Construction Manager-at-Risk Project Delivery for Highway Programs (2010)

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.

65 INTRODUCTION The term “guaranteed maximum price” (GMP) is often mis- understood. “Most Owners see having a Guaranteed Maximum Price (GMP) as equivalent to having a Stipulated Sum Cost” (Strang 2002). In the eyes of the uninitiated, the word “guar- anteed” implies that the owner will never have to pay more than the GMP. However, in CMR the GMP amount is a num- ber that corresponds to a quantified scope of work expressed in the design documents at the time the base cost estimate was completed. Therefore, if a substantial change in scope occurs, the CMR is due fair compensation for that cost of increased work. This leads to the impression articulated by Strang that the owner effectively shifts the risk for the total cost of the project to the contractor, regardless of the nature of the realized risks. The literature refutes this idea as patently untrue and supports the idea that it is important to lay out the details of how the GMP will be assembled and ultimately established in both the solicitation documents and the contracts for preconstruction and construction services (Armstrong and Wallace 2001; Strang 2002; Alder 2007; Bearup et al. 2007; Martinez et al. 2007; Trauner 2007; Ladino et al. 2008; Lee 2008). Kwak and Bushey (2000) furnish a very simple definition for the components of a GMP: “The GMP is composed of work, overhead, profit, and a contingency.” Breaking these elements out assists the owner to understand the relative magnitude of each component and furnishes a framework from which the owner can assess the reasonableness and realism of each element. A typical GMP clause from a transportation project RFP reviewed in the content analysis defines the GMP in the following manner: The Contractor’s Fee as an established percentage shall be applied to the Cost of the Work plus contingency. The sum of the Cost of the Work plus contingency, plus Contractor’s fee shall establish the basis of the Guaranteed Maximum Price (GMP) for the project prior to construction start (“CM/GC Fee Structure Sample” 2007). This definition is the simplest possible GMP: direct cost, contingency, and fee, which includes the contractor’s general conditions. Figure 18 illustrates the possible components of a CMR GMP based on those found in the case study projects. The figure is meant to be inclusive not restrictive. Therefore, some of the specific elements shown are not present in every CMR contract. However, the elements that are common to most transportation CMR project GMPs are identified as such. Figure 18 can be turned into a cost model for a specific project. Modeling the costs in context with the available budget before making fundamental design decisions is imperative to the success of horizontal CMR projects (Ladino et al. 2008). The model can then be used to validate the owner’s project budget at a point where design effort is not lost and where the CMR can furnish up-to-date market information that will help achieve the project’s function within the owner’s available funding (Ladino et al. 2008). In Figure 18, the least compli- cated GMP would have the following elements: • Project direct costs – Subcontract work package costs – CMR self-performed work package costs • Indirect costs: CMR’s general conditions/overhead costs • Profit: Percentage mark-up or lump sum fee • CMR project contingency. The industry has a number of variations on the basic GMP. Some add a separate pool for potential subcontractor-installed material purchases by the CMR to lock in the cost of volatile construction materials, such as asphalt, steel, etc., before the design has reached a point where subcontractor bids can be requested. Others divide the total project contingency into two or three parts to furnish additional cost control based on the wishes of the owner. Therefore, having defined the basic components of a CMR GMP, each will be discussed in detail. PROJECT DIRECT COSTS Developing the direct cost portion of a CMR GMP is highly dependent on the level of design development that has been completed at the time the GMP is established. These costs directly reflect the quantities of work. A design work package that is nearly finished can be estimated with greater confidence in the quantities shown on the plans than one that has just been started. Additionally, costs are also dependent on whether the work is being performed by the CMR’s forces or a sub- contractor. It is also dependent on the constraints imposed on the CMR regarding the selection of subcontractors, as well as how much, if any, of the work the CMR is allowed to self-perform. Guaranteed Maximum Price Timing Issues Selecting the point in time where a GMP is established is an extremely important decision. A GMP set late in the design CHAPTER SIX PROCEDURES FOR ESTABLISHING THE GUARANTEED MAXIMUM PRICE

process will have less contingency included than one required early in the design process. One of the reasons cited for using CMR rather than DB project delivery for large complex proj- ects is to reduce the “premium” the owner pays for requiring a firm fixed price at an early stage of design completion (Trauner 2007). The case study analysis found that nine of ten agencies required the GMP before 100% design. However, four waited until subcontractor bids were received on the major packages in the job, and two allowed the CMR to set the timing based on its assessment of the risk of quantity growth in those packages that were not complete. The other four agencies set the timing contractually. Table 29 is a collection of different GMP timing clauses found in the content analysis and is provided to demonstrate the range in possible options in use in the industry. The ten clauses in the table show a range of timing from as early as 15% design in Clause 9 to as late as shown in Clauses 4 and 5, which could be interpreted to mean after 100% design completion. 66 Construction Manager-at-Risk Subcontracting Issues Three issues regarding subcontracting were found in the case study analysis. The issues are postulated as the follow- ing questions that an agency would answer when it selects CMR project delivery: • Is the CMR allowed to self-perform a portion of the work? • If so, is there a limitation/constraint on the amount of the project the CMR can self-perform? • What constraints will be placed on the CMR’s ability to select subcontractors? The first question is a matter of policy rather than any underlying technical or contractual reason. The case study projects came from 10 different public agencies and only Management Reserve (controlled by an ownerís entity that is outside the owner’s project team) Owner’s Contingency (major design changes, scope creep, unforeseen conditions, force majeure, etc.) CMR’s Contingency (material & labor escalation, subcontractor availability, market- based issues, etc.) CMR’s Fee (construction profit) CMR’s General Conditions (overhead, taxes, permits, job site overhead, etc.) CMR’s Direct Project Cost (labor, materials, equipment for self-performed work packages) CMR’s Cost of Early Material Purchases to be installed by subcontractors. Subcontract Work Packages (labor, materials, equipment for subcontracted work packages) CMR Preconstruction Services Fee Design Fee/Cost O w ne r’s T ot al C on tra ct C os t To ta l P ro jec t C on tin ge n cy To ta l Pr ec o n st ru ct io n Co st G u ar an te ed M ax im u m P ric e Mandatory Component Optional Component FIGURE 18 Guaranteed maximum price model.

