In July 1996 the General Accounting Office told Congress that Massachusetts had just published an official estimate of $7.8 billion for the Central Artery/Tunnel project, and that the figure left out more than $1 billion of costs earlier estimates had included while making no allowance for inflation. Put those back and the comparable number was $10.4 billion. Before any of the famous overruns, the first quantity that moved on this project was the boundary of the estimate itself.
That is the reason so much writing about the Big Dig is unusable. The numbers quoted are real numbers taken from real documents, and they measure different things in different dollars across two decades. Reassembling them in an order that means something is most of the work.
Which dollars, and which decade
The clearest published reconciliation is the 2003 review by a National Research Council committee under the Board on Infrastructure and the Constructed Environment, carried out at the Massachusetts Turnpike Authority’s own request. It records the original estimate as $2.6 billion in base-year 1982 dollars, the figure attached to the 1985 final environmental impact statement, and the estimate at the time of writing as $14.6 billion in current dollars, which the same passage converts to $8.0 billion in 1982 dollars.
Both of those sentences are true, and only one pairing is a like-for-like comparison. Measured in constant dollars the project grew roughly threefold. Measured by pitting an unescalated 1982 figure against an escalated 2003 one it grew more than fivefold, which is the version that travels. Anyone using the second comparison is attributing to mismanagement a large amount of ordinary inflation across twenty-one years.
The thing being built was substantial on any measure: 161 lane miles, almost half of them underground, inside 7.5 miles of right-of-way. Completion was originally planned for 1998 and by 2003 was expected in 2005.
A total with an auditor’s name on it, and what sits outside it
For a figure with an auditor’s name attached, the October 2000 finance plan is the document to use. The Department of Transportation’s Inspector General reviewed it at Congress’s direction and reported on 29 November 2000 that the estimated completion cost of $14.075 billion, made up of $13.8 billion plus a $258 million contingency, was consistent with independent estimates prepared by FHWA and by Deloitte & Touche and appeared reasonable. FHWA accepted the plan the same day, against a substantial completion date of December 2004.
The Inspector General also published the assumptions the total rested on, which is what makes it checkable rather than merely official: a market discount rate of 7 percent, an escalation rate of 2.35 percent and a potential change allowance of 24 percent. Change orders at nearly a quarter of contract value were, on this project, a planning assumption.
Two boundaries around that number matter. Federal participation had been capped at $8.549 billion by a partnership agreement signed on 22 June 2000 and repeated in the appropriations conference report for 2001, which left the Turnpike Authority and its partners carrying every dollar above the cap. And the total is a project cost, not a public cost. Of the $9.122 billion spent by 30 June 2000, $1.114 billion came from grant anticipation notes, which is borrowing against future federal aid. Debt service on that borrowing sits outside every figure quoted here, and outside the totals most commonly repeated.
The reporting failure that wrote today’s rules
The federal requirement to file a finance plan at all was young. It arrived with the Transportation Equity Act for the 21st Century in 1998 and covered projects above $1 billion, and the Central Artery’s October 1998 plan set out total costs of $11.7 billion.
What happened next is the reason the modern regime looks the way it does. The project submitted its October 1999 update on 7 January 2000 and FHWA approved it on 1 February 2000. Later that same day the Turnpike Authority announced that costs would rise by a further $1.4 billion, an increase the approved plan had not mentioned. A replacement plan filed in March was rejected on 8 May 2000 because FHWA’s own review found that costs would likely exceed even that increase by another $300 million to $480 million. Revised federal guidance for finance plans on large projects followed on 23 May 2000, and the funding cap arrived in June.
Every element of the present system descends from that year. Under 23 U.S.C. 106(h) a build of $500,000,000 or more must produce a management plan for the work together with a financial plan refreshed every year against detailed estimates of the cost to finish. Federal guidance now expects a probabilistic cost estimate review before the first plan and requires the reported total to reflect the 70th percentile of that review. What such a filing looks like when it is working is visible in the Hampton Roads expansion, whose plan states in writing that its contractor is behind on production rates.
