Experiences of Federal Agencies with the Design-Build Approach to Construction (1993)

As noted in the introduction, completed questionnaires were received from seven agencies on 27 design-build projects. The characteristics of the projects reported on are summarized below. More detailed information on the projects can be found in the Appendix A.

The projects reported on included a wide variety of facilities; specifically:

• Seven general office buildings ranging in size from a 34,000 sq. ft. three-story addition to a new 974,000 sq. ft. building.

• One 921,000 sq. ft. federal courthouse building.

• One 192,000 sq. ft. dormitory.

• Three mail distribution facilities ranging in size from a 137,800 sq. ft. distribution center to a 698,000 sq. ft. combination mail and vehicle maintenance facility.

• One 440,900 sq. ft. vehicle maintenance and parking structure

• Five special industrial/laboratory facilities ranging in size from a 6,480 sq. ft. by 140 ft. high “canister rotation facility” to a 142,900 sq. ft. addition/alteration of a cryptologic support center.

• One non-building project involving the installation of replacement heating equipment in several facilities.

• One 90,000 sq. ft. commissary.

• One 74,000 sq. ft. outpatient clinic.

• Two “modular transitional bed [medical] buildings,” one 30,000 sq. ft. and the other 60,000 sq. ft.

• One youth activity center of unspecified size.

• One dining facility of unspecified size.

• One special operations headquarters of unspecified size.

• One comprehensive health care center of unspecified size.

Contrary to the widely held view that the design-build approach is most appropriate for relatively simple projects, only three of the projects reported on were characterized by respondents as being of low or medium-low complexity. Conversely, 12 projects were characterized as being of high complexity. The other 12 projects were characterized as being of medium or medium-high complexity.

The earliest contract-award date for any of the projects reported on was 1984. The latest contract-award date reported was 1992. The average contract-award date for all the projects was early 1989. For the committee's purposes, these contract-award dates are almost ideal because they mean that the projects were completed long enough ago that any serious design or construction flaws would have shown up by now, but recently enough that recollections about the projects should still be fresh.

The 27 projects reported on were located in 11 states, as follows: one project each in Massachusetts, Georgia, Kansas, North Carolina, Oklahoma, and South Carolina; two projects each in Illinois and Texas; three projects in Washington (state); four projects in California and New York; and six projects in Florida. Although the committee would have preferred somewhat more even geographical distribution, the distribution is acceptable since all regions of the country are represented.

The dollar amounts of the design-build contracts reported on ranged from a low of $725 thousand to a high of $276 million; the average contract amount for all 27 projects was $39.8 million. The average amount is relatively high because the sample includes a disproportionate number (four) of projects costing more than $100 million. Nevertheless, the distribution of contract amounts is good considering the relatively small size of the sample.

For nine of the projects reported on, the primary reason given for the use of the design-build approach was to save time. For another six projects,

the respondents indicated that the design-build concept had been used primarily to save money--or, as one respondent put it, “to get more building for the money.” On two other projects the design-build approach was used in the hope of saving both time and money; on one of these, however, the fact that the project was built with “nonappropriated ” funds was a major consideration. Design-build was used on three projects (all by the same agency) at the “direction” of Congress. One respondent reported that design-build had been used on two very large projects because the agency wanted to have them carried out by a “developer ” due to their “enormity.” Among the reasons cited for use of the design-build approach on the remaining projects were the following: To “take advantage of the contractors experience in state-of-the-art building concepts ”; to get better coordination of the design and installation of equipment; to minimize delays and claims and to centralize responsibility for design decisions; and to test the concept. (The reason for using the design-build approach was not indicated for one project.)

It is not difficult to understand why an agency would believe that the design-build approach could save time compared to the traditional design-bid-build approach, since with the design-build approach the delay between completion of design and start of construction can be eliminated--or at least greatly reduced. However, it is not self evident--and the respondents generally did not explain--why agencies expected to save money using the design-build approach. One possible explanation is that agencies thought that having contractors involved in the design process would result in more cost-effective designs; however, that is only speculation. The most important issue is whether the agencies expectations were realized, which will be discussed in the next chapter.