67 one would not permit the CMR to self-perform; that was the Fairbanks Airport project. The interview found that previous experience in Alaska with non-transportation projects created a perception that allowing the CMR to self-perform created a conflict of interest and was viewed as a disadvantage by the subcontracting community. This was because the CMR had to compete against industry for the work it wanted to self- perform and was thought to have “inside” knowledge. The issue was put to rest by not allowing the CMR to self-perform. Washington State had a successful protest of self-performance by a subcontractor on a building project that led to a change in the law to specifically limit the CMR to self-perform no more than 30% of the project (Septelka and Goldblatt 2005) and a report done by the city of Seattle (Denzel et al. 2004) confirmed the perception of unfairness when it stated: “The City should also recognize that the announcement of the intent to bid for “self-performed” work by a GC/CM can discourage other subcontractors from bidding, either because they per- ceive the GC/CM has an unfair advantage, or because they do not want to alienate the GC/CM who may provide future work for them. This can undermine the competitive process.” CMR Self-Performance Limits Another common premise for not allowing the CMR to self-perform is the idea that self-performance degrades the CMR’s ability to act in the owner’s best interest by imposing an internal need to realize profit on work packages that have not been assigned to subcontractors. Strang (2002) describes it this way: “For the general contractor the change will be the advocacy of the client’s position . . . CM At-Risk is still largely a position of representing the Owner’s interests, and if it is pursued as simply a general contract it will leave the Owner dissatisfied and the CM without future work.” The U.S. construction industry has evolved to a point where some general contractors routinely complete projects where 100% of the work has been subcontracted, whereas others have maintained the ability to self-perform. Thus, a policy or law that either disallows self-performance or caps the percent- age of self-performed work creates an artificial constraint on industry and unnecessarily reduces the pool of qualified competitors. The survey response from Washington State indicated that the state’s legislated 30% cap on CMR self- performance created industry resistance to using the DOT’s CMR authority because most contractors in that state would choose to self-perform a larger proportion of the project. On the other hand, some public agencies have legislation that requires a minimum level of self-performance on construction contracts (Brinkman 2007; “Supplemental Instructions for Form IC 730 . . .” 2008). That effectively eliminates those contractors that sub out all the work. The content analysis found 17 cases where CMR self- performance was mentioned. Three of those cases included a minimum percentage (45% or 50%) for CMR self-performed work. The others furnished no minimum or maximum limita- tions. However, limits may have been imposed by reference and as a result would not have been found in the content analysis. The city of Seattle report (Denzel et al. 2004) offers three reasons to permit the CMR to self-perform, terming them “significant benefits”: • The GC/CM can exert better control of the project schedule if they are self-performing parts of the work that are essential to the critical path for the project, espe- cially fundamental structural elements such as concrete or framing on which other subcontractors’ work depends; • Subcontractors may prefer that the GC/CM have a stake in the performance of the work; and • The ability to self-perform can be part of what makes the job attractive to top-notch GC/CM firms, since they customarily perform the type of work they bid for and have crews on hand to fulfill those responsibilities (Denzel et al. 2004). The previous discussion leads to the inference that competition can be enhanced if there is no constraint on the level of self- performed work imposed on the competing CMRs. Thus, a single answer that satisfies the first two questions would be: The CMR be allowed to self-perform as much of the work as it believes is appropriate for a given project. The case study results found limits that were both minimum and maximum percentages. Six of the seven case study high- way agencies had a limit of some form. The average of these limits without regard to their specific nature is 47%, with a low of 30% and a high of 70%. All of these limitations were imposed by legislation. None of the non-highway cases limited the CMR’s self-performance percentage. Therefore, from the information collected in this study, the imposition of a self-performance limit in any form appears to have no impact on the success or failure of the project. All the inter- viewees expressed the perception that CMR was working well on its respective projects. The city of Phoenix, Arizona, has completed more than 200 CMR projects while requiring a minimum of 45% CMR self-performance (Brinkman 2007). Therefore, if the perception of unfairness was indeed a real danger to competition, then the city of Phoenix would have been unable to attract competent CMRs and would have prob- ably abandoned the project delivery method altogether. The same argument applies to the UDOT, who at this writing has successfully completed 16 CMR highway projects and sees CMR as a major tool in its procurement toolbox (Alder 2007). Subcontractor Competition/Selection Constraints The third question in this section deals with the CMR’s control over the selection of subcontractors. The answer depends on the enabling legislation as well as state or agency policies regarding competitive bidding. One documented benefit of CMR is the ability to “preserve” competitive pricing by requiring the CMR to solicit bids from the trade sub- contractors (Uhlik and Eller 1999; Strang 2002; Alder 2007; Martinez et al. 2007). The case studies ran the gamut from

