Our review of the Department of Transportation's (Caltrans) Toll Bridge Seismic Retrofit Program (program) found that:
Caltrans neglected several important aspects of generally accepted standards for project management. Specifically:
Many of California's largest and most complex bridges are located in areas of high seismic risk, such as the San Francisco Bay Area (Bay Area). After the 1989 Loma Prieta earthquake, the Legislature imposed higher seismic safety standards for publicly owned bridges, and the Department of Transportation (Caltrans) already has retrofitted most of them. Caltrans determined that seven state-owned toll bridges would need seismic retrofitting. For the San Francisco-Oakland Bay Bridge (Bay Bridge), Caltrans decided that the west span needed retrofitting and the east span (East Span) needed replacing in order to satisfactorily serve as a "lifeline structure," which continues in service after an earthquake. Caltrans' 2001 cost estimate to complete all this work, established by Chapter 907, Statutes of 2001—Assembly Bill 1171 (AB 1171)—amounted to $5.1 billion, including a $448 million program contingency reserve to cover unforeseen costs. At that time, Caltrans projected that the program would be completed in 2009, but now the program faces a financial crisis and is jeopardized by sharp increases in cost estimates and major schedule delays. In August 2004, Caltrans informed the Legislature that the program could cost $8.3 billion—$7.4 billion for specific projects and a $900 million contingency reserve for possible cost increases on those projects—and would be completed no sooner than 2013.
The largest contributor to the estimated $3.2 billion cost increase has been the East Span component known as the signature span. Of this $3.2 billion, $930 million is attributable to the May 2004 bid to build the superstructure of the signature span. The remaining $2.3 billion is due to factors unrelated to the superstructure bid, such as $556 million in additional Caltrans support costs and the need for a $900 million program contingency reserve above the $7.4 billion Caltrans has allocated to specific projects. Increased timelines and schedule delays in contracting for the signature span have been the greatest factors in pushing out the program's estimated completion date. Further affecting the steep rise in cost estimates has been the unparalleled nature of retrofitting such complex bridges. Caltrans indicates that nowhere in the world have bridges as complex been designed or built to today's high seismic standards. This unprecedented retrofit program has created cost uncertainty by limiting the ability to draw from past experiences and to employ traditional estimating practices.
Many factors affected the rising cost estimates to complete the program. It is difficult to attribute dollar effects to specific causes because of this multiplicity of factors and the proprietary nature of the data supporting contractors' bids. Caltrans' standard contract provisions limit its access to the information supporting contractors' bid prices. It may not disclose such information, as it is acknowledged to constitute trade secrets and is not deemed a public record. However, our analysis suggests that various market and project developments have driven the cost increases. Volatile markets for materials and contractor services have yielded bids that include higher than expected steel and contractor overhead costs. Caltrans' efforts to increase competition among contractors by extending the advertisement period and extending the period to build the signature span's superstructure have lengthened the program's timeline, increasing the period during which support services are required and escalating capital costs for projects not yet started. In addition, Caltrans' experience with costs overruns and delays on this program and other recent bridge projects have led it to increase contingency reserve levels to cover the cost of known potential risks and unknown risks for individual projects and the overall program.
Unfortunately, the program's costs could go higher. Caltrans' August 2004 cost estimates assumed it would accept a $1.4 billion offer to construct the signature span's superstructure. However, because Caltrans let this offer expire, the program likely will experience further delays as Caltrans considers rebidding or redesigning this section of the East Span. Bechtel Infrastructure Corporation (Bechtel), a consultant that helped Caltrans develop its August 2004 estimates, agreed with Caltrans that failure to accept the superstructure bid likely would delay program completion and lead to higher costs.
Managing a program of this size, complexity, and cost requires a consistently high level of diligence in applying accepted project management practices. However, Caltrans has not fully incorporated generally accepted standards for project management. This report considers Caltrans' efforts in managing project risk, cost, and communications—areas in which it could have guided the program better to achieve the maximum chance of success. In the area of risk management, the report focuses on the East Span, which accounts for $2.5 billion of the $3.2 billion in cost overruns and the four-year delay. Although Caltrans took steps to identify and mitigate risks to the East Span project, such as hiring consultants to perform a risk assessment in February 2003, it lacked a comprehensive risk management plan for the East Span. Without a risk management plan, Caltrans never defined its risk management activities for the program. As a result, Caltrans lacked processes to identify, track, and quantify risks throughout the project's life.
