Final Report for NCHRP Report 574: Guidance for Cost Estimation and Management for Highway Projects During Planning, Programming, and Preconstruction (2006)

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.

10 CHAPTER 2 STATE OF PRACTICE The objective of Task 1 is to confirm the research team’s understanding of the problem, including key factors influencing project cost escalation, and to characterize the current state of Departments of Transportation (DOT) practice as related to cost estimation practice and cost estimation management. Project cost estimation and the management of project cost estimates are critical issues facing state DOTs. The three-element framework of strategies, methods, tools, in combination with project phases, and the impact of project complexity were used to structure the information collected. The team assembled state-of-the-practice estimation information by project development phase so that the final estimation guidelines can present tools to develop, manage (track), and document realistic cost estimates during each phase of a project. Figure 2.1 summarizes the inputs and outputs of this task. TASK 1 (CHAPTER 2) STATE OF PRACTICE LITERATURE FACTORS CAUSING COST ESCALATION GENERAL OVERVIEW OF CURRENT PRACTICEINTERVIEWS Figure 2.1. State-of-the-Practice Inputs and Outputs LITERATURE REVIEW AND ANALYSIS The literature review involved researching, gathering, reading, note taking, and processing information and literature relevant to cost estimation and the management of project estimates. The seeking process consisted of two methodologies. The first method consists of identifying sources through a number of search techniques and searches using keywords that identify areas of significance to the research. The following search engines were used: The general Internet, including Yahoo, Google, and Hotbot; Transportation Research Information Services (TRIS); Academic resources, such as LexisNexis and Engineering Village 2; Research Institutions, for example the Construction Industry Institute (CII) and Transportation Research Board (TRB);

11 Societies for journal and conference publications, consisting of the American Society of Civil Engineers (ASCE) and the Association for the Advancement of Cost Engineering International (AACEI); and Government publications, both federal and state, United States General Accounting Office, and state DOT research departments. Some representative search terms included: Estimating; Cost management; Cost overruns; Construction cost underestimation; and Project controls. Searches were made with each of these words individually and together with more specific terms such as highway, transportation, infrastructure, design, planning, and construction. The second method used to seek possible literature of interest was through the references noted in the literature that was examined. Efforts were made to gather the literature of interest through the search engines themselves, as well as through visits to various libraries, through interlibrary loan services, and document request functions. Each document gathered was read and notes taken on any information relevant to the research. Some sources of information found are summaries of a number of research efforts. In cases such as these, endeavors were made to obtain the original source of work rather than the summary. The notes were then filtered and information helpful in achieving the goals set forth in this research extruded. Over 100 documents have been reviewed and summarized by members of the research team. The documents consist of journal articles, reports, conference proceedings, and other documents (presentations, summaries) in the proportions shown in Figure 2.2. These documents and their abstracts have been uploaded to the controlled project website at: construction.colorado.edu/nchrp8-49. The articles on the website have been categorized by their application to project development phase and their relevance to estimation strategies, methods, and tools to allow for quick access by the research team. Figure 2.2. Categorization of Articles from Literature Review

12 The literature was analyzed with attention to cost estimation practice and cost estimation management. Additionally, it was analyzed in relation to the three-element framework and the DOT project development phases presented earlier. This organizational format supports the structure of the white papers for Tasks 1, 2, and 3 and provides supporting evidence for those strategies, methods, and tools identified through the Task 1 interviews and then derived during Task 2 from the critical review of the interview information. The majority of articles contained in the research team’s literature review to create the database of publications are from non-transportation industries, such as journals from the Association for the Advancement of Cost Engineering International (AACE International), the Project Management Institute (PMI), and the American Society of Civil Engineers (ASCE). This literature will be helpful in developing estimation strategies, methods, and tools that are not currently being employed by DOTs. Most transportation sector estimation literature focuses on cost estimation during the pre-construction phases with very little information available on procedures for estimating cost during the early stages of project development (Schexnayder et al 2003). Much of this literature does address problems or issues with cost estimation such as cost escalation. Further, the research team found that literature on cost estimation management practices in transportation is virtually nonexistent. The transportation literature more often addresses problems that are frequently associated with larger and more complex projects. The Federal Highway Administration (FHWA) is in the process of creating a set of guidelines for estimating major projects (June 2004). Major projects have cost estimates larger than one billion dollars. The draft guideline has established a set of key principles that a transportation agency should have in order to produce a reasonable estimate. The main principles are integrity, contents of a cost estimate, year-of-expenditure dollars, basis of a cost estimate, risk and uncertainty, project delivery phase transitions, team of experts, validation of estimates, revalidation of estimates, and release of estimates and estimation information. The principles state that the cost estimate should accurately reflect the all the projects cost components with proper adjustment for inflation, risk, and uncertainty. The estimators should act honestly, generate estimates using the best information available to them, and apply sound engineering judgment. Furthermore, the different project estimates should be well documented, approved, and undergo periodic reviews through out project development. The FHWA guideline also describes the elements that each project estimate should contain, and it includes a checklist to ensure the elements have been considered. Some of the elements identified are preliminary engineering, right-of-way, construction costs, and contingency. In addition to the checklist, FHWA identifies areas of cost estimation that should be considered during the earlier stages of cost estimation when the project is not well defined. For example, the guidelines recommend having documentation from the beginning of the project to the end, and it explains that estimating risk should be done during the initial estimates. The guidelines also state that transportation agencies must integrate quality control and assurance into the estimation procedures. STATE DEPARTMENT OF TRANSPORTATION PRACTICE After completing the literature search, it was apparent that the highway industry has little information published concerning cost estimation and the management of project estimates. Because of the scarcity of publications on cost estimation procedures and management in the

