A Guidebook for the Evaluation of Project Delivery Methods (2009)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Distinguishing Characteristics of Transit Projects Several types of project delivery methods are currently available to the owners of publicly funded transportation projects in the United States. It is important—especially in the case of large, complicated transportation projects—to select the most appropriate project delivery method. Contractual relations, contemporary laws and regulations, owners’ perceptions of risks, awarding mechanisms, and method of payment all influence the selection of a project delivery method. This guidebook in no way advocates one project delivery method over another. In fact, it is the expressed purpose of this effort to assist transit agencies in choosing the delivery method, from among the many project delivery methods, that is most appropriate for a particular proj- ect. In the material that follows, alternative project delivery methods will be compared with the traditional design-bid-build (DBB) project delivery method, which functions as a benchmark against which all other methods can be evaluated. The literature shows that the use of alterna- tive project delivery methods can accrue benefits for owners. However, the benefits of alterna- tive project delivery methods presented in the literature occur most often across a population of projects rather than on an individual project. Thus, the reporting of benefits found in the liter- ature should not be misconstrued as advocating one project delivery method over another. All project delivery methods have yielded both successes and failures. Selecting the wrong project delivery method is often a significant driver of project failure. Therefore, the reader should understand the results of the research reported herein as evidence that a given project delivery method may be used successfully on a specific set of projects, not as evidence that any particu- lar project delivery method is inherently superior to all others. Before describing various project delivery methods, it is important to note the features of major transit projects that distinguish them from other transportation projects. Transit projects are larger projects, usually in excess of $100 million. Transit projects, especially projects with fixed-guideway systems, usually consist of at least two large contracts: (1) civil and (2) systems. The nature of these two contracts and the specialization required for each are such that usually two different entities deliver these contracts. This circumstance makes coordination between these two entities of paramount importance to project success. Generally, in DBB projects, the owner hires a construction manager (CM) (this construction manager is a representative of the owner, i.e., the agency CM, as opposed to the construction manager at risk [CMR]) to coordi- nate these two separate contracts and manage the work. In design-build (DB) projects, the design-builder often subcontracts to separate systems and civil contractors or forms a joint ven- ture with them. Another feature of transit projects is that they are usually built in major urban population centers. This increases the complexity of dealing with various stakeholders. There- fore, a major criterion in choosing a project delivery method for a transit project is the delivery method’s ability to accommodate the needs of various stakeholders in a complex environment. 8 C H A P T E R 2 Background and Definitions

Whenever a commuter rail project is considered, a freight line may be in the mix where the owner will have to share the line with temporal separation or track separation. This circumstance makes coordination with the railroad company owning the freight line extremely important. The importance of reaching agreements with the railroad company and clarifying the details of the work and the responsibilities of the various parties cannot be overemphasized. The railroad com- pany usually wants to do the track work with its own forces on a cost-reimbursable basis, and this puts all the risk on the owner. This also increases the constructor’s risk because its work may be impacted by the railroad. This makes early involvement of the construction contractor very important to project success. Also, the railroad company tends to do the work at its own pace while considering project milestones; as a general rule, the agency does not enjoy the same degree of control that it exerts over the constructor with the railroad company. Another distinguishing characteristic of transit projects is that typically they incorporate fea- tures that are unusual in an engineering project, and thus transit projects may require the involve- ment of professionals from the fields of architecture, landscape architecture, and interior design, as well as engineering. The integration of “vertical” construction features (e.g., parking structures and transit stations) with “horizontal” construction features (e.g., track beds, bridges, and road- way elements) creates a need for a comprehensive set of design and construction services that is not normally found in transportation projects. Additionally, transit agencies’ need to integrate their facilities with other transportation modes demands another comprehensive set of design and construction service providers and requires a more flexible approach to design and construction than is required by single-mode transportation projects. These characteristics of transit projects drive the need for a “toolbox” of project delivery methods that permits a transit agency to select the appropriate project delivery “tool” based on the technical demands of a given project. Transit projects are not usually money makers (unlike some toll roads in the highway sector). Therefore, it is difficult to generate interest in potential public-private partnerships. Financial institutions, which are sometimes interested in supporting toll road and bridge projects, are usu- ally