In his office this afternoon, City Council member Mike O'Brien laid out a pretty compelling case (conspiracy theory?) that the state is focusing on getting the lowest bid, when they should be focusing on getting the lowest cost. If that low bid turns out to be too low, the city, O'Brien argues, will bear the risk.

It's gonna get wonky, so stick with me.

Every contractor on a major project like the tunnel is required to get a performance bond—essentially, an assurance to the client (in this case, the state) that they'll get paid up to the amount of the bond if the contractor fails to deliver the project (e.g., if the contractor goes bankrupt).

Until recently, the state required performance bonds to be equivalent to the entire cost of the project. In the case of the tunnel, that cost is $1.2 billion—the cost of digging the hole plus the cost of building the tunnel itself.

But in 2009, state legislators passed a law—sponsored, among others, by then-Sen. Fred Jarrett (D-41) and Sen. Mary Margaret Haugen (D-10)—saying that for projects costing more than $250 million (a category that includes only the tunnel, 520 replacement, the I-5 Columbia River crossing, and a portion of I-90 near the Snoqualmie Pass), the bond can be smaller than the total cost of the project.

Although the Surety Association of America determined after the new law passed that the market could support a bond of up to $800 million, the state ultimately agreed on a much smaller bond—$467 million.

How did that happen? The legislation says the bond must be big enough to protect the state from "one hundred percent of the state's exposure to loss," and directs the state transportation department to come up with guidelines for determining how large the state's maximum loss would be—the "worst-case scenario." Engineering firm Parsons Brinkerhoff—which also oversaw design and construction Boston's Big Dig, settling for $407 million in the fatal 2006 collapse of the tunnel—applied the guidelines and found that a bond of $467 million would be big enough to protect the state from any possible loss.

O'Brien says he thinks the $467 million number was arrived at not as a real assessment of risk, but as a way of getting the bond under $500 million. "I think that number was reverse-engineered," O'Brien said in his office today."To me, what it looks like—and I don't want to be a conspiracy theorist here—but I think they decided they wanted a $500 million bond and they said, 'What's the study we need to do to get to a $500 million bond?'"

O'Brien points to several oddities about the "worst-case scenario" in Parsons Brinkerhoff's report. According to the report itself, "The absolute worst case scenario assumed for this assessment is the complete failure of the tunnel boring machine during the construction of the tunnel," in which case "the [machine] will fail and be rendered useless, resulting in the project contractor abandoning the project and necessitating a new contractor to be found."

At the Brightwater sewage treatment plant, where two tunnel boring machines got stuck mid-tunnel last year, King County estimates it will cost more than $100 million to get one of the machines moving again and to resume digging the second tunnel with a new machine (the first one has been abandoned). However, the Brightwater tunnels are only 18 feet in diameter, compared to 54 feet for the downtown tunnel. According to PB's estimate, getting a new contractor and digging in to that tunnel will cost between $60 million and $160 million, depending on how many feet remain to be dug. Excavating an entirely new hole on the waterfront or First Avenue and digging out and replacing the machine, PB estimates, will cost between $158 million and $227 million.

O'Brien is skeptical of those numbers, noting that if the boring machine failed under a building on First Ave., for example, it would be astronomically expensive to dig a new hole, pull it out, and replace it. (O'Brien says the machine would have to be pulled out of the tunnel from above because the tunnel walls will be built behind it, making it impossible to pull the machine out the way it came in).

He also questions some of the firm's assumptions about the potential cost of various risks to the project. For example, the risk assessment notes that the state usually estimates the cost of mobilizing a new project at around 6 percent. However, the assessment estimates that demobilizing the project if the contractor defaults will only cost 0.5 percent of the project cost. "How did they come up with that?" O'Brien says. Similarly, getting the project back up and running again can cost as much as 10 percent of the remaining contract cost. However, the state estimates the cost of remobilization at just 2.5 percent. "It seems to me that if you're looking at the worst-case scenario, you ought to use the worst-case scenario," O'Brien says.

The county's contractor did not have to procure new machines to replace the ones that stalled in the tunnel.

"I don't think that anyone can argue that this project is not, relative to other projects, very risky."