Econ Papers Daily / 期刊论文
随机满意模型：特征刻画与（一般）识别
Random satisficing model: Characterization and (generic) identification
摘要
Satisficing choice is a well-known decision-making rule. However, as the literature shows, its predictive power is limited because identifying the underlying primitives is difficult. We consider a random satisficing model (RSM) where the decision maker is endowed with a preference, a random threshold, and a random search order. We characterize RSM with two axioms that use the concept of “virtual satisficing sets” introduced in the paper. We also show that the underlying preference can be generically identified when the distribution of search orders has full support and is known to the analyst. Further, we consider a possible generalization of the result when the analyst does not have full knowledge of the distribution of search orders, and when the stochastic search orders do not necessarily have full support.
中文摘要
满意化选择是一种著名的决策规则。然而，正如文献所示，其预测能力有限，因为识别其潜在基本要素较为困难。我们考虑一个随机满意化模型（RSM），其中决策者被赋予一个偏好、一个随机阈值和一个随机搜索顺序。我们利用本文引入的“虚拟满意化集”概念，以两条公理刻画了RSM。我们还表明，当搜索顺序的分布具有完全支撑且为分析者所知时，其潜在偏好可以被一般性识别。进一步，我们考虑了当分析者并不完全知晓搜索顺序的分布，以及当随机搜索顺序不一定具有完全支撑时，该结果的一种可能推广。