13 highway industry, the research team conducted a series of interviews with state DOTs to determine current DOT cost estimation procedures and estimation management practices. The surveys enabled the research team to acquire insightful data directly from the DOTs. After the DOT interviews, the research team interviewed other organizations to confirm the DOT information. Interview Protocol The interview protocol was designed to permit the research team to obtain DOT information concerning: Who is responsible for preparing and approving the estimates at each stage; How estimates are prepared and managed, where estimates are prepared; and What purpose the estimate serves. The research team used Schexnayder’s (2003) Project Cost Estimating Synthesis, and Dr. Schexnayder’s discussions with Arizona and Idaho’s Department of Transportations as a basis for developing the questions for the interview protocol. Then, the research team applied similar categories identified in the synthesis on Statewide Highway Letting Program Management (Anderson and Blaschke 2004). The categories, which are planning, programming, advanced planning/preliminary design, and final design, reflect typical phases in the project development process. These phases were also outlined and described in the interview package so the team could align definitions of the phases with those of the individual DOTs. Under each phase of project development, the questions were further categorized by topic. The subtopics included: estimate preparation, estimate review, estimate communication, and estimate/cost management. The questions in each section of the interview protocol are similar to allow the interview to be conducted on an individual basis or with a group of DOT personnel representing the different sections within the DOT responsible for each of the project phases. The interview protocol was pretested with two DOTs (Washington State and Florida) to ensure that the questions adequately covered the topic areas. The only change to the protocol was to split long-range planning from programming as the estimates in these two project phases have different purposes and frequently different individuals are involved in their preparation. Otherwise, the questions were deemed adequate and comprehensive. The interview instrument is provided in Appendix A. Interview Process The research team conducted interviews with state DOTs and other organizations. The research team relied on three sources to identify appropriate interviewees: 1) participants in the TRB Cost Estimating Workshop5; 2) members of The Technical Committee on Cost Estimating which is part of AASHTO Subcommittee on Design6; and 3) contacts established by Dr. Schexnayder when he prepared the synthesis on project cost estimation. A letter that briefly outlined the purpose of the project, provided some background information about the project, and requested a 5 TRB AFH35T, Special Task Force, Accelerating Innovation in the Highway Industry, Cost Estimating Workshop, Washington, DC, February 11, 2004. 6 The Technical Committee on Cost Estimating was created in Spring 2002 by the Standing Committee on Highways to provide a focal point for cost estimating issues within AASHTO.

14 list of individuals who would have appropriate knowledge for the interview was sent to the contacts identified from these three sources (see Appendix B). Initially, the research team assumed different individuals were involved at each project stage. Therefore, the letter included a form specifically requesting the names of individuals with knowledge of conceptual estimation, preliminary design estimation, and the engineer’s estimate. When the team collected the responses, the responses revealed this assumption was only partially true. The team did receive three different names from some of the DOTs. However, a few DOTs had two people listed for one project phase, and other DOTs listed the same person for several phases. When the interviews were scheduled, the research team contacted each participant from the DOT and gave them the option to perform the interview independently or as a group. Many DOTs requested that all representatives of the different project phases be present during the interview, while other DOTs preferred to complete the interview individually. The research team was able to send letters to specific individuals in all fifty states. The team received responses from 36 states. The research team selected specific DOTs for interviews based on prior knowledge of their practices from Schexnayder (2003) and judgment based on potential diversity in practice, size, and geographic location. Once the research team received responses to the contact letter, interviews were coordinated with the DOTs. The interview instrument was sent to the DOTs prior to the interview. It was also requested that each DOT send any supplemental information such as estimation procedures or manuals to the researchers prior to the interview. During the interview, the team began by providing background information about the project, seeking to understand how the agency defines project phases and when in the project development process estimates are prepared. After that, the interviewers proceeded with the questions from the interview instrument. Most of the interview discussions were more topical based than specifically following a question-by-question approach. The topical areas were estimate preparation, estimate reviews, estimate communication, and cost estimation management. At the end of the interview, the interviewers asked the agency to look over the questionnaire to make certain all questions were discussed. Once the interview was complete, the answers recorded from the interview were aligned with the corresponding questions. In addition, a summary page was written that presented an overview of the DOT’s estimation process, the strengths, and the weaknesses identified by the state agency (see Appendix C for an example). When the interviews were conducted, the researchers requested any documents that the DOTs might have related to their cost estimation practice and cost estimation management procedures and policies. Procedure manuals were obtained that described the steps to prepare an estimate using the DOT’s in-house estimation software. The DOTs also provided presentations and documents that describe specific estimation methodology developed by the DOT. Some of the acquired documentation shows typical pavement sections developed by the DOT along with their associated cost per mile factors. Several DOTs supplied spreadsheets that the DOTs use to prepare and document an estimate. The spreadsheets also have inflation rates that are applied to the estimates. One DOT furnished their estimation policies, which included documentation requirements for estimates prepared during each project development phase. This DOT’s procedure also covered cost estimation management practices as well by project phase. Their policies also cover the type of estimation methodology permitted and the approval requirements at each project development phase. Although not every interviewed DOT provided

15 documentation, most DOTs have procedural manuals. These manuals mostly cover estimation during final design. Data Characteristics The research team completed 18 formal DOT interviews (Figure 2.3). Two of these interviews were used to test the interview protocol. Minor changes were made to the protocol and then the remaining state-of-practice interviews were conducted as previously stated. In addition to the interview methodology, research members participated in a peer exchange at the Joint Summer Meeting of the Planning, Economics, Environmental, Finance, Freight, and Management Committees held in Park City, Utah in July 2004 as described in the following section. Representatives from 14 DOTs participated in the Park City meeting. Information was collected from a total of 23 DOTs during Task 1 through interactions between research team members and DOTs. As seen in Figure 2.3 contributions were made by departments from across the nation and representing a variety of program sizes and diverse attitudes, policies, and issues. Figure 2.3. States Represented in Task 1 The state-of-practice interviews began at the end of May and continued through August 2004. The team accomplished the interviews in several different ways, but primarily either over the telephone or at the agency’s headquarters. During the telephone and onsite interviews, either every project development phase was discussed or only a single phase. The type of interview

