Deconstructing the Big Dig
best practices for mega-project cost estimating
Associate Professor, Metropolitan College, Boston University
We study the cost performance of Boston's US$14.78 billion Central Artery/Tunnel Project, which proved to be one of the largest, most technically difficult and environmentally challenging infrastructure projects ever undertaken in the United States. Numerous obstacles and uncertainties during planning and implementation resulted in the project cost growing from US$2.8 billion to US$14.78 billion.
We first review the literature on cost escalation on mega-projects, and compare the conclusions and observations with the Big Dig. We then analyze detailed cost and schedule data for this mammoth project, beginning with a presentation of the how the costs grew over time. We present some relatively simple techniques for analyzing the data.
We analyze the claim that inflation was a major factor in the cost growth of the big dig, concluding that political rather than cost factors played a major role. We conclude with lessons learned, and present some recommendations for mega–project managers based on the lessons from the Big Dig.
The Big Dig
Boston's US$14.78 billion Central Artery/Tunnel Project, known affectionately as the “Big Dig,” proved to be one of the largest, most technically difficult and environmentally challenging infrastructure projects ever undertaken in the United States. The Big Dig was completed in 2007 and replaced Boston's inner-city infrastructure with new roads, bridges, and tunnels. It was conceived, planned, designed, and constructed over a period of 25 years, from 1982 through 2007. There were numerous obstacles and uncertainties during the planning and implementation of the project that resulted in schedule delays and significant cost overruns.
The Big Dig was conceived in 1985 as a way to alleviate the continuous traffic gridlock in the center of Boston. The original cost estimate was US$2.56 billion. Cost and schedule updates were completed annually, and, over time, the cost gradually increased to US$7.74 billion in 1992, to US$10.4 billion in 1994, with the most recent estimate for 2008 being US$14.8 billion (Massachusetts Turnpike Authority, 2008). There are many reasons for the increase in costs, but by far the most significant were the failure to properly assess the impact of unknown subsurface conditions, environmental and mitigation costs, inflation, and expanded scope (McCormack, 1997). The mitigation alone required 1,500 separate mitigation agreements, all of them unanticipated.
Cost estimation is critical and essential in the long-term financing of large mega-projects, especially since predicting all of the costs upfront is difficult to do. Despite the billions of dollars spent on mega-projects, it is troublesome to note how little comparative research is available on strategies and methodologies for cost estimation management.
The existing literature emphasizes that mega-projects may result in organizational mayhem and are burdened with conflicting interests, stakeholder disputes, and misguided decisions that result in delays and cost overruns (Kirkland, 1995; Genus, 1997; Gourvish, 2006). Another strand of the literature focuses on governance and support mechanisms (Crawford, Cooke-Davies, & Hobbs, 2008) and examines the different frameworks used in different countries (Klakegg, Williams, Magnussen, & Glasspool, 2008; Flyvbjerg, Bruzelius, & Rothengatter, 2003).
Bent Flyvbjerg, the author of several well-known studies on mega-projects, has noted that cost overruns in major transport infrastructure projects are widespread. The difference between actual and estimated investment cost is often 50% to 100%. Indeed, for many projects, cost overruns threaten the entire viability of the project. Flyvbjerg also notes that the underestimation of costs at the beginning is the rule rather than the exception.
Both the General Accounting Office (2003) and multilateral development banks (MDBs) observe that construction projects have a long history of underestimated budgets and cost escalation. The World Bank (1994) has noted that its projects are typically subject to more careful appraisal than most other infrastructure projects, but even with their relatively rigorous procedures, mega-projects show a consistent pattern of inflated costs.
A review of large public works projects over the last century concluded that they are consistently underestimated, and this phenomenon is attributed to the desire of the project advocates to have their projects approved (Flyvbjerg et al., 2003). Other important reasons for cost increases include inaccurate scope, unreasonably optimistic schedules, and political pressures to stay within budget (Chang, 2002). Large projects also have long lives and idiosyncratic features that contribute to their complexity (Esty, 2004).
Lessons learned from 1990s mega-projects led Congress to include a requirement in the Transportation Equity Act for the 21st Century (TEA-21) that every mega-project (defined as US$1 billion or more) receiving Federal funds have a financial plan that is updated annually. The financial plan must compare the original cost and schedule estimates to actual costs and schedules, and provide reasonable assurance that sufficient resources are available to complete the project as planned.
The Big Dig's Cost History
We begin with an analysis of the Big Dig financial data. Exhibit 2 shows the cost history of the Big Dig. The major feature of Exhibit 1 is the staggering cost growth (blue line).
Inflation on the Big Dig
The impact of inflation on the Big Dig is interesting. The impact of inflation is a critical factor in the underestimation of costs for many projects, and the Big Dig was no exception. It is claimed that a major cost escalation factor on the Big Dig was inflation on all project elements lasting more than a decade (Akinci & Fischer, 1998), and the project management team reported that about half of the cost growth was caused by inflation. In Exhibit 3, we present the published inflation rates over the life of the Big Dig. Using these data, we can project the expected cost of the project based solely on inflationary pressures.