not interested in transit investment, although that may change in the future. Finally, federal support for transit projects, often crucial to bringing the project into being, depends on specific steps that are not similar to other transportation projects. The Federal Tran- sit Administration (FTA) plays an important role in this process. Various transit agencies com- pete for federal dollars by preparing specific reports to the FTA. Depending on the rating that a project receives, it may be permitted to move to the next development stage. The owner agency must meet certain requirements to advance from project planning to final design and finally to construction. If, during various phases of project development and as project scope becomes more accurate, the rating of the project falls below the required threshold, there is a possibility that the project may be discontinued. The burden is on the owner agency to ensure that the proj- ect remains viable and meets federal requirements. Evolution of Current Alternative Delivery Methods in Transit Projects Public procurement law has historically limited public agencies to using DBB construction project delivery only. The current wide range of project delivery methods is a relatively recent development for publicly funded transit projects in the United States. The development of the public procurement laws limiting public agencies to use of the DBB project delivery method can be traced in part to the Brooks Act. Enacted in 1972, the Brooks Act (Public Law 92-582) states that design services on federally funded projects in the United States should be procured on the basis of qualifications only. Alternatively, numerous laws and statutes throughout the Background and Definitions 9

United States have limited the procurement of constructors to the lowest responsible, responsive bidder. The combination of these two procurement practices has helped solidify the proliferation of DBB in the public sector. This method was the traditional transportation project delivery method until the introduction of DB and design-build-operate-maintain (DBOM) in the Inter- modal Surface Transportation Efficiency Act of 1991.1 Another step was taken in 1996, when the Federal Acquisition Reform Act explicitly authorized the use of DB for federal projects. After that, the Transportation Equity Act for the 21st Century (TEA-21), Public Law 105-178, allowed state departments of transportation (DOTs) to award DB contracts if the enabling state-level legisla- tion was in force. Subsequent to the successful use of DB in several projects, many states passed new legislation and codes to allow alternative project delivery methods, i.e., DB and CMR. Adding the responsibility of operation and maintenance to DB projects created another delivery method, DBOM. The differences among delivery methods, the unique characteristics of each project, and the vast variety of parameters affecting the selection of a project delivery method, have made selec- tion of a project delivery method complicated for many owners. The purpose of this guidebook is to facilitate the decision-making process by clarifying the differences among the project deliv- ery methods and proposing a structured decision-making approach that incorporates all the per- tinent parameters. Definitions of the Delivery Methods Since the early 1980s, owners of construction projects have been putting greater pressure on the architecture/engineering/construction (A/E/C) industry to improve quality, reduce cost, and, more importantly, compress the period from project conception to project completion for all kinds of public and private facilities. As a result, both owners and the industry have experi- mented with various forms of project delivery with varying degrees of success. The adoption of alternative project delivery methods has added to the challenge of selecting the method most appropriate to the owner’s needs and desires as well as the project’s technical requirements. This report provides a set of standard project delivery definitions to help communicate the technical requirements for bringing a new project from the owner’s conception to operation and finally to decommissioning. Project delivery method is a term used to refer to all the contractual relations, roles, and respon- sibilities of the entities involved in a project. TxDOT defines “project delivery method” as follows: “A project delivery method equates to a procurement approach and defines the relationships, roles and responsibilities of project team members and sequences of activities required to com- plete a project. A contracting approach is a specific procedure used under the large umbrella of a procurement method to provide techniques for bidding, managing and specifying a project” (Walewski, Gibson, and Jasper 2001). The Associated General Contractors of America (AGC) (2004) defines “project delivery method” as “the comprehensive process of assigning the contrac- tual responsibilities for designing and constructing a project. A delivery method identifies the pri- mary parties taking contractual responsibility for the performance of the work.” Thus, different project delivery methods are distinguished by the way the contracts among the owner, the designer, and the builder are formed and the technical relationships among parties within those contracts. The Construction Industry Institute maintains that there are really only three fundamental project delivery methods: DBB, DB, and CMR (Construction Industry Institute 1997). While 10 A Guidebook for the Evaluation of Project Delivery Methods 1In 1992, the FTA announced the initiation of a Turnkey Demonstration Program (Federal Register Vol. 57, No. 157, 8/13/92) and later selected five projects for DB implementation. These projects were (1) the Los Angeles Union Station Intermodal Terminal, (2) Baltimore Light Rail Transit, (3) San Juan Tren Urbano, (4) Bay Area Rapid Transit in San Francisco, and (5) Hudson-Bergen Light Rail Transit.