16 along with the date it was conducted is listed in Table 2.1. Agencies that responded to the initial contact letter and were not interviewed have been notified that they might be contacted at a later date. The research team also requested estimation information such as manuals or guidelines from those agencies. Table 2.1. Type and Date of Interview State Highway Agency Date(s) Interview Type of Interview Peer Exchange 1 Arizona DOT January 2004 Interview Development 2 California DOT July 23, 2004 Onsite – All Phases √ 3 Connecticut DOT August 2, 2004 Onsite – All Phases 4 Florida DOT May 28, 2004 Onsite – All Phases √ 5 Georgia DOT July 6, 2004 Telephone – All Phases 6 Idaho DOT January 2004 Interview Development 7 Illinois DOT July 6, 2004 Telephone – Single Phases 8 Kentucky Transportation Cabinet June 14-17, 2004 Telephone – Single Phases 9 Michigan DOT July 27-28, 2004 √ 10 Minnesota DOT June 7, 2004 Telephone – All Phases √ 11 Missouri DOT June 7, 2004 Telephone – All Phases √ 12 Montana DOT July 27, 2004 √ 13 Nebraska Department of Roads June 16 & 18, 2004 Telephone – Single Phases 14 New York DOT July 15, 2004 Telephone – All Phases 15 North Carolina DOT July 12 & 29, 2004 Telephone – All Phases √ 16 Ohio DOT July 27, 2004 √ 17 Pennsylvania DOT July 8,2004 Telephone – All Phases √ 18 Texas DOT July 2, 2004 Onsite – Single Phases 19 Utah DOT June 1 & 14, 2004 Telephone – Single Phases √ 20 Vermont DOT July 27-28, 2004 √ 21 Virginia DOT July 12, 2004 Telephone – All Phases √ 22 Washington DOT May 21, 2004 Onsite – All Phases √ 23 Wisconsin DOT July 27, 2004 √

17 Utah Park City Peer Exchange In addition to the structured interviews a source of estimation information was the Park City Peer Exchange in Utah. The research team participated in a TRB sponsored “Peer Exchange” with the TRB Statewide Multimodal Planning Management Committee. This “Peer Exchange” occurred on July 27, 2004. The sponsoring committee invited the attendees. Participation involved 14 Department of Transportation representatives and eight representatives from other groups, including the FHWA and transportation consultants. A facilitator from NCHRP and facilitator from a transportation consulting firm were also present. The invited guests were given five questions to address prior to the peer exchange. The questions focused on the following areas: Major issues regarding planning or programming cost estimates Policies, procedures, techniques, and/or standards used in preparing planning or programming conceptual estimates Methods to insure the project scope is adequately covered under conceptual estimates Historical data used in preparing conceptual estimates Contingency and risk applications in conceptual estimation During the “Peer Exchange”, the research team made a brief presentation on the status of the research project. Each participating organization was then asked to discuss their major issues and briefly address the five estimation-related questions relative to project planning and/or programming. The discussion was recorded and most of the agencies provided written documentation. The research team summarized the discussion into 15 major issue areas and identified 11 approaches for managing the cost estimation process during the planning phase. Key members of the peer exchange group reviewed the research team’s summary of the issues and approaches and provided further comment and confirmation. A major comment from the Peer Exchange group related to the basic approach DOTs take when developing long-range plans. There appears to be two different types of long-range plans. One type is to prepare a “Policy-based Long Range Plan.” The Policy-based Long Range Plan does not identify specific projects and associated cost estimates; it only provides major categories of needs and their priorities. The other type focuses on specific projects to prepare a “Project-based Long Range Plan.” This latter type requires cost estimates for projects. The interrelationship between these two types of long-range plans is not clear. Hence, the research team is performing some additional investigation into policy and project based long-range planning in terms of their impact on cost estimation and cost management practices. The primary target of this investigation is DOT Planning Directors. The results obtain from the “Peer Exchange” was considered in the general description of the state-of-practice as well as strategy development. Other Organizations The research team feels that information pertinent to this project may also be obtained through discussions with organizations other than state DOTs. Therefore, contacts were sought with MPOs as well as with other transportation-engineering firms. Additionally, contact was made with non-transportation owner organizations.

18 Metropolitan Planning Organizations Texas Transportation Institute (TTI), Associate Director, Katie Turnbull, assisted the research team in identifying MPO contacts. The team has contacted several MPOs with the intent of conducting interviews with respect to their cost estimation procedures and cost management practices. At the time of the Interim Report, the team has interviewed the Maricopa Association of Government MPO and the Denver Regional County of Governments (DRCOG). Consultants The research team has also obtained information on cost estimation practice and cost estimation management from a transportation consultant’s perspective. The interview instrument provided the basis for collecting this input. Two major transportation industry design consultants participated in interviews: 1) Carter-Burgess and 2) Michael Baker. Representatives from both of these organizations answered the interview questions from their corporate perspective and involvement in the development and design of transportation projects. The interview information was documented and the results of the analysis are included in the state-of-practice. Non-Transportation Owners The University of Colorado contacted non-transportation owners who are members of the Rocky Mountain chapter of the Association for the Advancement of Cost Engineering International (AACEI). One interview was successfully complete with a non-transportation organization. In selecting non-transportation organizations to question during the course of this study, the research team decided that the selected organizations should be larger, owner types with capital project experience and large in-house engineering staff. One successful contact was with Coors Brewing Company, based in Golden, Colorado. Coors Brewing Company is a continuously expanding company that operates businesses in brewing, aluminum rigid container sheet, folding carton and flexible packaging, as well as ceramics. Coors is primarily known as a beer brewing company and as such it requires new facilities and maintenance of facilities involved with grain handling and storage, malting operations and storage, brewing, fermenting, storage of aging beer, and packaging and cold storage warehousing throughout the nation. Additionally, it constructs water collection and treatment facilities, waste treatment, steam generation, refrigeration, electrical systems, office buildings, and distribution warehousing (Berka and Daley 1992). The interview process with Coors was similar to the process used with DOTs. Before the interview, contact documents were obtained that provided information regarding the construction program at Coors. Using these documents, the research team became better acquainted with Coors project development processes. The participants discussed the project development phases of the transportation industry and the process of project development followed by Coors. Upon determining that the project development phases are similar in nature, the discussion turned to the state-of-practice survey instrument. The research team gained significant insight from the Coors interview. Some of the responses were similar to what was gathered from DOTs and but many differed and suggested different approaches to dealing with estimate development and control.

