Probability distortion with applications to investment bias

Abstract

Cognitive bias is thought to play an important role in cost and time overrun in infrastructure projects. It is common for project appraisers to make overly optimistic assessments of project assumptions which, in turn, leads to unrealistic assessments of time and cost. These sorts of appraisals are usually made using simple models that take both technical and economic factors into account. Often, there is some uncertainty in one or more model inputs and these uncertain inputs are represented with probability distributions which are propagated through the model to assess the impact on the output. Cognitive bias, however, can lead to overly optimistic assessments of input distributions, which feeds through to model output, giving an overly optimistic assessment of important indicators. In this paper, methodologies for distorting probability distributions are introduced with the intended purpose of debiasing. These distortions are then demonstrated using a simple spreadsheet model of a hypothetical district heating project.