The Office of the Inspector General conducted a review of a project to construct the Computer Science Research and Development Center (Computer Science Center) on the Amherst campus of the University of Massachusetts (UMass). Completed in July 1999, the Computer Science Center is a flat-roofed, steel-framed structure with an exterior cladding of metal and glass building panels. The three-story building encompasses approximately 80,000 gross square feet and houses the Computer Science Department, including departmental and faculty offices, research space, a main computer room, classrooms, and conference rooms.
The Division of Capital Asset Management (DCAM), the state's centralized construction management agency, was responsible for planning, procurement, and contract administration for this project. Design work began in 1994 when DCAM contracted with an architectural firm for a feasibility study to confirm the need for a new building and to prepare a program and cost estimate. The feasibility study, completed in December 1995, included a program for a building that could be constructed at an estimated construction cost of $10.4 million.
The study designer determined that UMass placed a high priority on completing the project quickly and recommended a design-build approach to shorten the project schedule. DCAM adopted this recommendation and used the state's modular building procurement law to award a contract to Suffolk Building Corporation (Suffolk) for the design and construction of the UMass Computer Science Center.
The study designer projected that the UMass Computer Science Center could be completed in 18 months using a design-build approach. However, the project took 37 months to complete. DCAM's contract with Suffolk contained a completion deadline of September 9, 1998 and a total design and construction cost of $9,231,000. Suffolk completed the project in July 1999, 10 months late. In the course of the project, DCAM approved change orders that increased the contract cost by $475,985, and in October 1999, three months after project completion, Suffolk submitted a claim to DCAM seeking an additional $2,733,674. In total, the change orders sought by Suffolk would increase the cost of the project by more than one third. In its claim, Suffolk alleged that DCAM and UMass caused project delays by failing to approve design submissions in a timely fashion, by delaying design decisions, and by requesting design changes. However, the Office's review found that the delays and cost overruns were largely caused by factors under Suffolk's control. In addition, the Office found that the Suffolk claim was inflated by charges for damages not permitted under Suffolk's contract, overstated indirect costs, and charges for undocumented design services that cannot be verified.
Alternative Approaches to Construction: Design-Build vs. Design-Bid-Build
In general, public building projects in Massachusetts must be constructed using the design-bid-build approach required by the state's designer selection and construction bidding laws. Contract procurement for a design-bid-build project proceeds in two distinct steps. First, the awarding authority selects a designer to prepare complete and detailed plans and specifications for the project. Second, the awarding authority uses the plans and specifications to solicit bids for construction. For a Massachusetts public building project, the awarding authority solicits filed sub-bids for work to be performed by subcontractors, as well as bids from general contractors. The contract is awarded to the eligible and responsible general contractor submitting the lowest bid.
In contrast to the two-step procurement approach used in a design-bid-build project, a design-build project entails the award of a single contract to one entity - a design-build contractor - that is responsible for both design and construction. The purported advantages of the design-build approach include: potential time savings produced by overlapping design and construction schedules, the design-build contractor's single-point accountability to the owner, increased coordination and collaboration between the designer and the general contractor, and improved working relationships between the general contractor and subcontractors.
In general, public awarding authorities in Massachusetts must obtain special legislative authorization to waive the designer selection and construction bidding laws in order to use the design-build approach. On the UMass Computer Science Center project, the study designer recommended that DCAM use design-build procedures authorized by the state's modular building procurement law to avoid the need to obtain legislative authorization for a design-build project. The study estimated that this approach would save nine months in comparison with a design-bid-build process.
The Procurement Process
In June 1996, in accordance with the study designer's recommendation, DCAM issued a request for proposals (RFP) for a modular building under M.G.L. c. 149, §44E, the modular building procurement law. By the proposal deadline of September 24, 1996, DCAM had received five competing proposals. The DCAM Commissioner designated an evaluation committee that collectively assigned a point score to each of the proposals based on the following criteria: quality of proposed design and materials, qualifications of the contractor and its subconsultants, and completeness and feasibility of the proposed schedule. The proposal submitted by Suffolk and a team of design subconsultants headed by the architectural firm of Whitney Atwood Norcross Associates, Inc. (WAN) received the highest score.