no restrictions on the subcontracting process to requiring the CMR to take competitive bids and award to the low bidder. Oklahoma has an extreme example of required competi- tive bidding in its code governing the construction of public buildings: Because the construction manager is “at risk” for the total contract amount, they are permitted to self-perform portions of the work, provided that they competitively bid the work as a lump sum (each work package) under the same terms and conditions as the other bidders . . . the construction manager will award to the lowest responsible bidder (Oklahoma 2007). Under this system, the CMR declares which work packages it wants to perform and literally bid against the industry for the right to self-perform. Thus, the CMR not only bids against its potential subcontractors for the right to self-perform, it also awards to the low bidder without respect to the sub’s ability to perform. This creates a problem that may have a detrimental impact on the project in that the code does not permit prequalification of subcontractors. Additionally, this prevents the CMR from seeking expert help during pre- construction because contacting a potential subcontractor for technical information or current pricing creates the perception of unfairness previously discussed. Hence, the CMR’s ability to use its professional connections to procure real-time pricing information is virtually eliminated. This policy also exists in Arizona and Florida. One Arizona RFQ contained the follow- ing statement: “The CMR may also compete to self perform portions of the work.” Therefore, to put this in perspective, the urge to require the CMR to select subcontractors on a low-bid basis and/or restrict subcontractor prequalification is motivated by the desire to “preserve” competitive pricing for most of the project’s work packages because the GC is now selected on a basis of qual- ifications. Both the literature and the case study output confirm that this motivation is misguided. The Construction Finan- cial Management Association (CFMA) conducted a study in 2000 comparing CMR costs with DBB costs and “showed no significant difference in costs to the owner between CMR and traditional low-bid lump sum GC-based contracts” (CFMA Annual Financial Survey, 2000 2000). This study was cited by two other studies that came to the same conclusion (Strang 2002; Cunningham 2005). That eight of ten case study proj- ects allowed the CMR to pick and choose its subcontractors without constraint demonstrates that practitioners have recognized the inherent value to their projects of the CMR’s long-standing commercial relationships. Additionally, Table 1 shows that the two significant CMR advantages that are most closely tied to subcontracting (“early knowledge of costs” and “ability to bid early work packages”) were each cited by 10 of the 15 authors, whereas “competitive bidding possible” was only cited 4 times. Therefore, the intersection of these three lines of information leads to the conclusion that owners who select CMR project delivery to control costs can allow their CMR the ability to select the subcontractors that it prefers. This is concluded for three reasons: 68 1. To get real-time pricing information, the CMR is able to communicate with the subcontractors it knows during preconstruction. 2. To obtain real-time technical information about best practices for subcontractor trade means and methods, the CMR is also able to communicate with the sub- contractors it knows to be subject matter experts during preconstruction. 3. Studies have shown that competitive pricing is “pre- served” without competitive bidding. Therefore, requir- ing the CMR to award subcontractor work packages to an open field of competitors does not appear to save money. Amplifying on the third point, agencies are reminded that in a DBB project delivery that GC is typically allowed to select a subcontractor whose price is marginally higher than the lowest quote. This is especially true if a given sub has a strong record of good performance on the GC’s past projects. A Massachusetts study that reviewed the cost of a very restrictive procurement law that required subcontractor selection on price alone, mirroring the Oklahoma CMR law, found that DBB construction costs increased as a result of excessive construction cost growth and claims. There were no long-standing relationships between the GC and its sub- contractors and as a result, the GC did nothing to discour- age the subcontractors’ claims (Gransberg 1999). Another study that compared project cost performance in the build- ing construction sector found that CMR costs were margin- ally less (1.5%) than costs for DBB projects (Sanvido and Konchar 1999). These prove the old adage that “lowest bid does not always equal lowest final cost” (Ellicott 1994). Therefore, this bit of wisdom from DBB could be revised for CMR to state: Requiring competitive subcontractor bidding in CMR project delivery does not equal lowest possible GMP. Construction Manager-at-Risk Self-Performance Costs First, it is to be noted that if the CMR is not allowed to self- perform work packages that these costs will not necessarily be zero. It is not uncommon for the CMR to furnish certain types of materials and equipment to the project to avoid sub- contractor mark-ups on those items. For example, if several subcontractors will require the use of a scissor-lift and scaf- folding to complete their work packages, the CMR will furnish those items at its own expense. It is also possible for the CMR to create a small work package for project clean-up and punchlist, planning to do these bits itself at project comple- tion rather than try to police all the subcontractors as they come and go. Therefore, this category of GMP costs might be better thought of as “CMR direct costs” to account for mis- cellaneous labor, equipment, and materials provided to the project by the CMR that are ultimately incorporated into or used to complete subcontractor work packages.

