The problem with probability

Abstract

Knightian uncertainty is the circumstance in which there is not enough information available to quantitatively measure in terms of probability the uncertainty that is present. Economic decision makers usually face that kind of uncertainty when attempting to determine their best route forward into the future. But instead of confronting Knightian uncertainty head on, economists generally have assumed it away by opting instead to employ what Knight called risk as measured by probability. But there are serious questions about whether probability has the capacity to coral in a meaningful way the uncertainty many decision makers actually face, and whether its use in models of decision making provides an appropriately relevant and realistic explanation of what is going on. This paper argues that it does not by examining the meaning of the probability concept as it relates to the true uncertain environment the decision maker faces. It also suggests that the alternative measure of surprise introduced by Shackle not only returns such analyses to the more realistic realm of Knightian uncertainty, but is also a workable and suitable replacement for probability in explanations of decision making that need to account for the unpredictability of the future.