Revealed reasoning

• 成果类型: Article
• 署名作者: Saponara, Nick
• 刊物名称: JOURNAL OF ECONOMIC THEORY
• ISSN/ISSBN: 0022-0531
• DOI: 10.1016/j.jet.2020.105096
• 发表日期: 2022
• 关键词: Limited reasoning; complexity; coarse contingencies; ambiguity

摘要

We model a decision maker (DM) making a choice under uncertainty who faces constraints on the set of contingencies she can consider, or reason about, when evaluating a given act (a mapping from states to outcomes). We model the DM's ability to reason about (coarse) contingencies when evaluating a given act using partitions of the state space. A key feature of the model is that the partition may vary with the act the DM is evaluating. This feature allows us to model a DM who, despite her inability to consider every individual state simultaneously, may completely understand the states of the world (in the sense that she may have a well defined prior on the full state space). We argue that this feature allows us to differentiate coarse reasoning from coarse contingencies (as it is typically modeled) and ambiguity aversion, both related phenomena, using only static choice of acts. Our main results axiomatically characterize this model, show how to uniquely identify beliefs in this setting, and uniquely identify the partitions that represent how the DM is able to reason. An application to reasoning about other agents' actions in a mechanism design setting is studied. (C) 2020 Elsevier Inc. All rights reserved.