69 On projects where CMR self-performed work is not pro- hibited, these costs will relate to direct costs of labor, equip- ment, and materials in much the same manner as they would be estimated for a DBB project. The important difference is that the agency delineate between direct costs and the CMR’s general conditions and overhead costs. The way to ensure that controversy does not arise during GMP negotiations is to publish a schedule of cost types that shows the CMR in exactly which category the agency wants various costs to be collected. Grand County, Colorado, has an excellent “CM/GC Fee Structure Sample” (2007), which is shown in part in Table 30. The example is not provided as an all-inclusive and defin- itive practice, but rather it shows how the agency can reduce the amount of potential controversy during GMP negotiations by structuring the GMP assembly process. This approach will also accrue the benefit of having the agency’s historical CMR cost data in a form that facilitates using it to estimate future CMR projects. It also gives upper management decision makers a reliable benchmark against which to measure the realism and reasonableness of the engineer’s estimates before advertising CMR projects. Lastly, it permits the agency to be able to quantify the cost impact of management decisions to shed specific risks such as obtaining permits in a manner where standard operating policy can be changed based on hard data rather than perceptions. Construction Manager-at-Risk Early Material Package Purchases The final common direct cost in CMR is the purchase of materials at the earliest possible point in design development Services Pre- Construction Fee Construction Fee General Conditions/ Overhead CMR Direct Cost Sub- Work Package Costs Owner Expense Contractor Home Office Staff and Services Corporate executives X X Safety manager X X Contractor Job Site Office Staff and Services Project manager X Engineering and layout X Project Cost Control Validation of project budget X CPM schedule with updates X X Subcontractor Selection Participate in setting sub prequal. X Set work package plans X Analyze subcontractor bids X X Contract Document Coordination Review for constructability X Coordinate design packages with work packages X Review of proper phasing X X Quality Control Prepare QC plan X Arrange QC testing and inspections X Independent verification testing X Acceptance testing X Insurance, Bonds, and Permits Builder’s risk X Utility development fees X X State DEQ general permit X Job Site Facilities and Services Temporary field office X Project signs X Temporary Facilities Temporary telephone X Final clean-up X X Project Work Package #1—Sitework Utility relocations X Earthwork—drainage X Project Work Package #2—Bridge A Drill piers X Precast concrete installation X After “CM/GC Fee Structure Sample,” Grand County, Colorado (2007). Shows all categories and typical examples of cost types. DEQ = Department of Environmental Quality. TABLE 30 GRAND COUNTY, COLORADO, SAMPLE CM/GC FEE STRUCTURE EXTRACT

to fix volatile construction material prices and reduce or eliminate the total amount of escalation contained in the GMP. Often these early material procurement packages are for work that will be completed by subcontractors. The UDOT uses this technique in its CMR contracts. On the I-80 case study project, it established a separate GMP for materials and thereby eliminated the need to add escalation to material pricing in the direct costs. The Memphis Airport did a similar thing for the steel that was in its project and estimates it had a significant savings as a result. The major issue with early material packages is that the associated design features need to be fairly complete to generate the quantities necessary to place the order. Memphis accomplished this by putting “date-certain” milestones for the engineer to bring the design for the steel procurement package to a point where reasonably accurate quantities could be surveyed. They then permitted the CMR to include an item in its contingency to pay for any changes that might occur between the time the steel was ordered and the final steel design. The steel was erected by a subcontractor who furnished labor and equipment only. Given the successful application of this approach in the case study projects, it appears prudent to consider including early material packages routinely on CMR projects. To do so necessitates a change to standard design contracts. These early packages are only as beneficial as the designer can make them by approaching the design process with a plan to get to the bottom of those features that will require large amounts of price volatile materials. It will also require that the preconstruction services contract be well-coordinated with the design con- tract. Seven of ten case study projects modified their design contracts to require the engineer to coordinate the design work packages to fit the construction bid packages. Memphis found some resistance from its design consultant who did not want to change its traditional design sequence to accommo- date the early steel procurement. Nevertheless, working with the CMR they were able to accelerate the necessary design effort and achieve the owner’s desire to lock in the price of steel as early as possible. CONSTRUCTION MANAGER-AT-RISK’S GENERAL CONDITIONS AND OVERHEAD COSTS The cost categories in Table 30 provide examples of typical CMR general conditions and overhead costs. The primary issue is determining what types of costs an agency is allowed to reimburse. Every state will have a somewhat different system for identifying these costs, but all federal aid projects will need to comply with federal contracting regulations. The federal model allows general conditions and overhead costs to be reimbursed if they are “allowable, allocable and reasonable” (Federal Architect–Engineer . . . 2005). The definitions for these three terms are as follows: • Allowable: “a normal cost that a firm would incur in the normal operation of that type of business.” • Allocable: “a cost that would be normally charged for the service to be received and benefits the contract.” 70 • Reasonable: “a cost that does not exceed that which would be incurred by a prudent person in the conduct of compet- itive business” (Federal Architect–Engineer . . . 2005). For example, sending a general superintendent to another state to receive training in accelerated bridge construction would probably qualify because training is a normal cost for construction firms, the contract would benefit from the train- ing, and it is prudent for a competitive business to train their personnel before trying a new technology. The important factor here is to ensure that competing CMR firms are aware of exactly what can and cannot be included in this portion of the GMP. Again, using a system such as the one illustrated in Table 30 is one technique to communicate those facts to the construction community. CONSTRUCTION MANAGER-AT-RISK FEE This portion of the GMP is not to be confused with the fee paid for preconstruction services. This is the profit that the CMR will earn by successfully delivering the project. The system shown in Table 30 shows certain types of costs recovered in the fee. This approach is at odds with the idea that fee is a function of both cost and risk and that a business is entitled to a profit on all its costs. Based on a Washington State study of CMR projects, “the average fee for a general contractor can range from 2% to 15% and the fee amount is contingent on such factors as project risk, contract conditions, competition, and project complexity . . . the fee percentage is typically larger for smaller projects” (Septelka and Goldblatt 2005; italics added). The salient idea in this quote is that the fee is to be determined on a