there are a multitude of names for project delivery methods throughout the industry, the Con- struction Industry Institute is essentially correct. Therefore, this report will focus its information on those three methods. The AGC also distinguishes between the delivery method and the management method. The management method “is the mechanics by which construction is administered and supervised” (AGC 2004). This function is either retained by the owner agency or is outsourced. An example of outsourcing the management process is to hire an agency CM to represent the owner’s inter- ests during design and construction. Theoretically, any management method may be used with any delivery method. For example, an owner may hire an agency CM to manage a DBB, DB, or even a CMR project. Graphics displaying the contractual relationships among the major stakeholders and their lines of communication are presented in Figures 2.1 through 2.3 to assist the reader in putting the contents of this report into proper context. Note that the lines of communication shown in the figures represent the ability to exchange information through formal and informal requests for information among the various entities in the project. Design-Bid-Build (DBB) DBB is the traditional project delivery method. In this method, an owner retains a designer to furnish complete design services and then advertises and awards a separate construction con- tract that is based on the designer’s completed construction documents. The owner is responsi- ble for the details of design and warrants the quality of the construction design documents to the construction contractor. Figure 2.1 shows that the owner is situated squarely between the designer and the builder in the DBB project delivery method. In DBB, the owner “owns” the details of design during construc- tion and, as a result, is financially liable for the cost of any design errors or omissions encountered in construction; this is called the “Spearin Doctrine” (Mitchell 1999). The construction phase of DBB projects is generally awarded on a low-bid basis. There is no incentive for the builder to min- imize the cost of change orders in this delivery method. In fact, there can be quite the opposite effect. A builder who has won a project by submitting the lowest possible bid may need to look to post-award changes as a means of enhancing profit on the project. One author states that the defining characteristics of DBB are as follows (Bearup, Kenig, and O’Donnell 2007): • There are separate contracts for design and construction. • Contractor selection is based entirely on cost. • Design documents are 100% complete. Background and Definitions 11 (Adapted from American Institute of Architects 1996.) Owner Designer Builder Contracts Communication Figure 2.1. Design-bid-build.

