A methodology for predicting company failure in the construction industry

This paper describes research directed towards the development of an operational system for identifying construction companies in danger of failure. The major component of the system combines financial ratio analysis and the statistical technique known as multivariate discriminant analysis, to produce a predictive model made up of seven variables, measuring distinct aspects of company financial structure, all transformed into a single value called the Z score. Good distinction between the scores of solvent and failed companies was provided. This technique is widely employed in the commercial sector with much of the work concentrated on failed and healthy companies. A secondary method was developed to reinforce the financial approach, whereby managerial performance aspects are weighted, combined and a cut-off, known as the A score value, determined to separate the two groups. The concept behind the A score is based on the belief that if a company is in financial difficulty the reason generally relates to inadequate management ability and errors perpetrated earlier. The A score is designed to address this aspect of failure prediction. By operating these two principal methods in conjunction, it is possible to predict with confidence who could be next to fail.