19 Data Analysis The data collected from all of sources enabled the research team to identify the core estimation assumptions that are the root causes behind cost escalation and lack of project estimate consistency and accuracy. The team was also able to identify specific estimation practices and cost estimation management approaches currently used or maybe more importantly not being used during each project phase. The data recorded during the interviews was analyzed by citing common practices approaches used by DOTs in the database. Documents provided by some DOTs were also studied. In many cases, these documents provided additional details of a DOTs current practice. The research team convened in College Station, Texas and spent two days analyzing the data. First, the team reviewed the factors leading to cost escalation. Through the interviews, the team was also able to determine problems that arise out of the agencies’ weaknesses, which in turn led to the development of factors influencing cost escalation. The factors were further proven through correlation with the information found in the literature review. The research team organized the information and characterized the state of practice as revealed by the interview data. The information was separated into project development phases and topical areas: preparation, review, communication, and management. Once the data was organized, the team developed trial approaches that described the state of practice. The conclusions that were drawn from the literature review and analysis of the interview data are described in the following section. CHARACTERISTICS OF THE STATE OF PRACTICE The state of practice is described first by identifying factors that lead to cost escalation. This discussion is followed by a discussion of DOT identified practices relevant to cost estimation procedures and cost management approaches. The description is general in nature and does not describe a particular approach of any DOT. Cost Escalation Factors Construction projects have a long history of cost escalation. (Federal-Aid 2003, Flyvbjerg 2002) The factors that lead to project cost escalation have been identified through a large number of studies and research projects as described in the literature. The factors driving cost escalation of project cost can be divided by project development phases: planning, and execution. As defined here planning involves all project development phases prior to bidding including long-range planning, programming, advanced planning/preliminary design, and final design. Execution entails contract bidding, award, project construction, and closeout. The factors that affect the estimate in each development phase are by nature internal and external. Factors that contribute to cost escalation and are controllable by the DOT are internal, while factors existing outside the direct control of the DOT are classified as external. This arrangement of factors is shown in Table 2.2, these factors are numbered for reference only and do not suggest a level of influence. Table 2.2 has been constructed to provide an over arching summary of the factors that have been identified from many sources and a better understanding of how project estimates are effected. It is important to note that one of the factors points to

20 problems with estimation of labor and material cost, but most of the factors point to “forces” that impact project scope and timing. Table 2.2. Factors Causing Cost Escalation of Projects* Planning Execution Internal 1. Bias 2. Delivery/Procurement Approach 3. Project Schedule Changes 4. Engineering and Construction Complexities 5. Scope Changes 6. Poor Estimating (errors and omissions) 7. Inconsistent Application of Contingencies 1. Inconsistent application of Contingencies 2. Faulty Execution 3. Ambiguous Contract Provisions 4. Contract Document Conflicts External 1. Local Government Concerns and Requirements 2. Effects of Inflation 3. Scope Creep 4. Market Conditions 1. Local Government Concerns and Requirements 2. Unforeseen Events 3. Unforeseen Conditions 4. Market Conditions * Note: these factors are numbered for reference only and do not suggest a level of influence. Planning—Internal While numerous internal factors can lead to underestimation of project costs at the planning stages seven primary internal factors have been well documented: bias, delivery/procurement approach, project schedule changes, engineering and construction complexities, scope changes, poor estimation, and inconsistent application of contingencies. Each of these factors separately or in combination with others can cause significant project costs increases. Bias is the demonstrated systematic tendency to be over-optimistic about key project parameters. It is often viewed as the purposeful underestimation of project costs in order to insure a project remains in the construction program. This underestimation of costs can arise from the DOT estimators’ or consultant’s identification with the agency’s goals for maintaining a construction program. The project process in some states is such that the legislature establishes a project budget by legislative act and that budget is based on preliminary cost estimates. Later if the department’s estimate is higher than the budget, the project may not be let. As a result, engineers and the DOTs feel the pressure to estimate with an optimistic attitude about cost - (Akinci 1998, Condon 2004 Bruzelius1998, Flyvbjerg 2002, Hufschmidt 1970, Pickrell 1990, Pickrell 1992). Delivery/Procurement Approach effects the division of risk between the DOT and the constructors, and when risk is shifted to a party who is unable to control a specific risk project

21 cost will likely increase. The decision regarding which project delivery approach, design-bid- build, design-build, or build-operate-transfer, and procurement methodology, low bid, best value, or qualifications based selection effects the transfer of project risks. In addition to the question of risk allocation, lack of experience with a delivery method or procurement approach can also lead to underestimation of project costs. Many DOTs are looking to reduce project schedules in order to quickly deliver much-needed projects to the traveling public, but accelerated schedules are only achievable at a cost. While the end results of applying different procurement approaches should be beneficial some hard lessons must be learned regarding the proper allocation of risks and what each new method entails, in terms of DOT responsiveness, expectations, and time (Harbuck 2004, New Jersey 1999, Parsons 2002, SAIC 2002, Weiss 2000). Project Schedule Changes, particularly extensions, caused by budget constraints or design challenges can cause unanticipated increases in inflation cost effects even when the rate of inflation has been accurately predicted. It is best to think in terms of the time value of money and recognize that there are two components to the issue: 1) the inflation rate; and 2) the timing of the expenditures. Many DOTs have a fixed annual or bi-annual budget and project schedules must often be adjusted to ensure that project funding is available for all projects as needed. Estimators frequently do not know what expenditure timing adjustments will be made (Board 2003, Booz·Allen 1995, Callahan 1998, Hufschmidt 1970, Mass 1999, Semple 1994, Touran 1994). Engineering and Construction Complexities caused by the project’s location or purpose can make early design work very challenging and lead to internal coordination errors between project components. Internal coordination errors can include conflicts or problems between the various disciplines involved in the planning and design of a project. Constructability problems that need to be addressed may also be encountered as the project develops. If these issues are not addressed, cost increases are likely to occur (Board 2003, The Big Dig 2003, Booz·Allen 1995, Callahan 1998, Hufschmidt 1970, Mass 1999, Touran 1994, Federal-Aid 2003, Transportation Infrastructure 1997, Transportation Infrastructure 2002). Scope Changes, which should be controllable by the DOT, can lead to underestimation of project cost escalation. Such changes may include modifications in project construction limits, alterations in design and/or dimensions of key project items such as roadways, bridges, or tunnels, adjustments in type, size, or location of intersections, as well as other increases in project elements (Board 2003, Booz·Allen 1995, Callahan 1998, Chang 2002, Harbuck 2004, Hufschmidt 1970, Mackie 1998, Mass 1999, Merrow 1981, Merrow 1986, Merrow 1988, Semple 1994, Touran 1994). Poor Estimation (errors and omissions) can also lead to underestimation, which subsequently translates into increases in project cost as errors and omissions are uncovered. Estimation documentation must be in a form that can be understood, checked, verified, and corrected. The foundation of a good estimate is the formats, procedures, and processes used to arrive at the cost. Poor estimation includes general errors and omissions from plans and quantities as well as general inadequacies and poor performance in planning and estimation procedures and techniques. Errors can be made not only in the volume of material and services needed for project completion but also in the costs of acquiring such resources (Arditi 1985, Booz·Allen

