An investigation of possible discriminating earned value variables in Department of Defense major acquisition program cancellation

Abstract

Some major defense acquisition programs (MDAPs) are cancelled while others are not. The research evaluates earned value (EV) data reported by 12 cancelled MDAPs and compares it to the EV data reported by 12 MDAPs that had significant cost overruns yet were not cancelled (referred to as troubled programs). The study investigates whether there are differences in the key EV metrics reported for cancelled and troubled but not cancelled programs and seeks to develop a model that captures the probability of cancellation based on EV information. The thesis hypothesizes that cancelled programs would have more unfavorable cumulative cost and schedule variances, greater cost growth in the estimate at completion (EAC), and more disparity between the contractor and program manager cost estimates. Mann-Whitney tests were used to identify significant differences in EV metrics between the two groups and a probit model was used to further explore the relationship between EV information and the probability of program cancellation. The Mann Whitney and probit results, viewed together, suggest that early cost growth in a program may be an indicator of program cancellation. The thesis also suggests ways to improve EAC using probabilistic methods, including quantifying the level of cost risk associated with the EAC.