Predicting construction contractor failure prior to contract award

This paper discusses items that can aid owners in predicting the chance of construction contractor failure prior to contract award and thus assist them in the evaluation process. A predictive contractor failure model has been developed and is discussed. The model predicts the probability of contractor failure at the project level; failure is defined as a significant breach of the contractor's legal reponsibilities to the owner (for example, bankruptcy or material breach of contract related to meeting desired project objectives such as cost, schedule, and quality). Data for the model development were collected using two questionnare surveys. The modeling method involves the use of discrete choice logistic regression. Results show that four variables are strong predictors of contractor failure: (1) The amount of owner-contractor evaluation; (2) whether cost monitoring was performed by the owner; (3) the level of support received by the project manager from the contractor's senior management throughout the course of the project; and (4) the early involvement of the contractor's project manager. The model was validated using an additional 36 projects.