Collaborative Research: Modeling Stochastic Intertemporal Preferences

该奖项资助人类决策的经济理论研究。这个项目的第一部分将解决常见的日常情况下，个人不断做出选择。在许多情况下，人们对各种商品的偏好可能每天都在变化。换句话说，他们的偏好是随机的。这项研究将提出一个新的理论模型，以帮助我们了解这些随机偏好如何影响选择行为。特别是，人们更早或更晚消费的偏好将是一个重要因素。 本项目的第二部分将研究随机偏好的福利含义。当偏好随着时间的推移而改变时，政策干预是否有益并有助于改善个人福祉尚不清楚。这项研究将提供新的工具和方法，为随机偏好的公共政策提供信息。总的来说，这个项目的结果将最终帮助研究人员为企业、政府机构和其他机构开发新的见解，这将使美国公众受益。 随机选择文献中的现有模型大多采用概率选择频率，而没有明确建模重复。第一部分将为随机选择提供一个重复选择的基础。其目的是将重复形式化，并制定一个易于处理的（即，递归）随机跨期偏好模型。现有的标准模型假设代理人的预期效用最大化，并有标准的（即添加剂的时间可分离）跨期偏好。然而，许多实验研究表明，这些标准模型的限制性太大，无法准确描述人类的真实偏好。这项研究将考虑到这一点，并允许更普遍的跨期偏好，如爱泼斯坦-津（1989）。从推理和估计的角度来看，这将是至关重要的，因为不考虑跨期结构的动态问题将导致有偏见的估计代理？喜好本项目的第二部分将研究随机选择背景下的动态一致性和平稳性。这方面的研究一直受到限制，部分原因是难以模拟这些概念的随机偏好。例如，只要偏好是随机的，平稳性到随机选择的自然延伸就总是被违反。这一点很重要，因为传统上对平稳性的破坏被视为动态不一致的实例和政策干预的理由。动态一致性和平稳性的随机选择的新的定义和表征是必要的，因为不精确的定义将导致政策干预，可能会适得其反或福利减少agentsThis奖项反映了NSF的法定使命，并已被认为是值得通过使用基金会的智力价值和更广泛的影响审查标准进行评估的支持。

This award funds research on the economic theory of human decision-making. The first part of this project will address common everyday situations where individuals make choices repeatedly over time. In many of these cases, people's preferences over various goods may change from day to day. In other words, their preferences are stochastic. This research will propose a new theoretical model to help us understand how these stochastic preferences affect choice behavior. In particular, people's preferences to consume earlier or later would be an important factor. The second part of this project will study the welfare implications of stochastic preferences. When preferences change over time, it is unclear whether policy interventions would be beneficial and help improve individual well-being. This research will provide new tools and methodologies that will inform public policy for stochastic preferences. In general, results from this project will ultimately aid researchers in developing new insights for businesses, government agencies and other institutions that will benefit the U.S. public.This research will study the dynamics of stochastic intertemporal preferences. Existing models in the stochastic choice literature have mostly taken probabilistic choice frequencies as given without explicitly modeling repetition. The first part of this project will provide a repeated choice foundation for stochastic choice. The aim will be to formalize repetition and develop a tractable (i.e., recursive) model of stochastic intertemporal preferences. Existing standard models assume agents are expected utility maximizers and have standard (i.e. additive time-separable) intertemporal preferences. However, many experimental studies suggest that these standard models are too restrictive to accurately describe true human preferences. This research will take this into account and allow for more general intertemporal preferences such as Epstein-Zin (1989). From the perspective of inference and estimation, this will be crucial as not taking into account the intertemporal structure of the dynamic problem will lead to biased estimates of agents? preferences. The second part of this project will study dynamic consistency and stationarity in the context of stochastic choice. Research on this has been limited partly due to the difficulty of modeling such concepts for stochastic preferences. For instance, a natural extension of stationarity to stochastic choice would always be violated whenever preferences are stochastic. This is important because traditional violations of stationarity have been viewed as instances of dynamic inconsistency and grounds for policy interventions. New definitions and characterizations of dynamic consistency and stationarity for stochastic choice are needed as imprecise definitions would lead to policy interventions that may be counter-productive or welfare-decreasing for agentsThis award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.