Abstract: We design an original laboratory experiment to investigate whether redistributive actions hinder the formation of Pareto-improving groups. We test, in an anonymous setting with no feedback, whether people choose to destroy or steal the endowment of others and whether they choose to give to others, when granted the option. We then test whether subjects join a group that increases their endowment but exposes them to redistribution. We conduct the experiment in three very different settings with a priori different norms of pro-social behavior: a university town in the UK, the largest slum in Nairobi, Kenya and rural Uganda. We find a lot of commonality but also large differences between sites. UK subjects behave in a more selfish and strategic way ñ giving less, stealing more. Kenyan and Ugandan subjects behave in a more altruistic and less strategic manner. However, pro-social norms are not always predictive of joining behavior. African subjects are less likely to join a group when destruction or stealing is permitted. It is as if they are less trusting even though they are more trustworthy. These findings contradict the view that African current underdevelopment is due to a failure of generalized morality.
Status: Working Paper