On the Dynamic Development of Social Ties: Theory and Application

We implement, expand with limited foresight, and test the performance of the theoretical social ties model of van Dijk and van Winden (1997). The model is estimated on various experimental data sets of repeated public good games involving different numbers of players. Our estimation results provide direct support for the proposed social ties mechanism, showing that the history of social interaction is an essential determinant of preferences. About a quarter of the subjects appear to be forward-looking. Both the within-sample and the out-of-sample predictive performance of the model turn out to be remarkably good, as it is able to track the often complex dynamic contribution patterns. This is true both for games played in pairs and in groups of four players, where the decision-making processes are much more complex. Moreover, we show that this model performs better than some other social preferences models that allow for a dynamic implementation. Our conclusion is that this simple, tractable, and psychologically grounded model of social preferences can account very well for dynamic behavioral patterns in repeated public good games.