22 1995, Carr 1989, Chang 2002, Harbuck 2004, Hufschmidt 1970, Merrow 1981, Merrow 1986, Merrow 1988, Pickrell 1990, Pickrell 1992). Inconsistent Application of Contingencies causes confusion as to exactly what is included in the line items of an estimate and what is covered by contingency amounts. Contingency funds are typically meant to cover a variety of possible events and problems that are not specifically identified or to account for a lack of project definition during the preparation of early planning estimates. Misuse and failure to define what costs contingency amounts cover can lead to estimation problems. In many cases it is assumed that contingency amounts can be used to cover added scope and planners seem to forget that the purpose of the contingency amount in the estimate was lack of design definition. DOTs run into problems when the contingency amounts are applied inappropriately (Noor 2004, Ripley 2004, Association 1997). Planning—External External factors that can lead to underestimation of project costs during the planning portion of project development include local government concerns and requirements, effects of inflation, scope creep, and market conditions. Again it must be recognized that each of these factors can act separately or in combination with others to cause significant project costs increases. Local Government Concerns and Requirements typically include mitigation of project effects and negotiated scope changes or additions. Actions by the DOT are often required to alleviate perceived negative impacts of construction on the local societal environment as well as the natural environment. Measures may include but are not limited to introducing changes to project design, alignment, and the conduct of construction operations. These steps are often taken to appease the local residents, business owners, and environmental groups. The required accommodation is often unknown during the early stages of project development. There are a multitude of examples of “drastic” measures that were taken to accommodate local government and citizen concerns as well as national concerns with two of the most notable examples being actions during the Legacy Highway project in Utah and the Big Dig in Massachusetts (Board 2003, Booz·Allen 1995, Callahan 1998, Chang 2002, Daniels 1998, Harbuck 2004, Hudachko 2004, Legacy 2004, Mackie 1998, Mass 1999, Merrow 1981, Merrow 1986, Merrow 1988, Parsons 2002, Schroeder 2000, Touran 1994). Effects of Inflation is a key factor in the underestimation of costs for many projects. The time value of money can adversely affect projects when: 1) project estimates are not communicated in year-of-construction costs; 2) the project completion is delayed and therefore the cost is subject to inflation over a longer duration than anticipated; and/or 3) the rate of inflation is greater than anticipated in the estimate. The industry has varying views regarding how inflation should be accounted for in the project estimates and in budgets by funding sources. In the case of projects with short development and construction schedules, the effect of inflation is usually minor, however projects having long development and construction durations can encounter unanticipated inflationary effects. The results of inflation effects are evident in Boston’s Big Dig. The original estimate for this project, which was developed in 1982 and based on the FHWA guidelines in the Interstate Cost Estimate (ICE) manual, excluded inflationary factors. Inflation is a large portion of the cost overruns experienced on the project (Akinci 1998, Arditi 1985, Board 2003, Booz·Allen 1995, Hufschmidt 1970, Merrow 1988, Pickrell 1990, Pickrell 1992, Touran 1994).

23 Scope Creep is similar to changes in scope; however, these changes are usually the accumulation of minor scope changes. Projects seem to often grow naturally as the project progresses from inception through development to construction. These changes can often be attributed on highway projects to the changing needs or growth of the population in the area to be served (Akinci 1998, Board 2003, Booz·Allen 1995, Callahan 1998, Chang 2002, Harbuck 2004, Hufschmidt 1970, Mackie 1998, Mass 1999, Merrow 1981, Merrow 1986, Merrow 1988, Semple 1994, Touran 1994). Market Conditions or changes in the macroenvironment can affect the costs of a project, particularly large projects. Often only large contractors or groups of contractors can work or even obtain bonding for a large project. The size of the project affects competition for a project and the number of bids that a DOT receives for the work. Typically, the risks associated with large projects are much greater, both for the owner and contractor, and that affects project costs. Inaccurate assessment of the market conditions can lead to incorrect project cost estimation (Summary of Independent Review 2002, Woodrow 2002). Execution—Internal Although this study focuses on developing better practices for early estimates, cost growth occurring during the construction of a project cannot be ignored and must be planned for during project development. Internal factors that lead to the underestimation of project costs during the execution of a project stem from poor project management and design documents. More specifically, these factors can include inconsistent application of contingency, faulty execution, ambiguous contract provisions, and contract document conflicts. Inconsistent Application of Contingency can be both an internal factor contributing to underestimation during the planning stage and a contributor to cost overruns during the execution of the project. During the project execution contingency funds, instead of being applied to their dedicated purpose are inappropriately applied to construction overruns and then not available for their intended purpose (Noor 2004, Ripley 2004). Faulty Execution by the DOT in managing a project is one factor that can lead to project cost overruns. This factor can include the inability of the DOTs representatives to make timely decisions or actions, to provide information relative to the project, and failure to appreciate construction difficulties cause by coordination of connecting work or work responsibilities (Board 2003, Callahan 1998, Chang 2002, Merrow 1981, Merrow 1986, Touran 1994). Ambiguous Contract Provisions dilute responsibility and cause misunderstanding between the DOT and project constructors. Providing too little information in the project documents can lead to cost overruns during the execution of the project. When the core assumptions underlying an estimate are confused by ambiguous contract provisions forecast accuracy cannot be achieved (Callahan 1998, Chang 2002, Department 1994, Harbuck 2004, Mackie 1998, Measuring 1998, Tilley 1997, Touran 1994). Contract Document Conflicts lead to errors and confusion while bidding and later during project execution they cause change orders and rework. (Callahan 1998, Chang 2002, Department 1994, Harbuck 2004, Mackie 1998, Measuring 1998, Tilley 1997, Touran 1994)

