In this experimental study, we ask subjects to allocate tokens between a private investment and a public investment. The latter investment reduces, for all anonymous members of the group, the probability of a loss. In a number of different treatments, the expected loss without public investment is maintained constant while the probability and initial wealth are varied. In some of these treatments, the participants play under incomplete information (ambiguity). The observed behavior allows us to reject the Nash equilibrium prediction of zero contribution to the public investment. The results of non-parametric statistics suggest little difference between treatments in the level and variation of voluntary contributions to the public investment. However, the results of parametric regressions using panel data methods applied to either latent or count variables suggest that in the various treatments subjects not only react differently to the occurrence of a loss but also show different behavior with respect to reciprocity and endgame, wealth and income effects.

Co-author Claudia Keser