Do We Need a Different Approach to Statistical Analysis of Research, Development, Test, and Evaluation Cost Growth?

Abstract

McNicol (2018; hereafter Acquisition Policy) obtained remarkably strong statistical results for a simple model of cost growth in Major Defense Acquisition Programs (MDAPs). Following previous studies, Acquisition Policy used Program Acquisition Unit Cost (PAUC), the numerator of which is the sum of Research, Development, Test, and Evaluation (RDT&E) cost and procurement cost. This paper asks whether the model used by Acquisition Policy characterizes RDT&E cost growth and growth in procurement cost individually as well as it does PAUC. It does not. As would be expected, the results for Average Procurement Unit Cost (APUC) are very similar to those for PAUC and are marginally stronger statistically. The results for RDT&E also are quantitatively similar to those for PAUC, but the explanatory power of the model is far lower, suggesting either much greater variability in RDT&E cost estimates or flaws in the model as applied to RDT&E. The paper concludes with suggestions for improving models of RDT&E cost growth.