Twelve of the projects reported on were carried out under fixed-price/lump sum contracts; however, one of these started out as a type of lease-purchase arrangement. The other 15 projects were all carried out under some type of cost-reimbursement contract with a cap on total costs. However, the committee is uncertain of the exact nature of these contracts since the respondents generally did not give any details on them and sometimes used non-standard terminology to describe them; for example, four contracts were referred to as “not-to-exceed-cost contracts ”; six contracts were referred to as “cost-plus-fixed-fee contracts with a bonded negotiated price cap”; one contract was referred to as a “cost-plus-fixed-fee contract with a guaranteed maximum price”; three contracts were referred to merely as a “guaranteed-maximum-price contracts”; and finally, one contract was referred to as a “guaranteed-

maximum-price contract with cost incentives.” The committee suspects that many of these contracts are very similar even though they have been given different names.

With regard to the process used to select contractors, all the respondents stated or implied that all selections were based on evaluations of contractor-submitted proposals;⁴ however, the procedures and factors used in making evaluations varied considerably.

The most basic issue was whether the proposal-evaluation procedures involved one step or two steps. Ordinarily, the difference between the two approaches revolves around the consideration of cost or price: With the one step approach, cost is just one of several factors that are considered simultaneously, whereas with the two-step approach, cost is considered separately after all the proposals have been evaluated with regard to all other factors. (However, as discussed later, cost apparently was not a factor at all in the selection of the contractor for some projects.)

The use of the two-step approach was explicitly mentioned in only three questionnaires, whereas the use of the one-step approach was indicated or implied in 21 questionnaires. (Three questionnaires did not indicate or imply which approach had been used.) These results were surprising since, in the past, federal procurement officers have tended to favor the two-step approach. The committee believes that some respondents might have implied the use of the one-step approach by mistake and that the two-step approach might have been used on more than three projects.

With regard to the factors used to evaluate proposals, cost (price) was the most commonly listed factor. However, two respondents emphasized that cost was not the most important factor. It is also interesting that cost was not even listed as a selection factor for two projects--one on which a two-step procedure⁵ was used and another on which a one-step approach was used. Presumably, in both cases a price was negotiated with the firm receiving the highest rating in the evaluation--the approach used with architect-engineer contracts under the Brooks Act. The factors other than cost that were mentioned frequently were contractor experience/qualifications, management plan, proposed design, and key personnel. It should be emphasized that, as

with other issues addressed in the questionnaire, respondents used a wide variety of terms when discussing contractor evaluation factors and that some of the differences noted by the committee could be more semantical than actual.

The number of proposals received for the projects reported on ranged from one to 13; the average for all projects was 5.67.

Ten of the design-build contracts reported on were awarded to general contractors; nine were awarded to teams (consortia) composed of a general contractor and an architect-engineer firm; three were awarded to firms specializing in design-build projects; two were awarded to modular building manufacturers; and three were awarded to developers.

As with other project characteristics, the diversity of responses on these two issues was surprisingly broad given the small sample size.

One issue frequently raised with regard to design-build projects is whether an owner should have a preliminary design developed before awarding a design-build contract--and if so, to what extent it should be developed--or whether it is better to give the design-build contractor just written criteria on the desired facility. For the 27 projects reported on in the committee's survey, the extent of design work performed prior to the award of the design-build contract⁶ was as follows:

• One project--80 percent.⁷

• One project--50 percent (development drawings plus specifications).

• Four projects--35 percent (preliminary design with complete specifications).

• One project--30 percent.

• One project--25 percent.

• Seven projects--20 to 25 percent (preliminary engineering report/study, floor plan plus specifications).

• One project--15 percent.

• Ten projects--10 percent (design/concept stage).

• One project--0 percent.

As with most other factors examined by the committee, the practices of the agencies with regard to the amount of design work performed prior to award of a design-build contract were surprisingly diverse.