This projected inflationary cost growth is also shown in Exhibit 1 (the red line). We see that the project's estimated cost should grow from the US$2.8 billion to around US$4.5 billion. This is nowhere near the final estimate of US$14.8 billion.
Common causes for cost escalation on the Big Dig included: the failure to include a cost for inflation in each contract; delays in project completion; and the actual rate of inflation being greater than the planned estimate. Other factors that impacted the Big Dig were:
- Financing shortfalls and interest rates
- Scope changes
- Shortages of materials and labor, price increases, and market changes
- Weak project managers
- Technical and design complexity
- Unexpected events and force majeure
- Political and legal risks
Exhibit 1: Big Dig costs over time (blue). The red line shows the projected costs using actual, published inflation rates.
Exhibit 2: Published inflation rates (Consumer Price Index, CPI) over the life of the Big Dig.
Analyzing Big Dig Costs
In this research, it is convenient to divide the construction into two periods. The first period is 1982 to 2000. In 1982, when the idea of the Big Dig was first conceived, the cost estimate was approximately US$2.56 billion (in 1982 dollars). By 1989, the estimated costs had risen to US$4.44 billion (in 1987 dollars), a 73% increase in 5 years. Four years later, in 1993, the cost estimation was US$7.74 billion—a 202% increase!
In 1985, the cost included the original Environmental Impact Statement and a conceptual design. That design did not include several major components that were added later: the Massachusetts Avenue Interchange, linkages to Logan Airport in East Boston, and work north of the Charles River.
The second half of the project was from 2000 to 2007. The year 2000 was important because it followed a 3-year budget freeze, which can be seen as the flat region in Exhibit 2. In 2000, the cost estimate was adjusted three times, resulting in a US$2.65 billion increase and bringing the projected total cost to US$14.08 billion. The project was completed in December 2007, 3 years later than originally scheduled, at a total cost of over US$14.8 billion—almost five times the original estimate.
Exhibit 3 compares the 1999 and 2003 budgets for the major components of the Big Dig construction packages. The Big Dig actually consisted of more than five major project segments, more than 119 separate contracts, and 60 major construction projects. However, as seen in Exhibit 3, the costs are dominated by two categories: the construction costs (CONS) and the design costs (JV).
Exhibit 3: Almost all of the costs on the Big Dig were incurred in two categories: construction (CONS) and design (JV).
The three major components of construction were I-93, I-90, and the I-90/I-93 Interchange. The financial costs of these three components are shown in Exhibit 4, and amount to approximately US$8 billion. The design (JV) accounts for another US$2 billion.
No single cause seemed to have caused the cost growth on the Big Dig. This can be seen in several ways. In Exhibits 3 and 4, all components show similar proportional growth. Cost growth was also distributed uniformly across contracts. This is illustrated in Exhibit 5, where the cost growth of the various contracts making up I-93 is again very similar.
An important conclusion therefore, is that the cost growth was uniform across the entire project. There was no single cause, but a systemic underestimation of costs.
Exhibit 4: The costs of the three major construction components of the Big Dig: I-93, I-95, and the I-93/I-95 interchange.
Cost data on the Big Dig was extracted from a series of reports, all of which were made available through the Freedom of Information Act. The major publications are the Cost and Schedule Updates (CSU) and the Project Management Monthly (PMM) reports.
Cost estimates were updated on a regular basis through the Project Cost and Schedule Updates (CSUs). Estimates were based on analysis and review by multiple levels of project management. The construction estimate process analyzed several factors, including the value of identified claims and changes to date, current schedule status, type and amount of remaining work, amount of remaining risk, status of design issues, and complexity of adjacent or follow-on contractor interfaces. For those items that were not yet fully scoped, the project construction controls participated in the development of those estimates.
In addition, project managers reviewed and quantified the cost of items such as work acceleration, the resequencing of construction work, mitigation, and security and protection.
Unanticipated site conditions included such events as uncharted utilities, obstructions, unexpected groundwater conditions, environmental problems, archeological discoveries, weak soil, and hazardous materials. Design changes occurred because electrical, mechanical, and systems designs lagged behind the civil/structural design and construction. Also, schedule delays in one segment required revised roadways in another area to support traffic staging and assure effective traffic flow. As milestones changed on one segment, there were delays on other segments that impacted costs in a major way.
Project costs were estimated by project element, which included design, right of way, construction, project management, insurance, force accounts, and contingencies. A change in one element could force an increase or decrease in other elements. For example, as scope and construction claims increased, insurance premiums also increased. As abutters, other government agencies and business owners made demands on the project in the form of mitigation requests. Compensation for government takings, including the relocation of numerous sewer and water lines, meant that agreements had to be entered into to address these concerns, often causing substantial delays and rework.
