This chart—stuck in the middle of a serious academic report about why big public works projects go over budget—made me laugh out loud…
It’s from this article by the ever-fascinating Danish academic Bent Flyvbjerg (for more, see here), who’s made a career out of figuring out why engineers and governments do such a bad job of estimating the costs of megaprojects.
The graph, obviously, is meant to be more illustrative than scientific. But it does a pretty good job of explaining Flyvbjerg’s central thesis: through a combination of over-optimism (delusion) and strategic misrepresentation (deception), planners and boosters tend to low-ball the costs of a big project; and the higher the political stakes, the more likely it is that deception is involved.
The problem Flyvbjerg identifies is hardly limited to public projects. This short blog post has already taken twice as long as I hoped it would—call me delusional. And as loath as I am to admit it, I’m probably guilty of some strategic misrepresentation. “I’m leaving right now, honey” often means that I hope to leave sometime in the next 10 minutes.
But scaled up to the size of a massive public project, those minor pecadillos balloon into crushing cost overruns. Boston’s Big Dig comes to mind. Or consider the Iraq war: a generous mixture of delusion and deception led the nation into an undertaking that’s been dozens of times more costly, in both money and lives, than the planners and boosters initially thought.
At some level, of course, I wonder how much outright deception there really is in mega-projects. Few people think of themselves as liars; but it’s easy to see people convincing themselves that wildly optimistic cost estimates are actually quite reasonable. Subjectively, “strategic misrepresentation” can feel a lot like honesty. (For a similar phenomenon, see last week’s Newsweek article on why people believe lies, even when shown a clear explanation of the truth.)
Flyvbjerg’s antidote to both delusion and deception in public megaprojects is called “reference class forecasting.” Rather than making cost forecasts based on how planners hope things will go, Flyvbjerg recommends looking at a group similar projects—a “reference class”–that have already been completed. The real-world experience of these projects, Flyvbjerg believes, gives a more reliable estimate of costs than the rosy estimates of project planners.
Has anyone tried the reference class approach for any of the big megaprojects in our own backyard? Not to my knowledge—which is an oversight that could spell trouble for the region’s taxpayers down the road.
joshuadf
For what it’s worth, civil engineers already use the “reference class” approach to estimating cost of projects. However, in any large project there is always of a lot of uncertainty about scope, cost of labor and materials, what incentives will come through, market conditions, etc.
Clark Williams-Derry
As I read it, Flyvbjerg recommends using reference class forecasting as (among other things) a way of “uplifting” cost estimates derived from standard time-plus-materials-plus-financing estimates. That is, select a set of similar projects; look at the initial cost estimates that were derived from *internal* factors for those projects; then look at the final cost of those projects; and use the ratio of actual-to-estimated costs to adjust the cost estimates upwards.But perhaps I’m reading his methods wrong.
Cary
Regarding the looming tunnel megaproject, here is a starting list of the delusions that have emerged so far.1. State legislators announced there will be no cost overruns, but if there are, individual property owners will be handed a bill by state government. As it is written, citizens, not the government, are expected to pay out of their own pockets (even though this vacuum of accountability seems unconscionable.) 2. The promise of It’s all underground! We won’t even know it’s happening! is already slipping. For instance, bad soils in Pioneer Square likely mean that this tunnel will be a cut and cover tunnel south of King—and 1st Avenue is going to be out of commission for quite a while. The ramps around the tunnel mouths are turning out to be pretty bleak, especially in Pioneer Square and the stadium area.3. The tunnel is officially a preliminary preferred alternative at this point, as the EIS hasn’t been completed yet. No other alternatives are being studied in the EIS, since politically the decision has already been made in advance of the data—exactly the nightmare Flyvbjerg warns against. If it turns out the tunnel is too costly, or serious constructability problems emerge, or energy (and steel, and concrete…) costs spike again, or there is too much negative environmental impact, there are no other options on the table. 4. The tunnel itself serves bypass trips avoiding downtown. If this project is expected to replace the mobility function of the current viaduct, it needs to include lots of transit and street improvements to provide local access. These projects were included in the January 2009 deal announced by the Governor, and have now disappeared from the plan. The promise to help guide stimulus money to Spokane St and Mercer St projects: didn’t happen. The promise to authorize local governments to raise MVET for transit: vetoed. The inclusion of new transit in the transportation modeling within the EIS: absent. Without new transit, we could likely see worse congestion downtown and more pressure for a bigger, faster arterial street on the waterfront, as all those trips not served by the tunnel have little choice but to drive.And there are likely more to come—when the EIS analysis and cost estimates are revealed later this fall.
Wells
Cary, you are awesome! =^)Here are a few particulars about the AWV project that are apparently more intentional deception than optimistic delusion:For over 5 years, WsDOT and SDOT proposed an alignment of the Waterfront Streecar line through the Wide Plaza. Only recently was the idea admitted unfeasible, (unsafe, operationally difficult). Were these agency heads ‘pandering’ to the streetcar popularity, knowing the alignment would be ultimately rejected? Several feasible alignments are possible. Why is a Waterfront Streetcar line off the table? Right after the Nisqually quake, WsDOT proposed the most expensive tunnels, 2x and 3x as expensive as the tunnels proposed after the 2007 voter rejection. Did WsDOT begin with the most expensive tunnels knowing their cost was prohibitive and they’d be rejected? The Deep-bore does not provide access to Interbay-bound traffic at Western/Elliott, about 40,000 vehicles daily or 2500 per hour, give or take, depending upon time of day. Even half this much additional traffic there will produce bumper-to-bumper gridlock all day long. Why do WsDOT’s latest video simulation show a 4-lane Alaskan Way when even a 6-lane version is not enough to prevent gridlock? Based on these questions, it’s hard to not conclude that WsDOT and SDOT have intentionally ‘deceived’ the public. Seattle’s next mayor should not be the candidate who goes along with questionable designs from untrustworthy DOT chieftains.
AnotherHomeOwner
This is actually an easy one:1) Uncertainty.You can do as many tests/studies as you want, but you will never find out exactly what the costs are until they are incurred.2) Politics.You can be as clean with your books as a whistle, but there will (almost) always be somebody in the office saying ‘cant we give these guys a better deal? I know we will lose money, but they need a better deal if we want them to sign our contract. We can milk them later’3) Unexpected ChangeMaybe that good worker that you rely on will have a death in the family and be unable to be there…or maybe that software you have relied on for so long is upgraded to become a piece of useless junk…or maybe the wood you bought for the project will be at an unexpectedly high price because of forest fires.In short, all things being equal, all things really arent equal after all.