Despite the general definition of DBB given above, DBB projects can also be awarded on a negotiated basis and a best-value basis (Scott et al. 2006); however, DBB transit projects awarded in either of these two ways usually require FTA approval and frequently violate local laws. For projects awarded on a negotiated basis or a best-value basis, the probability that the project will be awarded to a builder who has submitted a mistakenly low bid is reduced. Additionally, in both cases, the builder will be motivated to complete the project in a such a way that it be invited back to do the next negotiated contract or that will reflect well in the next best-value selection. Regard- less of the award method, DBB involves less builder input to the design than DB or CMR. Thus, the owner must rely on the designer or agency CM (and not the builder) for a constructability review, if there is any at all. Nonetheless, in this method the owner has full control over the details of design, which may be a requirement for some complex projects. DBB is also characterized by the greatest amount of familiarity in both the design and construc- tion areas. All qualified designers can compete for the design without restriction. Additionally, all constructors who can furnish the requisite bonding and meet any agency prequalification crite- ria are also able to compete without constraint. Design subconsultants and construction trade subcontractors are also able to compete with minimal restrictions. Finally, as DBB is generally viewed as the traditional project delivery method in the United States, it is well understood and accepted by owners and members of the design and construction industries. CMR or Construction Manager/General Contractor (CM/GC) CMR projects are characterized by a contract between an owner and a construction manager who will be at risk for the final cost and time of construction. In this agreement, the owner authorizes the construction manager to handle the construction phase and give inputs during the design development. The idea of CMR is to furnish professional management of all phases of a project life to an owner whose organization may not have those capabilities (North Carolina State Construction Office 2005). Typically, CMR contracts contain a provision in which the CMR stipulates a guaranteed maximum price (GMP) above which the owner is not liable for payment. Often, these contracts include incentive clauses in which the CMR and owner can share any cost savings realized below the GMP. Some states, like Oklahoma, take the GMP and con- vert it to a firm-fixed price contract and administer the construction as if it were a traditional DBB project thereafter (AIA 2005). CMR contracts can contain provisions for the CMR to han- dle some aspects of design, but generally, the owner retains the traditional responsibility by keep- ing a separate design contract and furnishing the CMR with a full set of plans and specifications upon which all construction subcontracts are based (see Figure 2.2). The CMR will usually be 12 A Guidebook for the Evaluation of Project Delivery Methods (Adapted from American Institute of Architects 1996.) Owner Designer Trade Subs CM At-Risk Contracts Communications Figure 2.2. Construction manager at risk.

paid for furnishing preconstruction services such as cost engineering, constructability review, and development of subcontractor bid packages. According to AGC (2004), the defining char- acteristics of the CMR are the following: • The designer and the CMR hold separate contracts with the owner. • The CMR is chosen on the basis of criteria other than just the lowest construction cost, such as qualifications and past performance. According to Bearup, Kenig, and O’Donnell (2007), additional defining characteristics are the following: • The CMR contracts directly with trades and takes on “performance risk” (cost and schedule commitments); • The schedule allows for overlapping design and construction; • The owner procures preconstruction services from the CMR; and • The owner expects the CMR to provide GMP and to commit to a delivery schedule. A final defining characteristic, noted in AIA’s “Construction Manager at-Risk State Statute Compendium,” is that “transparency is enhanced, because all costs and fees are in the open, which diminishes adversarial relationships between components working on the project, while at the same time eliminating bid shopping” (AIA 2005, p. 1). Constructability and speed of implementation are the major reasons that an owner would select the CMR method (3D/International, Inc. 2005). Additionally, CMR greatly facilitates phased construction when that is a requirement for a given project. Unlike DBB, CMR brings the builder into the design process at a stage in which definitive input can have a positive impact on the project. “The CM[R] becomes a collaborative member of the project team. Preconstruc- tion services include budgeting, cost estimating, scheduling, constructability reviews and value engineering studies.” (3D/International, Inc. 2005, p. 4). In CMR, the CM essentially becomes the general contractor at the time the GMP is established. While some experts attempt to distin- guish between CMR and CM/GC, due to perceived levels of risk, many agencies use these terms more or less interchangeably.2 The CMR can and is expected to provide realistic project cost esti- mates early in the project lifecycle. It is anticipated that after a certain amount of the design is complete and the project is sufficiently defined, the owner will enter into a contract with the CMR for providing construction services. Many states reserve the right to go out for bid if they think that the CMR’s price is not competitive (Minchin, Thakkar, and Ellis 2007).3 The timing of GMP negotiations varies among different agencies. In many cases, at least 60% of the design is completed before a GMP is established. In some cases, the design is 80 to 90% complete before a GMP can be effectively negotiated with the CMR. The timing of GMP negotiations depends on project complexity, agency rules, and external conditions such as inflation and the expected level of competition among subcontractors. In general, the CMR may feel that committing to a GMP while all the details of the design are not defined may involve incurring undue risk. Also, some agency rules may hamper early GMP negotiations. For example, if an agency insists on requiring a fully open competition for hiring of subcontractors, then negotiating an early GMP may be more difficult because some subcontractors may be reluctant to give their prices without Background and Definitions 13 2According to AGC (2004), there has been some confusion about the terms CM at-risk and CM/GC because of the assumption that the phrase “at-risk” connotes cost guarantee. Even if there are no cost guarantees, the CM is still at risk because the CMR holds the trade contracts (warranting the performance of the work). Because of this, some users choose to avoid the debate over the term “at-risk” and instead use the term CM/GC (p. 8). 3There are two types of CM arrangements, namely agency CM and CM at-risk. Our emphasis in this work is CM at-risk. Agency CM is not a project delivery method because the CM is not contractually responsible for deliv- ering the project. The role of agency CM is purely advisory, and thus, the agency CM is usually not at risk for the cost and schedule of building the project.

