This paper develops a formal model of exchange network stability that combines expected value theory (Friedkin 1995) with the economic literature on network dynamics. We identify stable networks up to size 8 for varying costs and investigate whether they are Pareto efficient and egalitarian. Only a very small number of networks are stable. Odd cycles and networks consisting of dyads and at most one isolate are the only egalitarian, efficient, and stable networks for a large cost range. We show that some of these results are generalizable to networks of any size and are independent of using expected value theory.