Revealed preference axioms for endogenous consideration set formation
研究了当决策者的考虑集可观察时，如何用类似显示偏好的公理来检验这些集合是否由内生形成，即决策者在注意力成本下最优选择子集，并将结果应用于Logit模型。
We consider a setting in which the consideration sets being formed by a decision maker are observable. We analyze the necessary and sufficient conditions under which the observed sets are consistent with endogenous consideration set formation. In particular, we rationalize the consideration sets as being optimally formed by a decision maker who faces costly attention and is forced to choose a subset of alternatives to pay attention to. We show that axioms similar to those from revealed preference theory allow us to do this. The most general model is characterized by a condition resembling the Strong Axiom applied on a domain of sets rather than individual alternatives. Since the idea of observable consideration sets seems realistic in a random choice framework in which we can interpret zero probability of being chosen as the alternative being omitted from the consideration set, we apply our result to this setting using the Logit model. This results in a representation theorem for a generalized version of the Logit model.