a 100%-complete design. This in turn makes the CMR hesitant to provide a reasonable GMP for fear that money will be lost if the subcontract bids come in too high. As the design selection process in CMR virtually mirrors the design process in DBB, imple- menting CMR does not inherently restrict competition among designers and design subconsul- tants (AIA 2005). Owners, at their own discretion, occasionally require the designer in a CMR project to have previous CMR experience, which may result in fewer qualified proposers. As the constructor is selected on the basis of qualifications and past performance and must also have the capability to perform preconstruction services, CMR project delivery can constrain compe- tition to those constructors that have previous CMR experience. Most public CMR laws require competitively bidding out the construction trade subcontract work packages. The central idea of CMR is to get the advantage of price competition in the subcontract work packages combined with the qualifications-based selection of the GC as CMR. Design-Build (DB) Design-build is a project delivery method in which the owner procures design and construc- tion services in the same contract from a single legal entity referred to as the design-builder. A variety of methods exist for selecting the design-build constructor. Common methods are the one-step and the two-step processes. The one-step process provides for competitive evaluation of technical proposals, with the contract award decision based on best value to the owner agency. The determination of best value is based on a combination of technical merit and price. The two- step process separates the technical proposal from the price. This method typically uses request for qualifications (RFQ)/request for proposal (RFP) procedures rather than DBB invitation-for- bid procedures. There are a number of variations on the DB process, but all involve three major components. First, the owner develops an RFQ/RFP that describes essential project requirements in performance terms. Second, proposals are evaluated. Finally, with evaluation complete, the owner must engage in some process that leads to contract award for both design and construc- tion services. The DB entity is liable for all design and construction costs and must usually pro- vide a firm, fixed price in its proposal (El Wardani, Messner, and Horman 2006; Ibbs, Kwak, and Odabasi 2003; and Graham 2001). Figure 2.3 shows that from the owner’s standpoint, DB simplifies considerably the project’s chain of responsibility. As in CMR, the builder has early constructability input to the design process. As the owner no longer owns the details of design, the owner’s relationship with the 14 A Guidebook for the Evaluation of Project Delivery Methods (Adapted from American Institute of Architects 1996.) Owner Design- Builder Designer Builder Contracts Communication Figure 2.3. Design-build.

