Abstract

The inconsistency of predictions from solution concepts of conventional game theory with experimental observations is an enduring question. These solution concepts are based on the canonical rationality assumption that people are exclusively self-regarding utility maximizers. In this article, we think this assumption is problematic and, instead, assume that rational economic agents act as if they were maximizing their implicit utilities, which turns out to be a natural extension of the canonical rationality assumption. Implicit utility is defined by a player's character to reflect his personal weighting between cooperative, individualistic, and competitive social value orientations. The player who actually faces an implicit game chooses his strategy based on the common belief about the character distribution for a general player and the self-estimation of his own character, and he is not concerned about which strategies other players will choose and will never feel regret about his decision. It is shown by solving five paradigmatic games, the Dictator game, the Ultimatum game, the Prisoner's Dilemma game, the Public Goods game, and the Battle of the Sexes game, that the framework of implicit game and its corresponding solution concept, implicit equilibrium, based on this alternative assumption have potential for better explaining people's actual behaviors in social decision making situations.
********************************************
Link:
http://www.ncbi.nlm.nih.gov/pubmed/21573055

Recap:
Early game theorists viewed 'rational' people as "exclusively self-regarding utility maximizers." This is actually correct in-so-far as it goes: rational people are as such. But you cannot fail to integrate -- like all early game theorists have failed to integrate -- that being an "exclusively self-regarding utility maximizer" doesn't preclude, but alternatively incorporates, character building (being a "self-made soul"). In order to draw good conclusions, you have got to start with good first principles -- you have got to start by viewing man through an Aristotelian (or a Randian) 'lense.' Only then will our history of empirical results make any sense.

Early game theorists viewed 'rational' people through the wrong 'lense', such as that afforded by Hobbes or Machiavelli. This early error on their part prevented them from making sense of the human cooperation found in game theory research, perhaps most notably in the Prisoner's Dilemma game.

Ed

Abstract

Cooperation is ubiquitous in the world surrounding us, from bacteria to Human interactions. In Humans, cooperation is often associated with various group decisions, resulting from their complex web of interrelated interests, associations or preferences. The existence of such social structures not only opens the opportunity of having diverse behaviors depending on the individuals' social position, but also for a dynamical allocation of contributions depending on the returns obtained from each group. Here, we address these issues by studying the evolution of cooperation under Public Goods Games in the framework of Evolutionary Game Theory where cooperative players are able to distribute their donations to their liking. As a result, cooperation is greatly enhanced when the community structure is described by homogeneous graphs, as cooperators become able to support cooperative groups and retaliate against those with poor achievements by withdrawing donations from them. Whenever the underlying network becomes complex enough to add diversity to the distribution of group sizes, directed investments do not optimize the emergence of cooperation, but they do enhance its robustness against the invasion of a minority of free-riders. We define a robustness index and show that directed investments expand the robustness of cooperation by about 50%.