On February 25, 1997, DCAM and Suffolk executed a contract in the amount of $9,231,000. The contract had an effective start date of March 18, 1997 and required the project to be completed by September 9, 1998.
Summary of Findings
The design-build approach to the UMass Computer Science Center did not produce the anticipated time savings: Suffolk did not complete the project until July 1999, 10 months after the contract deadline. The project also fell short of UMass's standards for quality in at least two significant respects. First, the underground steam line supplying the building with steam heat was directly buried rather than enclosed in a concrete tunnel in accordance with UMass's standards. According to the UMass Director of Capital Projects, the absence of a concrete tunnel will shorten the life expectancy of the steam line and burden UMass with higher maintenance costs over the life of the building. Second, Suffolk's failure to achieve a smooth, architectural finish for exposed building foundation and retaining walls was remedied by applying a coating to fill in and disguise defects. This coating has already begun to show signs of wear and is expected to cost more than a smooth concrete finish to maintain over the life of the building.
After the project was complete, Suffolk submitted a $2.7 million claim against the Commonwealth, contending that project delays and cost increases were due to "apparent conflicts between DCAM and UMass about what was required under the Contract, and what UMass wanted in the finished product." According to Suffolk's claim, UMass representatives were unable to make decisions on final design items on a timely basis and changed those decisions after they were made.
DCAM contracted with a consultant to analyze Suffolk's claim and, based on the consultant's recommendation, initially agreed to pay Suffolk $1.4 million to settle the dispute. After the Office raised concerns about the claim, the DCAM Commissioner did not approve the settlement offer.
A more detailed summary of the Office's major findings is provided in the following pages.
Although DCAM used the modular building procurement statute to contract for this project, the UMass Computer Science Center is not a modular building.
The Computer Science Center was not constructed as a modular building. Instead, it is a conventionally constructed building that used steel components provided by a pre-engineered metal building systems manufacturer for the third-floor and roof structure. DCAM therefore improperly avoided the separate designer selection and construction bidding procedures required for public building projects.
In the RFP, DCAM stated that a building constructed with steel columns and beams provided by a pre-engineered metal building systems manufacturer would be deemed a modular building for purposes of the modular building procurement law. However, the term "modular building," as defined in M.G.L. c.149, §44E, refers to buildings that are either factory-assembled into three-dimensional modules or portable structures. In the opinion of the Office of the Inspector General, a building constructed on site from pre-engineered steel columns and beams does not meet this statutory definition; thus, the definition of a modular building contained in DCAM's RFP permitted contractors to propose non-modular buildings. Moreover, the components provided by the pre-engineered building systems manufacturer for this project were limited to the third floor and roof structure. Regardless of whether a pre-engineered metal building is a modular building for purposes of the public construction bid laws, the first two floors of the Computer Science Center were conventionally constructed even by DCAM's definition.
Most delays and cost overruns on the UMass Computer Science Center were caused by factors under Suffolk's control.
The findings in this report show that, contrary to Suffolk's claim, the problems contributing to delays and cost overruns were not caused by major design changes initiated by DCAM or by UMass after the project began. Rather, most major problems encountered on the project were attributable to Suffolk's failure to ensure that design work was complete, accurate, and timely; Suffolk's continual efforts to reduce construction costs through design revisions; and Suffolk's failure to take timely steps to replace non-performing subcontractors.
A number of Suffolk's decisions that had significant impacts on both the project schedule and the construction quality appeared to be aimed at minimizing its costs. For example, Suffolk failed to obtain complete survey and site information or to submit a complete and accurate civil/site design early in the project. The lack of complete and accurate design documents led to construction problems that required Suffolk to redo work. Suffolk continued to revise its steam line design, in part to lower construction costs, for more than six months after being directed to proceed with direct burial of the steam line. Suffolk failed to replace its concrete subcontractor, despite serious performance problems that contributed to delays and compromised the project quality. Suffolk's problems with a drywall subcontractor and a subcontractor responsible for fabricating and installing the building's three metal stairways also delayed project completion.