design-builder must be based on a strong degree of mutual professional trust (Beard, Loulakis, and Wundram 2001). The design-builder literally controls this project delivery method. As a result, the DB project delivery method has proven to be highly successful in compressing the proj- ect delivery period and is therefore often used for “fast-track” projects (SAIC, AECOM Consult, and University of Colorado at Boulder 2006). Bearup, Kenig, and O’Donnell (2007) state that the defining characteristics of DB are as follows: • A single point of responsibility, • A schedule that allows for overlapping design and construction, • A design-builder that furnishes preconstruction services during the project design, and • An owner that expects the design-builder to provide a firm, fixed price and to commit to a delivery schedule. DB creates the greatest constraint on competition in that all parties to the DB contract are selected using qualifications and past performance as major selection factors. Because the owner transfers responsibility for all design and construction in the DB contract, it also loses the abil- ity to foster competition between design subconsultants and construction trade subcontractors. There is typically no requirement to competitively bid for subcontract work packages, and often the scale, complexity, and speed at which DB projects are executed precludes firms with no DB experience from being able to participate. Additionally, as the contract is awarded before the design is complete, DB can also create an unfavorable risk environment for subcontractors whose cost-estimating systems lack the sophistication to price work without completed con- struction documents. There are many variations on the DB method. Design-build-operate-transfer, design-build- operate-own (sometimes called lease-back), and DBOM, all require the DB contractor to remain with the project after construction is complete. DBOM is very similar to DB except that the DBOM contractor assumes the operation and maintenance risks of the project and is responsi- ble for operating the new facility according to a set of regulations and codes for a determined duration (Wiss, Roberts, and Phraner 2000; Kessler 2005). Statutory Authorization of Delivery Methods in Various States DBB has traditionally been used throughout the United States, and all 50 state codes have given full authority to transit agencies to use this method in their projects. Alternative delivery methods do not have this clear statutory support. Some states do not allow transit entities to use alternative delivery methods, some have given one-time authority to use an alternative method for a special project, a group of states have put some limits on the application of alternative deliv- ery methods, and a few states require transit agencies to obtain extra approval in order to use alternative methods. Developing pilot programs is a common approach in some states for imple- menting previously unauthorized project delivery methods, particularly DB. In order to update information on the legal status of alternative project delivery methods in various states, a thorough literature search was conducted on the laws of all 50 states. Several relevant keywords were searched using the LexisNexis search engine. All the state codes and statutes that deal with project delivery in transportation projects were examined. The results were then compared with the existing surveys of legal codes available in the literature (e.g., see Smith and Davis 2006 and AIA Minnesota 2006). The research herein shows that 37 states per- mit the use of DB in their transportation projects, leaving the agencies of 13 states without the authority to do so. The application of CMR is not authorized in 31 states, and only 14 states have Background and Definitions 15

fully authorized DOTs to use this delivery method. Five states allow the use of CMR with some restrictions or after obtaining extra approvals (Ghavamifar and Touran 2008). It should be noted that the laws governing the legality of alternative project delivery methods are evolving, and therefore the information given herein on the legality of alternative project delivery methods should be understood as subject to change. Also, while some state DOTs are permitted to use alternative delivery methods, it is not clear whether those states’ transit agencies are allowed to use alternative delivery methods. The purpose of the literature search was to provide an overall picture of the legal status of using various project delivery methods for transportation projects in the United States at the time this report was prepared. Each public agency considering the use of a specific delivery method should check the legality of the method carefully. FTA’s requirements for third-party contracting, described in Circular 4220.1E (FTA 2003), are sufficiently flexible to allow the agencies to select their contractors through competitive bidding and/or competitive proposal/RFP (price and other parameters considered). For DBB, Circular 4220.1E allows the procurement of services through sealed bidding or competitive negotiations. For DB, the grantees must procure DB services through qualifications-based competitive pro- posal procedures. So it seems