Essays on actor models in exchange networks and social dilemmas

A central problem in economics and sociology is that of explaining or predicting phenomena at the macro-level. One possible mode of explanation or prediction of macro-level phenomena entails examining processes internal to the system that involve actors at the micro-level. In this dissertation complexities were added to the model of the actor on the micro-level in order to better predict and understand two macro-level phenomena that receive a lot of attention in social science research in general, and sociological research in particular: outcomes of social dilemmas and outcomes in exchange networks. With respect to predicting outcomes of social dilemmas, it was derived from game theory that conditions for cooperation in indefinitely repeated social dilemmas become more favorable when the actors become more risk averse. In general, only weak evidence was found in favor of this hypothesis. The generally weak relationship between risk aversion and cooperation in social dilemmas could at least partly be explained by the small reliability of methods to assess risk aversion. One method to assess risk aversion, the tradeoff method, which in theory is robust against misconceptions of probabilities by subjects, is described and tested in the dissertation. Outcomes in exchange networks have been predicted with reasonable accuracy by a large number of theories that have fundamentally different assumptions with respect to micro-level processes like actor bargaining. Only a handful of experimental studies have focused on only a few aspects of actor bargaining in exchange networks. Therefore, there is an incomplete understanding of the effect of the exchange network on actor bargaining, and of how actor bargaining leads to exchange outcomes. In this dissertation an experimental study was carried out to examine which variables determine actor bargaining behavior, and to estimate and test the effects of these variables. The results of the experiment indicate that actors base their initial offer in the bargaining on their relative degree in the network and that they learn from previous experiences. Concession rates also depended on the actor's relative power, but the effects of unspecified actor characteristics were in general stronger than structural network effects as measured by actors' relative power. No effects of exclusion or inclusion were found. The decrease of concession rates in the first half of the bargaining period could be explained by the fact that some actors made large concessions early in the bargaining process, probably because they were afraid of exclusion. Finally, in this dissertation the actor model used in exchange research in general and in network exchange research in particular was examined in a fundamental way. It is argued that the non-reduced exchange representation, in which exchange relations are the endogenous dynamic result of two actor characteristics; endowments and utilities, should be reintroduced in exchange research, because: (i) the most often employed reduced representation is an invalid representation of exchange in some well-defined situations, and (ii) the non-reduced representation is more general than the reduced representation in the sense that it can be applied to a much broader range of issues in exchange research. Examples include the evolution of exchange networks and exchange of externalities