project-specific basis rather than by some arbitrary selection of a percentage. A simple way to avoid having to negotiate this fairly abstract point is to make it a part of the CMR selection process and set the fees upon award. Half of the case study projects established the construction fee in some manner before the contract was awarded. A typical transportation CMR contract clause for converting the fee to a fixed cost at construction contract award was found in the content analysis: At such time the GMP construction contract is executed, the Contractor’s Fee shall be converted from a percentage expres- sion to a stipulated sum amount within the GMP. This fee will not be subject to reduction if the Cost of the Work can be reduced through the efforts of the design/construction team via design refinement or procurement efforts. Abandonment or significant reduction in the scope or magnitude of the project will result in a negotiated reduction of the fee. Conversely, the fee shall not be increased for changes in scope which can be absorbed by the Contingency amount. The fee is only subject to increase should a significant additive scope change occur which would necessitate a change order to the GMP (“CM/GC Fee Structure Sample” 2007). Two agencies (Michigan and Oregon) published a fixed rate in the solicitation documents. Additionally, two of the interviewed contractors had completed projects with non- case study agencies that fix the maximum amount of fee in the same manner and indicated that they had no issues with the practice. One stated that it took one element of uncertainty

71 out of the project; that being how much fee the owner will consider fair and reasonable. The practice reduces the issue to the contractor’s business decision as to whether or not the proffered fixed rate of fee is a reasonable amount considering the risks and complexity of the given project. In other words, to fix the rate forces the construction industry to make a “bid-no bid” decision. As a result, the agency will know that those that do propose are willing to accept that profit level and the issue is no longer open to negotiation. Two other case study agencies require that the construction fee be proposed and evaluated in the selection process. The UDOT does not include a profit factor in its unit price GMP contracts. It requires the CMR to roll this into the unit prices that it furnishes. It also requires competing CMRs to propose unit prices for four to five major pay items as part of the selection process. The remaining five case study agencies negotiate this factor after award. The contractors interviewed were asked how they set this factor and with only one exception they stated that it was purely a business decision based on what they believed the agency would consider reasonable. This con- firms the sentiment expressed by the contractor that believed that the agency fixing the construction fee merely removes one element of uncertainty. Every contractor interviewed remarked that their major motivation was to leave the nego- tiations on good terms with the agency and thereby enhance their chances of winning the agency’s next CMR project. CONTINGENCY DEVELOPMENT From the owner’s perspective, contingency estimating is probably the least understood piece of the GMP. Most agen- cies are used to using a standard percentage of mark-up that is added to the engineer’s estimate to reach a project budget. For example, the U.S. Army Corps of Engineers mandates a 5% contingency (Engineering Instructions Construction Cost Estimates 1997) and the Riverside County, California DOT uses 10% (Estimating Guidelines for Roadway Construction Projects 1999). Hence, understanding exactly what a contin- gency represents is vital to being able to accurately develop one using a logical process. The literature has many definitions for contingencies from a variety of sources. However, there are two that fit the CMR GMP context very well. The first comes from the DOE and binds the contingency issue to the project’s characteristics: The amount budgeted to cover costs that may result from incomplete design, unforeseen and unpredictable conditions, or uncertainties within the defined project scope. The amount of the contingency will depend on the status of design, procurement, and construction; and the complexity and uncertainties of the component parts of the project (Cost Estimating Guide 1997; italics added). This definition narrows the contingency to only those costs that may occur as a result of uncertainties, not those that will occur. Additionally, the amount of contingency is not fixed in this definition, which states that the amount will depend on the project’s current status of completion. Thus, a project where no design has been completed would have a larger contingency than one where the construction is ready to commence. A project where 80% of the subcontract work packages have been awarded and the prices of 100% of the self- performed material have been locked in will have an even smaller contingency. This principle is illustrated in Figure 19. Thus, the DOE definition can be construed to mean that a contingency is the probable cost of the unknowns at the time the GMP is established. This is an important distinction and aids in determining how an agency will want to develop its contingency estimating policy. FIGURE 19 Contingency change as project advances to completion (Ladino et al. 2008). Need Design Design 0 5% 10% 15% Project Stage Identified Concept Schematic Design Development Start Construction Design Complete (Purchasing Complete) Milestones Construction Complete Budget / Baseline 20% A pp ro pr ia te C on tin ge n cy L ev el % (+ /-B ud ge t) Escalation Construction Contingency Design Contingency

As a CMR contract has three prime players, the owner, the designer, and the builder, it has become customary to split the project’s total contingency into logical proportions that relate to the specifics of its status from concept to ribbon cut- ting. For example, another report adds some CMR specifics to the contingency definition when the authors state: Design contingency accounts for estimating inaccuracy due to both quantitative error (take-offs) and qualitative error (design intent). Construction contingency accounts for inaccuracy due to both unforeseen site conditions and contractor risk. Owner contin- gency accounts for . . . things that are overlooked, scope creep, regulatory change, and so on. Escalation is different. Contingencies are for what may happen. Escalation is for what shall happen. Escalation accounts for the persistent inflation of construction costs. The value is reduced to zero when all [subcontractor] bids are in (Spata and Kutilek 2006). These authors differentiate between contingency and escalation using the DOE discriminator that a contingency covers what might happen, but as construction cost inflation is nearly certain during the design phase money to cover the change in prices resulting from inflation is better termed esca- lation. In seven of ten case study projects, the agency chose to break up the overall contingency and assign a contingency pool to itself and a separate one to the CMR. This demands that the agency define what uncertainties each contingency can be used to cover. To do this well requires an open books form of accounting for not only contingency expenditure but for the entire GMP. “Open books” is term that was used extensively in the CMR literature (Gambatese et al. 2002; Doren et al. 2005; Van Winkle 2007; Ladino et al. 2008). It indicates a level of collaboration that includes a sharing of the costs and “[s]haring project information openly, defining risk and profit appropriately, and creating a high level of trust among all the parties” (Doren et al. 2005). It also means that the project execution