In managing the project's cost, Caltrans has not followed generally accepted cost management practices to ensure that the project could be completed within its 2001 budget, approved by the Legislature in AB 1171. Caltrans did not regularly update its cost estimates for some components of the East Span or the entire program, including updating estimates for capital and support costs. Also, Caltrans did not use information about identified risks to regularly reassess its contingency reserves for potential claims and unknown risks. For example, Caltrans indicated to the Federal Highway Administration (FHWA) in February 2004 that its program support costs would be $766 million, $30 million less than the AB 1171 estimated amount. However, Caltrans' accounting records show that it already had spent $612 million in support costs by October 2003, leaving only $154 million to pay such costs for eight more years, through 2011. Just six months later, in August 2004, it raised its estimated support costs to $1.352 billion. Without updated cost estimates, Caltrans' program managers forgo the benefits of a detailed overview of the program's capital and support costs for all the bridges. Further, Caltrans indicates that since October 2001, when AB 1171 was passed, its only published program-wide cost update was the August 2004 report, which disclosed the $3.2 billion cost overrun. Had it been monitoring the program's costs regularly, Caltrans would have realized much earlier that the program was exceeding its budget under AB 1171.
Finally, Caltrans has neglected communications planning and management, failing to inform significant stakeholders regularly of relevant changes in its estimates of program costs and cost overruns. State law requires Caltrans to provide periodic status reports to the Legislature, but Caltrans provided no statutorily required annual status report for 2003 and no statutorily required quarterly status report in 2004 until August of that year. It chose not to disclose program information according to the regular reporting schedule established by law and disclosed the large cost overruns long after it should have known that the program likely would exceed its budget. As a consequence, Caltrans placed the Legislature in the awkward position of having to try to devise a funding solution six weeks before the bid on the signature span's superstructure was set to expire.
In November 2003, Caltrans submitted a legally required financial plan update to FHWA showing that the program's projects were going beyond the AB 1171 cost levels and that less than a 3 percent program contingency reserve remained—$122 million—to fund further cost overruns in the eight years left to complete the project. In response to FHWA's questions about the financial plan, Caltrans did not reveal the probable extent of estimated program costs: at the time of the report, Caltrans' internal analyses showed that it would likely exceed the AB 1171 budget. Based on internal Caltrans' reports and the amounts it eventually reported to the Legislature in August 2004, Caltrans should have known about the huge cost overruns. For example, although Caltrans had advertised the contract for the signature span's superstructure at $733 million, internal analyses showed that as early as August 2002 this contract could be as high as $934 million, while later estimates placed its potential price at more than $1 billion. Further, the uncommitted balance of $122 million in the contingency reserve was grossly insufficient given that Caltrans had not received the superstructure bid, the East Span's skyway was only 31 percent constructed, and the Richmond-San Rafael Bridge retrofit costs were underreported by $43 million to $78 million.
In fact, in its report to FHWA, Caltrans claimed it would save $130 million in three areas, thus allowing it to assert that it had a program contingency reserve of $122 million. However, by August 2004, Caltrans reported for those same three areas (unrelated to the capital costs of the superstructure bid) that it would not save $130 million from AB 1171 estimates and that it would need $748 million more than AB 1171 estimates for a total of $878 million more than it reported to FHWA just six months earlier.
To ensure that it properly manages the risks associated with large construction projects, Caltrans should continue to revise its risk management practices, but ensure that its efforts include:
To ensure that it follows generally accepted practices for cost management, Caltrans should:
To ensure that it keeps its stakeholders informed on the status of projects, Caltrans should:
The Legislature should consider revising state law to require that Caltrans submit its quarterly reports within a certain period after each quarter, such as 45 days, to ensure that the information that Caltrans provides is current. The Legislature should also consider changing state law to require that the quarterly reports provide a program-wide summary of the program's budget status for both capital outlay and support costs.
In reviewing the options that Caltrans presents for completing the East Span, the Legislature should consider requesting that Caltrans provide sufficient detail to understand the financial implications of each option. Specifically, this detail should include for each option a breakdown of the costs for capital outlay, support, and contingencies at the project and program level. Further, to place each option in perspective, Caltrans should provide a reconciliation of each option to the figures it presented in its August 2004 report to the Legislature.
Caltrans and the Business, Transportation and Housing Agency provided clarifying comments to the report, and Caltrans indicated the steps it would take to implement the report's recommendations. The Metropolitan Transportation Commission had no comments on the report.