24 Execution—External External factors that lead to the underestimation of project costs during the execution of a project stem from those items that are primarily out of the control of the highway agencies. External factors in the project execution stage include local government concerns and requirements, unforeseen events, unforeseen conditions, and market conditions. Local Government Concerns and Requirements can affect the project costs during the execution phase. Similar to the effects during the planning phase, mitigation actions imposed by the local government, neighborhoods, and businesses as well as local and national environmental groups during the construction of a project can extend the project duration affecting inflation allowances or add direct cost. By not anticipating these changes, DOTs can be plagued by project cost increases (Board 2003, Booz·Allen 1995, Callahan 1998, Chang 2002, Hall 1980, Mackie 1998, Mass 1999, Merrow 1981, Merrow 1986, Merrow 1988, Pearl 1994, Sawyer 1951- 52, Summary of Independent Review 2002, Touran 1994, Woodrow 2002). Unforeseen Events are unanticipated and typically not controllable by the DOT, occurrences such as floods, hurricanes, tornadoes, or other weather related incidents. Typically these are called “acts of god.” These acts can bring construction to a standstill and have been known to destroy work creating the need for extensive rework or repair. Events controlled by third parties that are also unforeseen include terrorism, strikes, and changes in financial or commodity markets. These actions can have devastating results on projects and on project costs (Akinci 1998, Arditi 1985, Callahan 1998, Chang 2002, Hufschmidt 1970, Merrow 1981, Merrow 1986, Merrow 1988, Semple 1994, Touran 1994). Unforeseen Conditions are notorious for causing cost overruns. Unknown soil conditions can effect excavation, compaction, and structure foundations. Contaminated soils may be present. Utilities are often present that are not described or described incorrectly on the drawings. There are a multitude of problems that are simply unknown during the planning stage and which can increase project cost when they become apparent during construction (Akinci 1998, Arditi 1985, Callahan 1998, Harbuck 2004, Hufschmidt 1970, Merrow 1981, Merrow 1986, Merrow 1988, Semple 1994, Touran 1994, Transportation 1999). Market Conditions affect the project costs during the execution phase similar to the effects during the planning phase. Changing market conditions during the construction of a project that reduces the number of bidders, affects the labor force, and other related elements can disrupt the project schedule and budget (Board 2003, Booz·Allen 1995, Callahan 1998, Chang 2002, Hall 1980, Mackie 1998, Mass 1999, Merrow 1981, Merrow 1986, Merrow 1988, Pearl 1994, Sawyer 1951-52, Summary of Independent Review 2002, Touran 1994, Woodrow 2002). Current Practice in Cost Estimation Practice and Cost Estimation Management Departments of Transportation attempt to mitigate the factors leading to cost escalation through their prescribed cost estimation practice and cost estimation management systems. These practices and systems are employed across the spectrum of project development, from the conception of an idea to address a need to construction of the project. DOTs also have requirements related to planning and programming their projects and eventually committing funds to projects as the target letting date approaches. As a consequence of this requirement, cost

25 estimates must be prepared to support long-range plans, authorized programs, and funds for State Transportation Improvement Programs (STIP). According to FHWA, requirements the long- range plan must be at least 20 years (Anderson and Blaschke 2004). The first three years of this long-range plan is typically the STIP. The STIP must be at least three years. A DOT’s authorized program varies between four years and twelve years where the first three years are the STIP. In some DOTs, the STIP may be longer than three years and may constitute the authorized program. Other states may have projects that are programmed in later years, that is, beyond the STIP such as, for example, 10-year authorized program where the first four years are included the STIP. Those years beyond the authorized program would include up to 20 years or more of projects depending on DOT policies and procedures. The DOT must therefore align their estimation practices and cost management systems to fit within their long-range planning, priority programming, and preconstruction processes. The first project development phase is long-range planning. Most of the DOTs interviewed employed conceptual estimation techniques based on cost per mile factors, while a smaller number of DOTs used a typical or similar project to arrive at a planning estimate. If a project has structures, the DOT would use a cost per square foot of bridge deck for this project component. The DOTs use this planning estimate as the stated “order of magnitude cost” of the project when their transportation project needs list is developed. The estimation procedure for the programming estimate varies among the DOTs. These cost estimates often become the stated project cost included in the department’s authorized program, and in many cases the program and project costs must be approved for funding at this point by the legislature. Parametric estimation techniques are used for this estimate based on concept drawings and factors covering significant cost elements in the project scope such as pavements, bridges, and right-of-way. The advanced planning/preliminary design phase begins when the DOT commits resources to developing design documents for a project. The estimation procedures used during the early project design phase depend on the completeness of the design, that is, percent of design complete. At the early stages of design, estimates are prepared in a manner similar to the programming estimation approaches (parametric based on lane mile factors, bridge deck square foot/yard, or similar projects). As the design becomes more definitive, the estimation procedure evolves from a parametric estimation process to a line-item approach. These estimates are often used as the basis for project funds included in the STIP. Preliminary design estimates are typically prepared before each formal design review (30%, 60%, and 90% design reviews are required by many DOTs). The final estimate is the engineer’s estimate, which is created when the design is 80 to 100 percent complete. The engineer’s estimate is used to evaluate the bid prices submitted by the contractors. Table 2.3 summarizes general characteristics of state DOT cost estimation practice and cost estimation management characteristics. These characteristics are further explained in the following sections.