method is transparent to all parties and it discour- ages “hidden agendas” (Ladino et al. 2008). This leads back to the joint development of the preconstruction cost model. If all parties understand the intricacies of the cost model, they will more closely understand the impacts as estimated costs are replaced by actual costs. Van Winkle (2007) describes it as follows: Open Book accounting is a two-edge sword. Pricing knowledge benefits the owner to confirm cost reasonableness and serves as a basis for change order pricing. However, every variance may be viewed as a change to GMP . . . The solution is to define who owns each risk (Van Winkle 2007). Another author describes the process as follows: Open books accounting eliminates hidden agendas and provides the following • Greater savings opportunities, • Better planning, • Designing to contractor’s strengths, 72 • Improved quality and value, • Shorter project schedules, and • Creation of a “win-win-win” environment (Ladino et al. 2008). Contingency Identification and Management Figure 18 shows three possible contingency options based on the findings of the case study review and the literature review. Those types are: • CMR’s contingency (also termed contractor’s contin- gency or construction contingency) • Owner’s contingency • Management reserve. Three of ten case study projects used a single project contingency. Five had separate contingencies for the CMR and the owner and two had all three. The two projects that added a management reserve to the separate owner’s and CMR’s contingencies were the Fairbanks Airport project and the Miami Intermodal Center. The Alaskan management reserve is a fund controlled by a panel of individuals who are not directly involved in executing the project. Its stated pur- pose was to furnish resources to take advantage of previously unseen opportunities to improve the overall operations of the airport as well as to resource force majeure events resulting from Alaska’s challenging climate and the need to import much of the material from the lower 48 (Storm 2007). The Florida management reserve was created to fund specific owner design changes to the GMP scope (Minchin et al. 2007). In all case studies, the contingency accounting system was transparent and there was some system in place to authorize the use of contingency funds for their intended purpose as well as cross balance between contingency pools if necessary. CMR’s Contingency The CMR’s contingency can be the only contingency in the GMP if that is the agency’s desire, in which case its purpose is to cover all eventualities. However, if the agency decides to separate the CMR’s contingency from an owner’s contingency, it effectively turns the CMR’s contingency into a construction contingency focused on the uncertainties of the market. This may include a separate escalation component that is reduced, as shown in Figure 19, as material pricing and subcontractor work packages are finalized (Spata and Kutilek 2006; Trauner 2007; Ladino et al. 2008). Typical uncertainties assigned to the CMR’s contingency are: • Labor availability (Ladino et al. 2008). • Material pricing (Trauner 2007; Ladino et al. 2008). • Schedule delay costs not attributed to the owner (Touran 2006; Trauner 2007). • Subcontractor coordination/conflict issues (“Request for Proposal Model . . .” 2007). • Other items as may be appropriate.

73 Owner’s Contingency In CMR project delivery the owner holds the design contract and, as a result, the design contingency discussed previously is normally assigned to the owner’s pool. Some agencies have chosen to break a design contingency out of this pool of funds (Spata and Kutilek 2006). In this case, the agency might give control of that contingency to the design consultant and use it as a sort of GMP on the design contract to discourage scope creep. None of the case study projects with separate owner’s contingencies used this technique. Typical uncertainties assigned to the owner’s contingency are: • Design errors and omissions (Cost Estimating Guide 1997). • Scope creep (Cost Estimating Guide 1997; Spata and Kutilek 2006). • Owner-directed scope enhancements (Minchin et al. 2007; Ladino et al. 2008). • Force majeure (“Request for Proposal Model . . .” 2007; Storm 2007). • Regulatory change (Spata and Kutilek 2006). • Other items as may be appropriate. Management Reserve The management reserve can be an interesting feature if used. The fundamental concept is to identify a source of funding to cover the cost of changes, improvements, and operational requirements that affect the project but do not spring from the execution of the project’s intended scope of work. In federal procurement jargon, this might be called a “cardinal change contingency. A cardinal change occurs when the proprietor effects alterations in the works so drastic they effectively require the contractor to perform duties materially different from those originally bargained for” (Whitten 2004). Thus, including this form of contingency in a CMR contract gives the owner much wider latitude to take advantage of unforeseen opportunities as they arise. In another author’s words, manage- ment (owner) reserve is budgeted for discretionary purposes (Van Winkle 2007). An example of this occurred in the Weber County Commuter Rail project in Utah. The CMR was able to create a substantial savings during preconstruction through value engineering. The savings flowed to the owner’s contin- gency where it was used to add park-n-ride structures that were not in the original scope of work; that is, execute a cardinal change to the benefit of the project. Although UTA did not have a separate management reserve fund, its owner’s contingency was structured in such a fashion that it could and did function as one. Another example from the building industry shows just how creatively a management reserve can function. A CMR project was nearing design com- pletion and the CMR informed the owner that it could com- mit to a GMP that was substantially lower than originally estimated. Therefore, there appeared to be no need to use the project’s management reserve. The owner decided to release the funding to the CMR with instructions to work with the architect and use the available capital to qualify the building for certification at the highest level of sustain- ability (Winters 2004). Another situation in which a management reserve is appro- priate is when the agency needs to establish a GMP at an early stage of design. In transportation this sometimes occurs for bond-funded projects. In this case, “a design contingency is often held outside the GMP [making it a management reserve] to be drawn against as the exact scope of the work becomes better defined” (“Construction Delivery Approaches” 2007; italics added). Doing this allows the CMR to reduce its contingencies because funds are available from the man- agement to increase the GMP amount for significant scope changes. The Miami Intermodal Center case study project used its management reserve in this manner. Minchin et al. (2007) described the function of this type of contingency as follows: There is a contingency within the GMP to cover unexpected but justifiable costs, and a contingency above the GMP allows for owner changes. As long as the subcontracts are within the GMP, they are reimbursed to the CM, so the CM represents the owner in negotiating inevitable changes with subcontractors. The key element in the CMAR system on this project is the contingency fund (10% on this project). Without that, an adversarial atmo- sphere would appear on the project. Instead of the prime contractor or the CM looking for changes as on a DBB project, the sub- contractors are doing so, but a strong CM insulates the owner from this problem (Minchin et al. 2007; italics added). A third situation