On the Value of Information and the Measurement of Risk

摘要

This paper is based on the observation that the usually used expected monetary value in estimating the value of information and the measurement of risk sometimes is unreliable, and sometimes even misdirects a decision, specifically in the cases where the expectation is infinite but the probability of success is very small. In this paper we first introduce the definition of confidence expected monetary value with different methods: confident function methods, computer simulating methods and confidence interval methods. And then, Utilizing the proposed confidence expected monetary value we further discuss the corresponding confidence value of information and confidence risk premium to improve on the reliability of their original estimations. Finally, we analysis some pical investment problems including the Morton Davis' long-term investment problem with the new tools and give some investing strategies.