26 Table 2.3. Summary of Cost Estimation Practice and Cost Estimation Management Characteristics Project Development Phase Cost Estimation Practice Cost Estimation Management Estimate Purpose Estimate Preparation Estimate Reviews Estimate Communication Cost Management Planning (Conceptual Estimate) Estimated funds needed for long range plan Cost/Mile & Percentages Internal Review or MPO Review Point or Range Promote Transparency and Integrity Programming (Parametric Estimate) Estimated funds for project in authorized program Cost/Mile, Percentages, & Defined Line Items Internal Review Point Checklist & Tracking System Advanced Planning/Preliminary Design (Parametric and Line-Item Estimate) Estimated funds for project in STIP Defined Line Items & Percentages Peer and Team Reviews Point Management Accountability & Scope Control Final Design (Detailed Engineer’s Estimate) Estimated construction cost to compare with bids Completely Line Item Committee Review Point Checks within Estimating Software Cost Estimation Practice The initial project phase, planning, identifies the need for a project. This need has little definition, which affects the estimation method used to arrive at an estimate of project cost. The main method or approach used for long-range planning estimates is lane-mile cost factors. The cost per mile factor is developed using different methods such as historical lane-mile sections, similar projects, or volumetric factors. The cost is based typically on historical data derived from the bid prices (not actual project cost), either award or averages of several bidders. The long- range planning estimate is often prepared using only basic computerized tools, a Department developed Excel spreadsheet. Many of the spreadsheets used are templates with predetermined formulas and historical data incorporated into the spreadsheet. The DOTs also have a cost per mile document that lists project types, such as a four-lane roadway in a rural area, and the cost that corresponds with the project type. A few DOTs have developed typical and/or standardized sections that correspond to a cost per mile section. Other project elements such as right-of-way, engineering, environmental, and miscellaneous items are incorporated into the planning estimate as a percentage of the total project cost or as a contingency factor. For example, in the case of preliminary engineering, 0.5 to 8 percent is added depending on project complexity, and the utility cost is 3 percent of the total cost. The estimate may or may not be inflated to the midpoint of construction year. In most cases, the planning

27 estimate undergoes very little review within the Department. If the estimate is reviewed, the review is conducted by another person on the estimation team or by an engineer in a district office. However, if the estimate was prepared for a metropolitan planning organization (MPO) it may be reviewed by the MPO during the project selection process for the long-range plan. Programming estimates are produced in a similar manner as the long-range estimates but these estimates are based on more specific definition of project scope. The programming estimate amount often becomes the DOT’s cost number included in its authorized program. DOTs typically use cost per mile factors, and percentages to create the programming estimate as was the approach for preparing the planning estimate. However, this estimate is evolving into a parametric estimate and beginning to include defined project items, especially for the major cost items such as paving and structures. A parametric estimate is an estimate that is based on a broad breakdown into key components of the project and parameters like length of project, width of roadway, or depth of pavement. This information is derived from conceptual drawings. Furthermore, some DOTs use conceptual and parametric estimation software that has been developed by the department. Other “Add-on” elements, such as local government concerns, environmental issues, and externally imposed requirements, also receive their first recognition in this estimate. To produce the programming estimate, historical bid data is often the primary source of cost information. The data utilized may be sorted by state, region, or DOT district. Some DOT databases have the capability of being arranged by market area, terrain, and project type. The programming estimates can be created in current year dollars and then inflated to some mid-point of construction time period. After the programming estimate is complete, it does not usually go through a formal review process but typically members of the project team review the estimate internally. If a change has occurred that causes the estimate to increase, then the changes above certain percentages initiate another review of the project within the DOT. The preliminary design estimate is an amplification of the programming estimate. For this estimate, DOTs begin to create increasingly more detailed line-item estimates. At this phase, actual design quantities begin to replace previous quantity assumptions. Once the project is in the design phase and the right-of-way limits set, the right-of-way and utility costs can be refined based on specific design information (e.g., parcels). As the level of design increases, the estimate is further refined. A preliminary design estimate is updated when the scope reaches established design milestone or a significant element in the scope has been identified. At some point, the preliminary estimate is the basis for funds included in the STIP. Project estimate preparation can also follow major milestones of project development, such as project initialization, conceptual plan/environmental document completion, preliminary plan completion, right-of-way plan completion, and contract plans completion (PS&E). The difference between each estimate produced during design development is that more line items are identified, as the project scope is refined. At the preliminary design stage, the estimation calculations may be performed using a spreadsheet or in-house computer software. The same historical data used in the programming estimate is applied to the preliminary estimates. The design team is ultimately responsible for the quality and accuracy of the estimates they create. However, the review process begins to become more formalized as design proceeds. Peer and project team reviews, which are often led by the project manager, occur. The project manager approves the estimate, and the district or region often reviews it. The Department’s central office

28 will review the estimate if it has increased beyond specified limits. The cost growth limits that trigger additional reviews or approvals are established by internal DOT policies. The engineer’s estimate is the final estimate before a project is advertised, and it is used to judge the contractors’ bids. This estimate is performed using complete plans, specifications, and other project information. Estimation software such as AASHTO’s Trns·port software or an in-house program is used to generate the engineer’s estimate. There are basically three approaches used to develop the final line-item (pre-bid) engineer’s estimate (Contract 2001). The use of historical data from recently awarded contracts is the most common approach. Under this approach, bid data are summarized and adjusted for project conditions (i.e., project location, size, quantities, etc.) and the general market conditions. However, this method is the most susceptible to outside factors such as inflated bid prices from contracts with little or no competition (Contract 2001). The detailed estimate approach based on specific crews, equipment, production rates, and material costs (also termed cost-based estimation). This is similar to the way a construction contractor would estimate a project. This approach requires the estimator to have a good working knowledge of construction methods and equipment. While adjustments for current market conditions may be required, this approach typically produces an accurate estimate and is useful in estimating unique items of work where there is insufficient bid history (Contract 2001). The third approach combines the use of historical bid data with actual cost development. Most projects contain a small number of items that together comprise a significant portion of the total cost. These major contract items may include Portland cement concrete pavement, structural concrete, structural steel, asphalt concrete pavement, embankment, or other specialty items. Prices for these items are estimated using the detailed approach and adjusted for specific project conditions. The remaining items are estimated based on historical prices and adjusted as appropriate for the specific project (Contract 2001). If the design team prepares the engineer’s estimate, then it undergoes a district or regional review and more than likely a central office review. Some estimates are reviewed by estimation committees that are composed of personnel that have specific knowledge about different aspects of a project and ranges of experience. If the agency’s central office prepares the engineer’s estimate, then they also review the estimate. Project Complexity Most DOTs describe project complexity by preservation projects, medium sized to large rehabilitation and reconstruction (mid-range) projects, and large mega projects (greater than $100 million). Project complexity is also characterized by the project’s anticipated cost. For example, one DOT divides their projects into smaller maintenance projects estimated to be less than 5 million dollars, widening projects ranging from 20 to 30 million dollars, and large projects ranging from 60 to 80 million dollars. DOTs do not consider projects such as preservation projects to be a significant issue in cost estimation, because the DOT typically has a good idea of the project elements and quantities associated with preservation projects. For the mid-range

