Network formation with heterogeneous agents and absolute friction

We present a model of endogenous network formation with absolute friction and heterogeneous agents. The individual payoffs from a given network are determined by the difference of an agent specific utility function that depends on the number of his/her direct links and the sum of his/her link-costs. These link-costs decompose in a symmetric function that represents the social and geographical distance between the two agents and an agent specific function representing the partner’s wealth and status. From a theoretical point of view, we define a new stability concept that is situated between the notions of pairwise stability (see Jackson and Wolinsky (1996)) and strong stability (see Dutta and Mutuswami (1997)). We show that our model has a unique stable network and we demonstrate that it is also strongly stable. As such, we provide uniqueness and existence for a whole range of stability concepts situated between our new stability concept and strong stability. From a practical point of view, we provide an algorithm that reproduces this stable network from information on the individual payoff structure. We illustrate the use of this algorithm by applying it to an informal insurance data set from the village of Nyakatoke in rural Tanzania.