Exhibit 5: The growth of the contracts making up I-93.
Big Dig Cost Estimation and Tracking
Major elements of the project that were not included were: environmental mitigation, complex work with ground freezing and tunnel box construction in the Fort Port Channel, tunnel roofing for South and East Boston, rebuilding the Dewey Square Tunnels, new interchanges at Logan Airport, temporary ramps, and substantial mitigation costs for Gillette World Headquarters and abutters all along the alignment. Direct costs for environmental requirements and scope changes were more than US$5.5 billion (Anderson et. al., 2006).
The major lessons learned about cost estimation on the Big Dig are:
• The importance of calculating inflation from the inception of the project.
• Monitoring the impact of scope changes.
• Critical factors in cost estimation can be developed by studying the Big Dig.
• Although factors differ from one project to another and unpredictable events can occur, there are similar factors that cut cross all projects that are important to address.
Based on these lessons, we can recommend the following actions for other mega-projects:
1. Study the historical data from mega-projects—the patterns in the data are valuable indicators of trouble.
2. Recognize the limitations of the assumptions in historical projects with comparable characteristics.
3. Identify the attributes of the project that will grow and change over time.
4. Recognize that the accuracy of cost estimates vary throughout the project.
5. Adopt a baseline for cost control during inception and update the baseline when schedule, scope, and quality change.
6. Estimate inflation at the inception of the project for the entire duration of the project.
7. Enforce project standards and requirements on all project contractors.
8. Utilize management reserves solely within the framework for which they are maintained.
Finally, the most important recommendation is to establish an open and transparent process throughout the project. This is a necessity for accurate and realistic cost estimation and budgeting. This should result in more successful projects and greater benefits for all project stakeholders.
Akinci, B. & Fischer, M. (1998). Factors affecting contractors' risk of cost overburden, Journal of Management in Engineering, 14(1), 67–76.
Anderson, S., Molenaar, K. & Schexnayder, C. (2006) Final Report for NCHRP Report 574: Guidance for Cost Estimation and Management for Highway Project During Planning, Programming, and Preconstruction. National Cooperative Highway Research Program, Transportation Research Board of the National Academies, 52, 55-56.
Chang, A. S. (2002). Reasons for cost and schedule increases for engineering design projects. Journal of Management in Engineering, 18(1), 29–36.
Crawford, L. Cooke-Davies, T., & Hobbs, R. (2008). Governance and support in the sponsoring of projects and programs. Project Management Journal, 39, S43–S55.
Esty, B. (2004). Modern project finance. New York: John Wiley & Sons, Inc.
Flyvbjerg, B., Bruzelius, N., & Rothengatter, W. (2003). Mega projects and risk, an anatomy of ambition. Cambridge, UK: Cambridge University Press.
Flyvbjerg, B., Holm, M., & Buhl, S. (2002). Underestimating costs in public works projects, error or lie? Journal of the American Planning Association, 68(3), 279–295.
General Accounting Office. (2003). Federal-aid highways cost and oversight of major highway and bridge projects: issues and options, GAO-03-764T, Washington, D.C.
Genus, A. (1997). Managing large-scale technology and inter-organizational relations: The case of the Channel Tunnel. Research Policy, 26, 169–189.
Gourvish, T. (2006). The political economy of the Channel tunnel: An international business-government perspective. Business and Economic History Online, 4.
Kanabar, V., & Warburton, R. (2008). MBA fundamentals: Project management. New York: Kaplan Publishing.
Kirkland, C. (1995). The Channel Tunnel - Lessons Learned, Journal of Tunnelling and Underground Space Technology, Elsevier Science Ltd., 10(1), 5-6.
Klakegg, O. J., Williams, T., Magnussen, O. M., & Glasspool, H. (2008). Governance frameworks for public project development and estimation. Project Management Journal, 39, S27–S42.
Massachusetts Turnpike Authority. (1994–2008). Cost and Schedule Update (CSU) rev. 6, Central Artery/Tunnel Project, Boston, MA.
Massachusetts Turnpike Authority. (1999-2007). Project Management Monthly. Central Artery/Tunnel Project, Boston, MA.
McCormack, J. W. (1997). Report to the Legislature on Managing the Central Artery/Tunnel Project: An Exploration of Potential Cost Savings. The John W. McCormack Institute of Public Affairs, University of Massachusetts, Boston.
Project Management Institute (PMI). (2008). A guide to the project management body of knowledge (PMBOK® guide)—Fourth edition. Newtown Square, PA: Author.
World Bank. (1994). World Development Report 1994: Infrastructure for development. Oxford, UK: Oxford University Press, p. 17.
© 2009, Virginia Greiman and Roger Warburton
Originally published as a part of 2009 PMI Global Congress Proceedings – Orlando, Florida.