Reference other projects. Frick points to a couple ideas for controlling mega-project costs. Scholar Bent Flyvbjerg, who has studied infrastructure cost overruns around the world—and who often boils them down to political deception—has promoted the idea of basing costs on a “reference class” of similar projects already completed. The fear with that is project leaders won’t bother to keep costs down if they know they can hit a certain number, but Frick says that possibility bothers her less than the uncertainty surrounding costs that goes on right now.
Widen early cost ranges. Giving a precise cost number out to multiple decimals, as the state legislature did with its $1.285 billion estimate in 1997, makes the figure seem more scientific and precise than it really is, and creates that much more public frustration when the costs keep rising in the future. “In the early planning stages, ranges in the projects would be really important to provide,” she says.
Track progress more closely. Frick also suggests that officials pay more attention to “transaction cost economics”—an approach that “analyzes project development over time,” she writes, in an effort to identify the precise “political and economic origins” of new costs. This fuller accounting also considers costs that often go overlooked, such as the time and energy that go into public participation. Without better cost estimates, projects will continue to suffer from the type of strategy described to Frick by one senior engineer:
“Basically at the onset of a project I think the higher ups prefer a dollar amount and schedule that doesn’t shock the public.”
Which, as the Bay Area knows, only makes the shock that much worse when it finally arrives.
The typical resident is going to look at this and ask how in the world this was allowed to happen. Large infrastructure projects have a lot of moving pieces but the change in price is still hard to understand. Of course, there may be a political penalty for adhering to this advice – a higher projected cost upfront is likely to limit support. Yet, going with an unreasonably low projection with no cost range borders on dishonesty.