Expert elicitation and Bayesian analysis of construction contract risks: an investigation

Abstract

Formal risk analysis techniques applied in managing construction project risks tend to focus on risks that lend themselves to ‘objective’ methods of economic analysis. Although subjective probabilities and Bayesian methods are applied successfully in other industries to manage ‘subjective’ risks similar to those encountered in construction contracts, very little is reported on the application of such methods for analysing risks in construction contracts. An investigation has been carried out into using subjective expert opinions and Bayesian methods in construction contract risk analysis. The elicitation of expert opinions about the risk of encountering adverse ground conditions on construction sites is examined. Prior distributions of expert opinions about the risk form the critical component in the Bayesian analysis of risk. The conclusion of this is that although construction professionals have difficulty estimating intermediate and tail values of probability distributions, elicitation techniques can be used to develop prior distributions of contract risks which, coupled with sample information on the risks, would enable the Bayesian analysis of risks. Effective strategies for systematically analysing significant individual contract risks often covered under an ‘arbitrary’ project contingency sum can now be developed to enhance the risk management efforts of contractors and construction experts.