What failed underground
The construction defects surfaced in two events. In September 2004 a slurry wall panel in the I-93 mainline tunnel breached, and federal and state prosecutors opened the investigation that would run for the next three years.
The second event was fatal. About 11:01 p.m. on 10 July 2006, a section of suspended concrete ceiling in the I-90 connector tunnel detached and fell onto a passing car, some 26 tons of concrete and suspension hardware in all. The front-seat passenger was killed.
The National Transportation Safety Board found the probable cause to be the use of an epoxy anchor adhesive with poor creep resistance, a formulation not capable of sustaining long-term loads, which deformed and fractured over time until anchors pulled free. The chain behind that single material choice is long. The designer and the management consultant had not identified creep as a critical long-term failure mode or accounted for it in the specifications and approval process. The construction community generally lacked understanding of creep in adhesive anchoring systems. The manufacturer had not supplied complete and accurate information about the suitability of its fast set product for sustained tension, and had failed to recognize that the anchor displacement found in the high occupancy vehicle tunnel in 1999 was creep caused by the same product. Nobody kept monitoring the anchors afterwards.
Had the Massachusetts Turnpike Authority, at regular intervals between November 2003 and July 2006, inspected the area above the suspended ceilings in the D Street portal tunnels, the anchor creep that led to this accident would likely have been detected, and action could have been taken that would have prevented this accident.
What the contractors admitted
On 23 January 2008 the joint venture that had managed construction agreed to pay over $407 million to resolve criminal and civil liabilities, with 24 design consultants adding $51 million, for a total recovery to the United States and the Commonwealth of $458 million including interest. Most of it went into a new state trust fund for non-routine repair and maintenance of the tunnels.
The admissions matter more than the sum. The joint venture acknowledged allowing concrete to be placed in the I-93 slurry walls when specifications had not been met, failing to document or correct the deficiencies, and certifying the safety and substantial completion of that tunnel in documents that were not true and accurate. On the connector tunnel it acknowledged observing epoxy bolts creeping out of the roof and failing to investigate the cause adequately. The settlement left liability open for a further ten years in the event of a catastrophic failure.
The rule that came out of it
The safety board’s recommendations were specific about the gap. It asked FHWA to seek legislation authorising a mandatory tunnel inspection program similar to the national bridge program, and then to build one that identified critical inspection elements and set an appropriate frequency. MAP-21 supplied the authority. The National Tunnel Inspection Standards were published on 14 July 2015 and took effect on 13 August 2015.
They require a national tunnel inventory, routine inspection of tunnels on all public roads including tribal and federal ones, written reports of critical findings with notification to FHWA within 24 hours, inspector training and a national certification program, and timely correction of deficiencies. Routine inspection defaults to 24 months. An owner may inspect more often without asking, and less often only after a written case considering tunnel age, time since major rehabilitation, complexity, traffic, geotechnical conditions, functional systems and known deficiencies. FHWA modelled the whole structure on the bridge standards; the parent regime and its own 2022 rewrite are covered in how bridges are inspected and rated.
The part the reforms did not touch
Both regimes address what happens after a project is designed. Neither improves the number the public argues about, which is generated years earlier, before any review can be held against a design that does not yet exist. The 1982 estimate of $2.6 billion, carried in the 1985 environmental impact statement, was a base-year figure for a project whose scope was unsettled, and it anchored expectations for two decades regardless.
The pattern has not gone away. The Interstate Bridge Replacement program’s 2020 conceptual estimate reached a high end of $4.8 billion; the same corridor is now put at $13.5 billion to $15.2 billion. No rule written after Boston would have flagged the first figure, because early estimates are not what the rules govern. The general problem of comparing such numbers runs through the projects and case studies section and is set out directly in the survey of the largest current builds.
The Big Dig’s real bequest to American practice is a filing cabinet: annual plans, probabilistic reviews, tunnel inventories, certified inspectors. It made megaprojects legible without making them cheaper or faster. Legibility is worth having on its own, and the case for it is simply what the alternative looked like on 1 February 2000, when a federal agency approved a finance plan in the morning and learned of a $1.4 billion increase before the end of the day.