Forecasting costs and completion dates for defense research and development contracts

It is hypothesized that the actual final cost and time required to develop a new technological system can be forecast from initially negotiated costs and times and other administrative details that are available when the contract is negotiated. This hypothesis was tested by examining multiple regressions of total cost and time as functions of data available at the end of negotiations for 73 contracts administered by one division of the Air Force. Factors correlating with over 70 percent of the cost growth were identified: 1) the initial cost and period of performance, 2) whether or not the entire system is investigated, 3) whether or not a significant change in scope of the effort is anticipated, 4) whether or not the effort is classified as a study (in contrast to hardware development), and 5) who is the procurement officer. Factors correlating with one-third of the growth in the time required to perform the contract were identified: 1) what type of performance incentives and limitations are specified, if any, 2) whether or not the study is classified as “level of effort,” 3) the initial cost and period of performance, 4) the project branch, and 5) whether or not a significant change of scope is anticipated. Regression equations and coefficients are reported for the division studied.