that if a specific state’s laws allow an alternative project delivery method, the federal regulations will not prevent the agency from undertaking such procurement. Existing Selection Approaches for Project Delivery Methods Selection of the appropriate alternative project delivery method is a complex decision- making process. The decision should be made as early in the design phase as possible, preferably in the project scoping process and certainly before the final construction estimates for the project are ready. The decision will be made when the owner still has little information about the exact outcome of the project and the project plans are not detailed enough to be reliable grounds for judgment about the project. In this situation, having a framework for decision-making is vital for transit projects. This framework should be simple, comprehensive, rational, and objective. The literature review of this research report shows that some experts have concentrated on this issue and have developed a list of criteria and some decision-making frameworks (Debella and Ries 2006; Garvin 2003; Gordon 1994; Ibbs, Kwak, and Odabisi 2003; Konchar and Sanvido 1998; Mahdi and Al-Reshaid 2005; Oyetunji and Anderson 2006). Several of these researchers have chosen specific projects and have based their selection methodology on the characteristics of those projects. The relevant literature can be divided into two groups: (1) literature that compares project delivery methods on the basis of observed performance measurements collected from a group of projects and (2) literature that provides a list of criteria and a framework for decision-making. One of the best examples of the first kind of literature is a paper by Konchar and Sanvido (1998) in which a set of criteria is defined for a performance comparison of different delivery methods (i.e., DB, DBB, and CMR) in 351 building projects. These criteria are mostly objective and measurable, such as cost growth, construction speed, and schedule growth. Some criteria are also defined to incorporate the quality performance of the delivery methods, such as diffi- culty of facility start up, number and magnitude of call backs, and operation and maintenance costs. According to Konchar and Sanvido (1998, p. 9), “when all other variables were held con- stant, the effects of project delivery system indicated design-build projects to be at least 5.2% less than design-bid-build projects and 12.6% less than construction management at risk projects on average in terms of cost growth.” Konchar and Sanvido (1998) divided the projects into six dif- ferent groups (e.g., light industrial, complex office, and heavy industrial) in order to see clearer trends in each group. Taking this into account, the paper does not have enough data to distin- 16 A Guidebook for the Evaluation of Project Delivery Methods

guish between the performances of different delivery methods in transit projects. However, two studies comparing DB and DBB project performance in the federal building sector did make direct comparisons (Allen, Gransberg, and Molenaar 2002; Gransberg, Badillo-Kwiatkowski, and Molenaar 2003). One study compared 54 DBB projects with 34 DB projects and discovered that DB projects had 16.4% less cost growth and 19.0% less time growth than similar DBB proj- ects (Gransberg, Badillo-Kwiatkowski, and Molenaar 2003). Another study, which looked at 110 Navy projects, also found that DB projects performed more efficiently, with 18.0% less cost growth and 60.0% less time growth (Allen, Gransberg, and Molenaar 2002). Additionally, a recent NCHRP study of best value contracting also furnished a direct comparison of the per- formance of transportation project delivery methods (Scott et al. 2006). While that study did not include CMR projects, it did include DBB projects awarded on a best value basis, which paral- lels the CMR delivery method. The NCHRP study found that DB projects had 4.7% less cost growth and 9.3% less time growth than DBB. Best value projects had 2.0% less cost growth and 18.5% less time growth than DBB. Other researchers, such as Debella and Ries (2006) and Ibbs, Kwak, and Odabisi (2003), have used a methodology similar to that of Konchar and Sanvido (1998), but they have narrowed down the scope of their research either to special kinds of proj- ects or fewer performance measures. The second kind of literature mentioned above, literature that provides a list of criteria and a framework for decision-making, has focused on the decision-making process. This literature proposes mechanisms for decision-making and defines the necessary criteria and frameworks so that the most important project parameters are identified and used in the decision-making method. The frameworks are primarily intended to be simple, rational, and comprehensive. They range from basic flowchart methods (Gordon 1994) to more sophisticated processes based on methodologies such as multiple linear regression, the Analytical Hierarchy Process (AHP) (Mahdi and Al-Reshaid 2005), or the Simple Multi-Attribute Rating Technique with Swing Weights (SMARTS) (Oyetunji and Anderson 2006). Gordon (1994) created a procurement method selection model that uses a flowchart for select- ing the best contracting