouse.

In a year-long field experiment with a total of 70 households in 4 villages for a total of 280 households of which 16 are excluded for polygamy, study participants were interviewed privately (i.e. without the presence of their husband or wife) in five rounds, exploring their spending patterns. After an initial baseline evaluation, study participants were surveyed in four additional rounds, which took place once every two months. One week before each survey, lottery sessions and drawings were performed in each of the participating communities. Participating household members were randomly assigned to two possible treatment groups: some received public prizes drawn in a lottery during a community hall meeting, and others won assets privately by drawing numbered bottle caps from a bag. Prizes included either in-kind transfers such as livestock (chicken or goat), or cash transfers worth up to 75 Ghana cedis.