Socially Interdependent Investment

We report the results of an economic experiment on how social preferences and behavioral imitation lead to the social clustering of investment decisions, even if the expected return of the underlying asset is known. During the experiment, subjects are asked how much of their endowment they wish to allocate in an asset to which there is a 50% chance the amount they invest will be tripled and a 50% chance their investment will be lost. To study the social interdependence of investment decisions we use a 2 x 2 factorial design varying: (i) whether the subjects initially observed high or low investment social anchors, (ii) whether information about the investment decisions of other subjects in their social group is provided. Overall, we find strong evidence that individuals' investment decisions are malleable to the investment decisions of their peers, which in turn leads to the social clustering of investment decisions. Furthermore, social anchors shape initial investment levels, leading to mean investment rates that are double in size in the high anchor treatment relative to the low anchor treatment.