Game-Theory Models for Exchange Networks: Experimental Results

Abstract

The goal of current exchange-network literature is to develop algorithms, loosely based on rational choice, that predict how resources are distributed through exchange networks and which positions have power to accumulate resources. These objectives closely resemble those of N-person cooperative games with transferable utility, which are based on formal explicit models of rational choice. Experimentally, power is exhibited when a position can amass a favorable proportion of available resources by negotiating a division with another network member. Game-theory solution concepts that address the question of power in networks are introduced and compared to network-exchange models to evaluate the effectiveness of the game-theory solutions and those of exchange theory in predicting results observed in experiments. Experimental data show that there is a utility in incorporating game theory into the discussion of exchange in negatively connected networks. Furthermore, the use of game theory leads to a more comprehensive understanding of many processes to which exchange theory is insensitive.