Equilibrium Selection in Repeated Business-to-Business Matching Markets

A two-sided matching framework is applied to repeated business-to- business procurement matches. Both static and dynamic solutions concepts--- namely Gale-Shapley deferred acceptance algorithm, learning dynamics, and genetic algorithms--- are used to obtain solutions. The settings under investigation include both full information and limited information settings. We show that under certain conditions the dynamic predictions refine the core of the matching market. According to the theoretical predictions, organizational buyers would be better off in buyer-proposing settings than in seller-proposing settings, whereas sellers would prefer the opposite. The salience of this theoretical prediction is tested and confirmed using experimental data under different environments.