APPENDIX C: A&E PROJECT PERFORMANCE STATISTICS
During the study reported here, a limited and informal survey was made to determine whether data are available that might be used to develop indicators and benchmarks of A&E performance. It was found that many large users of A&E services, in both the government and private sector, do maintain records that might be analyzed, but these records are typically developed in an idiosyncratic manner, depending on the specific characteristics of the project and the personnel involved. There is seldom any effort to develop consistent definitions and uniform data collection procedures.
Some federal agencies have collected some data on construction experience that is relevant to A&E performance. The study committee reviewed two such data sets in an effort to determine whether more-explicit guidelines could be developed in this study.
One agency provided the committee with a data set describing that agency's experience with the development of 151 facilities construction projects completed between January 1985, and March 1992. The construction contract award values of these projects ranged from $55,000 to more than $12 million. The range of project types was broad, and they included new building construction, structural and interior repairs, subsurface facilities, roads, and other infrastructure systems. Project
durations ranged from approximately 100 days to more than five years.
Change orders were reported on more than 90 percent of these projects. For approximately 50 percent of the projects, ten or more changes were reported, from all sources. About 73 percent of projects had changes attributed to design, and approximately 65 percent had more than one such change. About 21 percent had more than five design-related changes, and about 14 percent had ten or more. However, approximately 20 percent of the projects, regardless of the number of change orders, reported no cost growth at all, and costs declined below contract amount in about 5 percent of the cases. For approximately 50 percent of projects, cost growth was less than 6 percent, and it exceeded 16 percent of contract value in fewer than 25 percent of cases.
A second agency provided data on 133 projects, which ranged in size from $190,000 to $24.5 million. The range of project types was broad, similar to that of the first agency. It included projects from child care centers to wastewater treatment systems to road improvements. Project durations ranged from 200 days to just over three years. Nineteen percent of these projects had no change orders at all, and 23 percent had ten or more changes, from all sources. While changes were generally less frequent than in the first agency's case, 71 percent of projects had changes attributed to design, a fraction similar to that of the first data set. Thirty-three percent had five or more design-related changes, and 20 percent had ten or more. All projects experienced some cost growth, regardless of the number or source of change orders. Cost growth exceeded 6 percent in 43 percent of the cases, but only 20 percent of projects experienced cost growth of 16 percent or more.
The committee concluded that these two limited analyses were inadequate, by themselves, to support development of
solid indicators of A&E performance. Nevertheless, the analyses do provide evidence supporting rules of thumb that were suggested by committee members' experience. Changes and cost growth are not unusual, but only about 20 to 25 percent of projects experience cost growth in excess of 15 percent of initial contract value. And when change orders are required, they are more likely than not to be attributed to design problems. However, the data sets were too small and not clearly representative of the broad range of government building experience, and the definitions too uncertain to draw conclusions regarding how likely it is that design changes will be the primary cause of cost growth.
On the basis of their survey and analyses, the committee concluded that efforts to develop valid indicators of A&E performance must start with development of a well thought-out study design to guide collection of data for analysis. This conclusion became the basis for the study committee's recommendation that professional organizations and the government should work together to develop a database and conduct analyses of A&E performance.