Managing Owner’s Risk of Contractor Default

Abstract

The objective of the study presented in this paper is to provide owners with a decision-making mechanism that will free them from automatically taking the typical “transfer the risk to a surety” option and will allow them to make intelligent and economical decisions that include retaining or avoiding the risk of contractor default. The methodology involves using artificial neural network (ANN) and a genetic algorithm (GA) training strategies to predict the risk of contractor default. Prediction rates of 75 and 88% were obtained with the ANN and GA training strategies, respectively. The model is of relevance to owners because once the likelihood of contractor default is predicted and the owner’s risk behavior is established, the owner can make a decision to retain, transfer, or avoid the risk of contractor default. It is of relevance to surety companies too as it may speed up the process of bonding and of reaching more reliable and objective bond/not bond decisions. The comparative use of the ANN and GA training strategies is of particular relevance to researchers.