The design-build approach was not appropriate for the UMass Computer Science Center project.
The design-build approach streamlines design and construction, in part by giving the design-build contractor responsibility for and control over final design decisions. The primary advantage of this approach from the owner's viewpoint is that it can, if correctly managed, accelerate the schedule. The major disadvantage is that the owner must give up control over final design, increasing the risk that the completed building will not meet its needs or standards for quality.
Some of the problems encountered by the Computer Science Center project can be attributed to the design-build approach, which opened the door to disputes over the project design requirements. For example, the RFP called for a concrete tunnel for the steam line but contained little design information about such a tunnel. The lack of design specificity allowed Suffolk to argue that the tunnel was not required. The dispute between DCAM and Suffolk over the RFP requirements contributed to project delay and gave Suffolk an opening to submit a $1.2 million change order for the tunnel, seeking a 13 percent increase in the price of the contract. In the end, DCAM abandoned its initial position that the RFP design required a tunnel and instructed Suffolk to directly bury the steam line piping without a concrete tunnel. As a result, the steam line has a shorter life expectancy and will likely cost more to maintain. The steam line tunnel dispute highlights the importance of developing adequate design parameters in an RFP for a design-build project to avoid potentially costly disagreements over the RFP requirements.
Differing interpretations of the design requirements for the atrium lobby stairway also contributed to delay. From the outset, UMass envisioned a decorative, open stairway with a stainless steel handrail in the building's atrium lobby. The RFP described this design concept, which was to be a focal point of the building. However, Suffolk's proposal featured a lower-cost design alternative with a gypsum board half-wall and wooden handrail. UMass rejected Suffolk's proposed design deviation before the contract was executed. Suffolk's next proposed alternative, a perforated metal panel rather than the stainless steel handrail, was also unacceptable to UMass. In the end, UMass got the higher-quality, open stairway with the stainless steel handrail described in the RFP, but delays in developing the final design held up project completion and increased the cost.
The Computer Science Center's cable tray system is another example of a problem that is more likely to occur on design-build projects, which tend to rely on performance specifications rather than the complete design specifications used for design-bid-build projects. The RFP contained a performance specification that required the building's cable tray system to be accessible to allow the building's users to easily reconfigure the computer cabling. After Suffolk began to install the cable trays, UMass determined that the installation did not provide adequate clearance for access. The ensuing dispute between DCAM and Suffolk over what constituted an "accessible" cable tray system led to delays and added costs. This dispute highlights the limitations of performance specifications for buildings, where it is often difficult to describe required functions with adequate specificity.
In sum, disputes over design requirements and problems with the quality and timeliness of Suffolk's design submissions highlight the disadvantage inherent in giving a construction contractor control over final project design. UMass's efforts to ensure that the building met its standards for quality and functionality were met with resistance by Suffolk and resulted in disputes, despite clear contract terms that required Suffolk to submit final design work to DCAM for approval.
This project also shows that the design-build approach, which is intended to foster collaboration between the designer and the contractor, does not guarantee a harmonious relationship between them. The Office found that Suffolk and WAN were engaged in a dispute over WAN's design fee almost from the outset. As the project progressed, Suffolk contended that errors on the part of WAN's design team had resulted in higher construction costs, escalating the amount of the fee dispute.
Top DCAM management did not support efforts of DCAM project personnel to enforce contract requirements and maintain control over the project.
DCAM recognized that relying on a design-build contractor to develop the final project design increased the risk that project quality could be compromised. DCAM's contract characterized the contractual arrangement as a "special relationship," which it described as follows:
In order to accomplish the purposes of the Agreement efficiently, the parties hereby declare that a special relationship of trust, confidence, and professionalism exists between the parties and will continue to exist throughout the duration of the Agreement.
In addition, DCAM developed a comprehensive set of contract requirements and project management procedures to ensure adequate control and oversight. The contract required Suffolk to submit complete plans and specifications for DCAM's review and approval prior to beginning the related construction work. In addition, DCAM contracted with an independent design firm, Helene-Karl Architects, Inc. (HKA), to perform a comprehensive review of Suffolk's design submissions for completeness and accuracy.
