Predicting the Likelihood of Cost Overruns: An Empirical Examination of Major Department of Defense Acquisition Programs

Abstract

This article provides a method for predicting the cost of a major acquisition program five years after program development approval. The authors’ work extends the effort begun by Asher and Maggelet by adjusting for changes in median cost growth factors for each acquisition domain. Procurement average unit costs at program approval and five years afterwards were compared for 101 major United States Department of Defense acquisition programs. Individual program cost growth ratios were gathered into 28 sub-groups, based on the product and year of approval. The authors employed a least squares regression to obtain the equation’s coefficients. In 26 of the 28 groups of data, a comparison of predicted costs to actual costs resulted in very low p-values, leading to the conclusion that the revised formulation is a valid predictor of program cost at the five-year point in product development.