29 projects, DOTs do not have a consistent estimation procedure. DOTs consider project characteristics such as the project’s location, but additional costs included to reflect project complexity are dependent on the estimator’s judgment and experience. For large or mega projects, DOTs are forced to consider project complexity. DOTs conduct constructability reviews, value engineering reviews, and evaluate several alternative design concepts for mega projects. Due to the complexity of the mega projects, DOTs have to perform some conceptual development before they can select an appropriate alternative and cost for that alternative. One DOT has developed separate policies for major projects and minor projects. For major projects, this DOT requires a draft scoping memorandum, a final scoping memorandum, more approval signatures, and extensive environmental documents. Cost Estimation Management Scope creep and major scope changes are significant factors that drive project cost increases. Managing the scope and schedule of a project is an essential element in alleviating cost escalation. To control and overcome these problems, project scope and schedule must be communicated and managed. In order to manage a project’s scope and schedule, therefore cost, departments attempt to promote transparency by identifying major issues contiguous to the project and project uncertainties. When long-range planning estimates are created, most states do not have regulations requiring the estimates to be approved by the legislature. The long-range estimates are simply communicated to higher Department management as a single number dollar amount. Only occasionally is the estimate communicated as a cost range, which typically is result from the performance of a risk analysis. During the programming phase, the project is often included in an authorized program with the project estimate. The project is usually assigned a tracking or project number. The project may or may not be included in the STIP at this point. Since the programmed estimate is often the first estimate entered into the tracking system, future estimates are compared to the baseline estimate. Departments have project checklists for standard project elements. These checklists call the planner’s attention to important cost items like pavement, right-of-way, demolition, traffic control, utilities, and engineering. As an estimate is prepared, the estimator makes certain that the costs of the listed items have been included in the project estimate. The project cost is further managed by the use of a risk charter. A risk charter is a list of identified risks that may be encountered during the life of the project. The charter may address the likelihood of the risk, the cost and schedule implications of the risk, and mitigation technique suggestions. The goal of the risk charter is to reduce the number of risks on the list to as few as possible, by mitigation strategies or project design changes. If the project was not added to the STIP during the programming phase, then it is included during the preliminary design phase. The estimate in the STIP is a single number, and it is open to public awareness. Therefore, management is held accountable for the estimate in the STIP. This accountability causes management to increase control over scope modifications. During the preliminary design phase, the project estimates and checklists created in the programming stage are being updated to reflect recently defined project elements. The scope and estimate is monitored with checklists and red flags caused by changes.

30 Release of the DOT’s Engineer’s Estimate In many States, the engineer’s estimate is not released to the public before the letting. What is allowable concerning release of the DOT’s estimates is usually defined by State statute and, in many cases, out of the DOT’s control. Once the bids have been submitted to the DOT, the department uses estimation software to compare the engineer’s estimate with the bids. By Law or internal rules DOT’s require the bids to fall within a certain range of their engineer’s estimate, or they will not award the contract. After the bids have been compared to the DOT’s estimate, the total amount of the DOT estimate is usually released to the publicly, most DOT’s however, do not release the detail item prices of their estimate. Overall Estimation Characteristics The organizational structure of a state DOT affects the development of the project estimates. A department’s organization determines where the estimate is created, who reviews and approves the estimate, when the estimate is communicated, and the process of how the estimate is prepared. In most cases, the initial project estimate is prepared at a district or region office and that office retains responsibility for project development and creating subsequent estimates. When the project reaches the later stages of development, it is handed over to the region or central office for letting. Although the districts or regions lead the project development process, the region or central office provides the districts with oversight, but this can often be minimal. States with large construction and maintenance programs are extremely decentralized, and their districts perform almost as separate entities. A few states have unique characteristics such as requiring at least one person retain responsibility for the project throughout its life or having a State Estimator’s Office that oversees all project estimates. The organization of the DOT also affects how historical cost data is managed. Some agencies have separate databases for each district. However, most DOTs have created a central database that can be accessed by all districts within the state. To generate unit-price data departments systematically compile bid data from past project lettings. These data are broken down by bid line item. The database users can sort the historical bid process by state, region, or district averages. Furthermore, users have the ability to organize the data by market area, terrain, or project type. Another characteristic that is unique for each DOT is how they define and apply contingency. Contingency covers a range of issues such as scope changes, scope increase, high-risk elements, and unforeseen site conditions. For every project development phase, the amount of contingency incorporated into an estimate is established by the DOT. The three methods to determine contingency are listed below: 1. Fixed Percentage, 2. Sliding Scale, and 3. Structure/Formal Analysis. A fixed percentage is a single percentage that is applied to every estimate prepared, and the set percentage could range from zero to ten percent. If a transportation agency uses a sliding scale,

31 then they apply a large percentage to the conceptual estimate and decrease the percentage as the project scope is defined. For example, 50% is added to the long rang planning estimate, and then 25% is added to the programming estimate. During the preliminary design phase, the percentage continues to lower until the design is complete, at which time contingency is not included. The final contingency application is a contingency determined by a structural/formal analysis, such as a Monte Carlo simulation. The state agency performs the risk analysis that identifies the level of risk for each project. Then, the analysis is related to the amount of contingency needed to sufficiently cover the risk. Once the project has entered the final design stage, most agencies do not include any contingency regardless of their methodology. CONCLUSIONS This chapter has described the state of practice in highway cost estimation practice and cost estimation management. The research team arrived at this state-of-the-practice review through an exhaustive literature review and in depth interviews with DOTs, MPOs, transportation consultants, and non-transportation owners. Additionally, the team has defined the factors that cause cost escalation and summarized them into 18 total internal and external factors that occur during the project planning and execution phases. Chapter 3 will provide a critical review of the state of practice through an analysis of how the current DOT cost estimation practices and management relate to the factors that cause cost escalation.