method. Within the flowchart are a number of drivers that direct the owner’s attention to the most important issues in project delivery method selection. A multi- media education compact disc and delivery selection tool have been developed (Loulakis 2005). The tool integrates training on project delivery selection systems with a matrix-style decision framework that owners can complete to make an informed delivery selection. Skitmore and Mars- den (1988) presented a multi-attribute analysis technique and a discriminant method for select- ing delivery methods. The multi-attribute method uses utility factors to evaluate the suitability of a delivery method with respect to a client’s priority criteria. Kumaraswamy and Dissanayka (1998) propose a client advisory system with an expert system front end that will gather project information and model the project profile to generate a list of delivery options. Finally, Oyetunji and Anderson (2006) use a SMARTS approach for delivery selection. The approach utilizes a matrix that has 20 criteria, each with a given weight. The owner rates these criteria and goes through the required calculation, which gives a single rank to each delivery method. The deliv- ery method with the highest rank should be chosen for the project. Looking at both kinds of literature, one finds that many of the important parameters that affect the decisions early in the project fall into one of four groups: project-related parameters, agency- related parameters, legal parameters, and lifecycle issues. Project-related parameters are those parameters that pertain to project duration: estimated cost, quality level, risks, limits on schedule growth, complexity, and so forth. Agency-related parameters mainly consist of the status of the agency; the role of a project in the strategies of the agency; and the organization of the agency, i.e., availability of funds, sophistication of the agency’s employees, flexibility needs in the construction phase, level of risk assumption, importance of preconstruction services, and quality level expecta- tion. The legal parameters mainly cover legal and contracting issues, such as statutory authority to Background and Definitions 17

use alternative project delivery methods, the level of competition in the market, and the permits needed for the project. Lifecycle issues cover the costs of maintaining and decommissioning the facility and the ability to minimize the energy and environmental effects of the project. In the parameters mentioned above, the ability to transfer the risks of a project to entities other than the owner is a characteristic that is related to both the project and the owner agency. This parameter includes the level of risk and uncertainty of the project and the ability of the owner to assume the risks or transfer them (risk-averse or risk-prone agency). Project delivery methods have different mechanisms for risk distribution among the entities involved. In summary, the existing body of knowledge in the area of project delivery, along with specific information col- lected on transit projects during interviews, provides a solid foundation for developing a new selection system for project delivery methods that is tailored to the needs of transit owners. Timing of Project Delivery Method Selection Transit projects, especially those that receive federal funds, follow several steps during their development. These steps can be summarized as follows: • Alternative Analysis—Draft Environmental Impact Statement (AA/DEIS), • Final Environmental Impact Statement (FEIS), and • Full Funding Grant Agreement (FFGA). The first two steps roughly coincide with conceptual design (5 to 15% of the design effort) and preliminary engineering (25 to 30% of the design effort). The timing of the FFGA, which repre- sents the federal government’s commitment to fund the project, depends on the project deliv- ery method; the FFGA can come at the end of preliminary engineering or at final design. In selecting a project delivery method, the owner should realize that the window of opportunity to select some methods will close as the project moves through various stages of development. Table 2.1 maps project delivery method selection with project development phase. It can be seen that selecting a project delivery method other than DBB should be done relatively early. Most of the benefits can be realized by engaging the constructor as soon as possible. The decision point for project method delivery selection should not be confused with the time that the constructor is engaged. As an example, an owner may decide to engage a DB contractor at the end of preliminary engineering or even later in order to clarify the project scope and reduce the uncertainty. However, the owner should have decided on the type of delivery (e.g., DB) much earlier, so that the design documents can be developed to properly accommodate the type of delivery method. 18 A Guidebook for the Evaluation of Project Delivery Methods Table 2.1. The timing of project delivery method selection. Project delivery method At the end of conceptual design At the end of preliminary engineering At the end of final design Construction DBB NF CMR NF DB NF NF DBOM NF NF = Desirable = Feasible NF = Not feasible