English
Valencià
Forward Induction (FI) is a game theoretic hypothesis that assumes individuals infer future intentions of others from past actions in strategic interactions. In game theory, two primary justifications exist for the concept of a forward induction equilibrium. One perspective regards it as a refinement of the Nash equilibrium concept (Kohlberg and Mertens, 1986; Van Damme, 1989), while the other adopts an epistemic approach (Battigalli and Siniscalchi, 2002 & 2003) in which FI is interpreted as a set of assumptions governing players’ belief revision processes. However, both these justifications begin with the unrealistic assumption that all players are rational and believe that all others are also rational. Our objective is to provide a cognitive process model of forward induction in which rationality and belief revision serve as the foundational determinants of FI signaling. By integrating data on choices, beliefs, and eye movements, we demonstrate that «Rationality,» defined as the ability to optimally respond to the expected actions of counterparts, and «Belief Update,» defined as the capacity to revise beliefs about the expected behavior of counterparts when signaling is possible, constitute necessary and sufficient conditions for the application of FI. We conducted an eye-tracking experiment with 160 participants and found that only a fraction of them (approximately 1/3 of the participants) meet the criteria of «Rationality» and «Belief update» and exhibit a visual analysis pattern consistent with FI signaling. In line with our hypothesis, these participants consistently converge to the FI equilibrium. Conversely, the remaining participants fail to meet both of these assumptions, exhibit a visual analysis pattern inconsistent with FI signaling, and consequently, do not converge to the FI equilibrium.
Resumen
Deja un comentario