The mainly used welfare criterion in the social network literature is Bentham´s utilitarian concept. The shortcomings of this concept are well-known. We compare the outcomes of the utilitarian concept with the Nash social welfare function. By using a Taylor approximation we deduce a formula which allows the direct comparison of both concepts. The implications of welfare considerations of important network formation models are evaluated by using the multiplicative concept. We introduce a new symmetric connection model which is related to Nash´s welfare function in the same way as the original model is related to the utilitarian function. Based on the observation that heavy tail distributions like the power law distribution and the Pareto distribution can be explained by multiplicative structures we propose to use multiplicative utility functions in social network models. Furthermore, multiplicative utility and welfare functions together exhibit favorable characteristics both in normative and positive terms. Many empirically observed social networks have structures which are better modelled by multiplicative functions. From the normative perspective, multiplicative functions might be attractive since the Nash product introduces some form of justice.
social networks welfare efficiency Nash product efficiency justice