Are Universal Preferences Possible? Calibration Results for Non-Expected Utility Theories

经济学和管理科学的标准模型将组织的原子视为单位（个人、家庭、公司），这些单位的特征是他们的禀赋（财富、技能、知识）和他们的偏好。当从彩票菜单中选择时，决策者会根据这些偏好选择最好的彩票。最近，Rabin发现，在预期效用理论中，相对于小彩票的中等程度的风险厌恶（例如，拒绝赢得110美元或失去100美元的均等机会）导致相对于大彩票的荒谬程度的风险厌恶（例如，拒绝赢得1200万美元或失去2000美元的均等机会）。对拉宾研究结果的一种可能解释是，这是反对预期效用理论的又一论据。所提出的研究的目的是表明，类似的论点适用于所有已知的风险下决策模型（即已知概率的彩票），在某种程度上，甚至适用于不确定性下的决策模型（其中概率未知）。这些结果将对过去25年发展起来的模型提出理论上的挑战，这些模型的标准假设是，每个决策者都有一个普遍的偏好关系，而不是彩票。这项研究提出的一个可能的含义是，每个经济主体都应该有一组这样的偏好，至少在风险和不确定性下做出决策时是这样。尽管这一观点在心理学上广为人知，但尚未被经济学家完全接受。提出的研究将为这种直观的想法提供数学支持。

Standard models in Economics and Managerial Sciences consider atoms of organizations as units (individuals, families, firms) who are characterized by their endowments (wealth, skills, knowledge) and by their preferences. When choosing from a menu of lotteries, a decision maker will choose the best lottery according to these preferences. Recently, Rabin discovered that within expected utility theory, a moderate level of risk aversion with respect to small lotteries (e.g., a rejection of an even chance to win $110 or lose $100) results in absurd degree of risk aversion with respect to large lotteries (for example, rejection of an even chance to win $12 million or lose $2,000). A possible interpretation of Rabin's results is that it is yet another argument against expected utility theory. The aim of the proposed research is to show that similar arguments hold against all known models of decision making under risk (that is, lotteries with known probabilities), and to a certain extent, even against models of decision making under uncertainty (where probabilities are unknown). Such results will present a theoretical challenge to the models that were developed in the last 25 years, where the standard assumption is that each decision maker is possessed with one, universal preference relation over lotteries. One possible implication of the proposed research is that each economic agent should be characterized by a set of such preferences, at least when making decision under risk and uncertainty. Although this idea is well known in psychology, it is not yet fully acceptable by economists. The proposed research will give a mathematical support for this intuitive idea.