where a management reserve would be helpful is a project where construction has started without a full suite of permits. If one of the permits creates a scope increase or major delay, the management reserve could be established to specifically assign that risk to the owner. Its magnitude might be able to be estimated based on an analysis of possible outcomes from the permit review (Touran 2006). Again, by quantifying the uncertainty and planning to resource it, the CMR will not need to inflate its contingency to cover that risk. There may also be some time savings to the owner in terms of the funding and contracting process by having a “preapproved” source of funding for a possible but not prob- able scope change. GUARANTEED MAXIMUM PRICE ADMINISTRATIVE ITEMS The final establishment of a GMP is a contractual transaction and as such will involve more than just the mere mutual agreement on a number. To effectively monitor and adminis- ter the GMP, back-up documentation that defines exactly what is represented by each number accompanies the form that displays the GMP itself. One author furnished a nearly comprehensive list of what is recommended to be included in the GMP documentation: • “a list identifying drawings and specifications, addenda, and other documents used in preparation of the GMP;

• a list of allowances and definition of materials and labor included in such allowances; • the construction manager’s assumptions and qualifications; • a breakdown of the GMP into costs of the work (trade cate- gories, holds, general conditions, insurance) and the construc- tion manager’s fee; • alternates included in the GMP and those available (with cut- off dates) for subsequent acceptance by the owner; • the date of substantial completion and other milestones (if applicable); and • proposed incentives (if any) for the construction manager’s performance” (Martinez et al. 2007). The point made by Martinez with this list of inclusions is that the GMP represents a cost estimate plus contingencies that were developed at a specific point in time and a specific level of design development. Understanding all the logic behind the numbers becomes extremely valuable if circum- stances arise where the numbers need to be adjusted. Having reviewed and approved the items shown previously, as well as others that would be appropriate for a given project, cre- ates a common ground from which to negotiate a fair and equitable adjustment of the GMP. “There has to be an ele- ment of trust and some give and take between the owner, designer and general contractor/contract manager [CMR] in order for this process to work” (Storm 2007). Preserving that element of trust is most difficult when changes are being negotiated. Thus, making sure that the administration is both complete and thorough is essential to success. A major aspect of GMP administration is the process of reaching a “final GMP.” Progressive Guaranteed Maximum Prices Design detail drives the amount of contingency that is contained in the GMP. Some agencies, such as UDOT, use a progressive GMP to keep project contingencies as low as possible. In essence, a progressive GMP is nothing more than breaking the project down into phases or work packages and asking the CMR to generate individual GMPs for each phase/ package as its design is completed. The final GMP becomes the sum of the individual GMPs plus any remaining project- level contingencies. This allows the design to progress without undue pressure and allows the CMR to furnish GMPs on phase design packages as soon as they are ready. “Practitioners have recommended that the GMP is more accurate when certain design elements are completed to 100 percent, rather than having all design elements partially completed, allowing the CM to lock in subcontractors and reduce the estimation involved in developing the GMP” (Trauner 2007). The three most experienced case study agencies (UDOT, city of Glendale, and Pinal County) used progressive GMPs. The interviews with these agencies and their contractors con- firmed that this was a key feature in controlling costs on CMR projects. On the Utah project, the phasing associated with the progressive GMP is credited with saving a full construction 74 season (Alder 2007). UDOT asked its CMR to furnish the following three progressive GMPs: 1. “Early Action Package”—consisted of maintenance of traffic, steel, and Mechanically Stabilized Earth wall panels. 2. Eastbound I-80 with Accelerated Bridge Construction costs. 3. Westbound I-80 plus remaining ancillary items. The city of Glendale project had substantial amounts of utility coordination and a large number of affected commer- cial businesses. Therefore, it used the progressive GMP to get the CMR started on the utilities, while the consultant sorted out the details of the design with the various business owners and public officials. Again, the CMR was not asked to furnish a GMP for undesigned features of work and this allowed it to keep the project on budget and to finish two months early. Pinal County used the progressive GMP to stay in touch with material price changes. The CMR furnished a forecast of key material prices for each design package and the agency used this to build the CMR’s contingency as well as to prioritize the design effort. Given the success of these three cases, the use of a progressive GMP appears to be very attractive. This leads to the conclusion that agencies planning to use CMR seriously consider incorporating a progressive GMP into their procurement package. Shared Savings Clauses The literature about GMP contracts discusses the advantages and disadvantages of including a shared savings clause, and a critical reading of it leaves the reader unconvinced as to its real value in influencing contractor behavior to minimize the overall project cost. In one paper written by a consulting engineer, the author advocates the use of shared savings as incentives (Cunningham 2005). In another, coauthored by a construction contractor and a public owner, they discourage shared savings clauses as unnecessary and ineffective (Bearup et al. 2007). Looking deeper, a study of $6.6 billion worth of CMR projects built by the state of Washington found that half the projects included shared savings incentives. However, the average incentive earned was less than 1% of the CMR contract value (Septelka and Goldblatt 2005). A second study of all CMR projects completed by the city of Seattle concluded: The City has used incentives in most of its GC/CM [CMR] proj- ects to encourage achievement of “social goals” such as small business participation, as well as cost savings. Some owners feel that offering a portion of the “savings” to the contractor will provide incentive to the contractor to increase diligence and efficiency to reduce costs. However, it is unclear that savings clauses actually result in a lower price at project completion (Denzel et al. 2004). These two studies tend to substantiate Bresnen and Mar- shall’s (2000) more philosophical conclusion that “clients and contractors are complex social entities, and their behav- iors are not necessarily modified simply by the existence of incentive schemes at an organizational level such as pain/gain sharing mechanisms.” To use the figures from the Septelka

