An experimental investigation of the eductive abilities of the agents in negative feedback environments

The goal of this thesis is to study the eductive-type of reasoning in negative feedbacksituations. We address it through an experimental approach. We construct this thesis aroundthree questions: what mechanism is the eductive type of reasoning based on? Is this type ofreasoning more likely to succeed in some particular situations? Can we find particularconditions improving the performance of the eductive reasoning? We identify the negativefeedback environments as stabilizing situations for the eductive reasoning. Therefore, afterdefining the concept of eductive reasoning and characterizing negative feedback situations,we introduce and test in the first part of this thesis a negative feedback beauty contest game.Repetition, elicitation and circularity are the conditions of its success within a marketsituation. Therefore, in the second part of this thesis, we are interested in the application ofthis game in cobweb markets. Our thesis shows that, in negative feedback situations,reflective beliefs turn faster into intuitive beliefs, because through an eductive type ofreasoning, the equilibrium is scanned several times, and useful information is increased.Consequently, a market with a negative feedback structure is stable and the agents within thistype of market hold coordinated beliefs.