Date of Award


Document Type


Degree Name

Master of Science


Department of Systems Engineering and Management

First Advisor

Scott T. Drylie, PhD


The accurate forecasting of escalation is vital to generating useful cost estimates. This research investigates various methods for their accuracy in forecasting an index, PPI3364, which measures price changes in aerospace product and parts. This is done over two different spans of time: fiscal years and FYDP (five out years). Also, this study of methods for forecasting was conducted from the perspective of the practitioner in the PPBE process. Global Insight is the current source for forecasting escalation for the aerospace products and parts industry, but there are concerns, expressed by the Air Force Cost Analysis Agency, regarding the accuracy of these forecasts. Therefore, this paper looks at a range of alternative methods to provide context for the level of Global Insight’s forecasting accuracy. The annual data has six alternative general forecasting methods being tested, many of the general methods having variants amounting to fourteen total methods being tested. There are two general alternative methods for the FYDP data. All forecasting methods are assessed for accuracy based on root mean squared error and then compared to one another to understand which methods outperform others. The outcome of this research is that many methods were found to be more accurate that Global Insight and may serve as alternatives for the forecasting of price changes in aerospace products and parts. The top performing individual method for reducing RMSE is Global Insight adjusted for systematic error which we refer to as Error Adjusted Global Insight. However, this method is augmented by using all the actual escalation and Global Insight forecast data available to building a regression equation where Global Insight’s forecasted values are used to predict the actual escalation in PPI3364 for both the annual and FYDP data. This research finds that the highest performing method for the forecasting of annual data that is not augmented using data that would not be available to the practitioner at the time of forecasting was a 5-year front-weighted moving average.

AFIT Designator



A 12-month embargo was observed.

Approved for public release: 88ABW-2023-0315