75 and Goldblatt study (2005), if $3.3 billion worth of CMR projects only generated a fraction of 1% savings, then Bresnen and Marshall are correct in their finding that it takes some- thing more substantial than a contract to modify corporate behavior. Thomsen (2006) summarizes the relative value of shared savings clauses succinctly when he says: It’s a process and a set of agreements based on incentives vs. consequences, morality vs. legalities . . . There must be a percep- tion that repeat work will follow good performance. By far the most important incentive that an owner has is the promise of repeat work (Thomsen 2006; italics added.) Reviewing the case study project output, six of ten projects included a shared savings clause. In the highway group, three of seven used the shared savings incentive. Once again the three most experienced agencies, UDOT, Glendale, and Pinal County, did not share savings with their contractors. More importantly, the contractors on those projects expressed the view that a shared savings clause took more time to admin- ister and verify than it was worth. One contractor stated that the owner was paying them a preconstruction services fee to find ways to save. Therefore, a shared savings clause created an unintended conflict of interest. The content analysis only found one solicitation with shared savings indicated. Finally, the author of this report was a member of the legislative task force that drafted of the Oklahoma CMR law and was initially surprised when the contractor’s association representatives come out forcefully against including shared savings incentives in the act. This also came as a surprise to the state agency and design consultant representatives. The contractors’ motivation was to eliminate one obstacle to developing trust in the rela- tionships. The contractors believed that if there was a shared savings clause, the owner would suspect that the CMR would inflate the GMP as a means of making additional profit by means of the incentive. Again, circling back to the two studies in Washington State, little, if any, benefit was realized by the public agencies that used this approach. Thus, if the Oklahoma AGC and the previously mentioned paper by the contractor could be considered representative of the larger construc- tion industry, a conclusion can be reached that shared savings clauses are not as valuable as some would believe. These two lines of information are squarely intersected by the case study project findings that nearly half the cases, including the three agencies with the most CMR experience, omitting shared sav- ings from their CMR contracting program. The lack of shared savings findings in the content analysis validates the notion that shared savings adds little value to a CMR project. Failure to Reach Agreement on Guaranteed Maximum Price Negotiating a GMP for a large, complex, fast-moving high- way project is not a trivial task. It requires both sides to come to the table prepared to demonstrate the validity of their numbers and their assumptions. Hence, there comes a time on the odd project when agreement is not possible. The final administrative aspect of a GMP contract is the question of what happens if the owner and the CMR are unable to negotiate a final GMP. The case study interviews asked each agency and contractor for their individual policies for this occurrence. The Glendale City Engineer stated that he did not know what they would do as it has never been an issue, an encouraging comment for those agencies who are thinking about trying CMR for the first time. Of all the agencies interviewed, only one had ever had that experience. Its solution was to com- plete the design and advertise the project as a DBB project. The reason for the failure to reach an agreement was the agency’s belief that certain material prices were overstated. When the bids were opened on the project, the lowest bid was higher than the CMR’s rejected GMP. This brings the issue of the owner’s lack of real-time pricing information to which the CMR has access to center stage. Most DOTs depend on their bid tabulations as their primary source of estimating data. This information is rea- sonably current, but is at least 30 days behind the marketplace when published. In the summer of 2008, the price of liquid asphalt doubled inside of two months (“WSDOT Market Analysis” 2008). Therefore, negotiating a CMR paving project during that period would create difficulties if the owner had not kept current on the volatility in asphalt prices. UDOT retains an independent construction estimator to literally develop a parallel GMP using the same documents as the CMR (Alder 2007). This particular consultant is a retired construc- tion contractor who has access to the same sources of pricing information as the Utah contracting community. The ODOT retained a contractor who was not involved in its CMR project to do exactly the same thing for them. This appears to be a redundant service; however, DOTs can recognize the limita- tions of their staff to approach a cost estimate the way a con- struction contractor will as well as the limitations of their bid tab estimating database. Based on the case study outcomes, there appear to be two methods for completing a project where the CMR and owner cannot agree on a GMP. The first is to take advantage of the concept that the owner still controls the design contract and advertise the project for competitive bids. This is the method advocated by UDOT (Alder 2007) and the method used by seven of ten case study agencies. In this manner, the agency would pay off the preconstruction services contract and not award the planned construction contract. The other method that is used by Pinal County is to open negotiations with the contractor who was ranked second by the selection panel. This would appear to be the more expeditious of the two methods in that taking the statutorily prescribed time to advertise and award the construction contract would create a several month delay. The second method would also not require the designer to complete fully biddable construction documents. One of the verified savings in CMR is in design cost and design time (Uhlik and Eller 1999; Thomsen 2006; Alder 2007). Most of these savings are because the CMR involved in the design

process can signal the engineer to stop designing when it has sufficient technical detail to get decent bids from its sub- contractors and build the job properly. During the UDOT interview, the agency indicated that it was saving roughly 40% on its design costs compared with DBB. This becomes a two-edged sword if GMP agreement is impossible. The designer may ask for additional compensation to take the design from where it was told to stop to where it needs to be to permit competitive bidding. SUMMARY Establishing a GMP in a CMR contract is perhaps the most important aspect, because it quantifies the project’s costs for the owner and the potential profit for the contractor. The industry has many different manners in which this critical aspect of CMR contracts can be established. Conclusions A number of conclusions are drawn from the analysis. • Allowing the CMR to self-perform those work packages that it is well-qualified to perform does not increase costs. • Allowing the CMR to prequalify and select subcontrac- tors without constraint preserves the CMR’s ability to 76 use its long-standing commercial relationships for the benefit of the project. • Progressive GMPs add value to CMR contracts by allow- ing the CMR to establish incremental GMPs as design packages are ready. • Shared savings clauses do not appear to create a signif- icant incentive and could possibly create a loss of trust if misunderstood. Effective Practices Several effective practices were reported: • Splitting the contingency between the owner and the CMR appears to make accounting for contingency allo- cation less onerous. • An open books approach to contingency calculation and allocation enhances the spirit of trust between the owner and the CMR. • Those projects where the agency wants to mitigate material cost escalation risk can include an early work package to allow the CMR to procure volatile construc- tion materials as early as design progress allows. This also argues for a progressive GMP that includes as a minimum the early materials package and the remainder of the job.