Dynamic preference foundations of expected exponentially-discounted utility
在定时风险框架下为期望指数折现效用模型提供了动态偏好基础，提出了新的公理条件，并给出了五种新的偏好基础。
Abstract Expected exponentially-discounted utility (EEDU) is the standard model of choice over risk and time in economics. This paper considers the dynamic preference foundations of EEDU in the timed risks framework. We first provide dynamic preference foundations for a time-invariant expected utility representation. The new axioms for this are called foregone-risk independence and strong time invariance. This class of dynamic preferences includes EEDU as a special case. If foregone-risk independence is strengthened to a new condition called conditional consistency, then an EEDU representation results. Alternative approaches for extending exponential discounting axioms to risk are considered, resulting in five new preference foundations of EEDU.