Project records show that DCAM project personnel attempted to enforce design review and other contract requirements, including requirements for Suffolk to obtain approval for its proposed schedule. DCAM project personnel frequently reminded Suffolk of its schedule for submissions and documented Suffolk's failure to comply with the schedule. Despite these efforts on the part of DCAM project personnel, Suffolk failed to meet its own proposed schedule for submissions or to submit complete, revised design documents for approval as required by the contract.
In many instances, top DCAM management waived these contract requirements. For example, top DCAM management gave "partial and conditional" approval to construction documents for the civil/site design, over the vehement objections of HKA, despite significant unresolved design problems. Similarly, top DCAM management effectively waived the final approval requirement for the structural design by giving Suffolk permission to proceed with foundation work before the structural design was complete. Given this lack of support from top DCAM management, it is not surprising that DCAM project personnel were unable to maintain effective control over the project.
Suffolk's failure to adhere to its proposed schedule suggests that it was unconcerned about the possibility that DCAM would enforce the $1,000 per day liquidated damages provision for late completion. At the conclusion of the project, the DCAM resident engineer recommended deducting $202,000 in liquidated damages from Suffolk's requisition. DCAM records contain no explanation for the decision made by top DCAM management to pay Suffolk's last requisition in November 1999 without withholding any money for liquidated damages.
Top DCAM management failed to conduct a rigorous assessment of the merits of Suffolk's $2.7 million claim.
Project records indicate that DCAM project personnel conducted an assessment of Suffolk's proposed change orders throughout the project, verifying the reasonableness of and seeking documentation for claimed extra costs. At the end of the project, DCAM project personnel had approved approximately $476,000 in change orders to Suffolk.
Many of the extra costs rejected by DCAM project personnel in proposed change orders were later incorporated into Suffolk's $2.7 million claim. In dealing with the claim, top DCAM management took an approach that differed dramatically from the careful assessment done by DCAM project personnel. DCAM hired a claims consultant under an agreement that contained no written instructions or scope of work to analyze the claim. The work product produced by DCAM's claims consultant and DCAM's subsequent negotiations with Suffolk reflect a lack of any substantial assessment of the legal merits or the costs included in Suffolk's claim. For example, the claims consultant recommended and top DCAM management agreed to pay the majority of Suffolk's claimed costs for 42 weeks of delay. However, the claim consultant's analysis contained no supporting evidence or rationale for recommending that DCAM accept responsibility for 42 weeks of delay. Moreover, the findings in this report show that the majority of delays were attributable to factors under Suffolk's control. In addition, the Office's review found that major cost items were not documented and could not be verified and that the contractual and legal basis for many elements of the claim is doubtful. For example:
- Suffolk's claim included $314,000 for undocumented design services that cannot be verified;
- Suffolk entered into an agreement with one of its subcontractors that gave Suffolk an incentive to inflate the subcontractor's claimed costs;
- Suffolk acknowledged that its claim had overstated its general conditions costs by $287,108.
The failure of DCAM's claims consultant to closely scrutinize this claim suggests that the $1.4 million settlement that top DCAM management agreed to pay Suffolk represents a decision to seek a ready compromise rather than to ensure that taxpayers do not pay more than they should for this project. The public is ill-served by this kind of compromise, not only because it may result in paying too much for this project, but because it will convey the message that the Commonwealth does not carefully analyze construction claims and will pay a premium to resolve a dispute. This message is an invitation to inflated claims on future state construction projects. Moreover, contractors who succeed in getting questionable claims approved may believe that they can lowball future bids with impunity and recoup their costs through change orders. By not scrutinizing change orders and claims and by not enforcing the requirements of its own contracts, the Commonwealth undermines fair competition for its future construction contracts.
DCAM decided not to execute the $1.4 million settlement agreement after learning of the Office's concerns. Instead, DCAM retained a law firm to assess the consultant's report and recommendations. The law firm has reportedly recommended conducting a new analysis of Suffolk's claim.