Assessing the Accuracy of Cost Estimates using Statistical Techniques

Abstract

Accurate cost estimates are vital to sustaining Navy Expeditionary Combat Command (NECC) operational forces. Currently, NECC needs a method to assess the accuracy of future cost estimates that will not disrupt its existing cost estimation process during the POM cycle. Accuracy is the sole metric used in this study to analyze and assess the effectiveness of a cost estimate. This research uses statistical techniques to assess the cost estimate’s accuracy, including single variable linear regression, Monte Carlo simulation, and a proportional scale-down model. The sample of sustainment cost data spans from FY 2016 to 2021. The point estimates from the regression model revealed that the total aggregate cost decreases by just under $16.6 million in constant year 2018 dollars (CY18$) each fiscal year. I applied the point estimate and standard error to the Monte Carlo simulation to produce a normal probability distribution with a range of possible outcomes (i.e., probability of occurrence linked to each value within that range). For the FY2025 scenario simulation, the mean value representing the most likely estimate is $303,649,744 (CY18$), and the standard deviation is $7,767,962 (CY18$). The proportional scale factor model breaks down the aggregate estimate to a lower level of programmatic detail for additional analysis. Applying statistical techniques to the existing process will help ensure that operational forces receive sufficient resources to accomplish the mission.