Rendering of the Transbay Termainal
Current shot of the terminal site.
It was announced last week that the Transbay Transit Center at First and Mission will cost $300 million (or 19%) more than the initially anticipated $1.9 billion price tag—which is awful, but it could be worse. As Willie Brown wrote (or we should say, dictated) in his Chronicle column this weekend, "We always knew the initial estimate was way under the real cost. Just like we never had a real cost for the Central Subway or the Bay Bridge or any other massive construction project. [...] If people knew the real cost from the start, nothing would ever be approved."
Is Willie right? Are public works projects inevitably destined to balloon in costs?
We've all heard horror stories like Boston's Big Dig, which came in at more than twice its initial projected cost. And when BART's transbay tube was completed in 1969, its cost had ballooned to $180 million—$47 million over budget. But just how frequent—and how extreme—are public-works project overruns?
Turns out that they're about as common as dirt and our dirt pile isn't even that high. According to data from Oxford professor Bent Flybjerg's 2007 paper, "Cost Overruns and Demand Shortfalls in Urban Rail and Other Infrastructure," which analyzed 258 major transportation infrastructure projects from around the world, the mean overrun was 27.6% (inflation-adjusted). What's more the mean overage for transportation projects was a whopping 44.7%. If the Transbay Transit Center came in that high, it wouldn't be $300 million over budget—it would be $713 million over.
The Eastern Span of the Bay Bridge is a prime example of how bad things can get. In 2006, authorities signed off on a budget of $1.43 billion. As of today, the estimated total cost will be more than four times higher: $6.3 billion. Even adjusting for inflation, that's an increase of 380%. But it's not just the marquee projects that have problems either. In April, Muni had to pay $100 million more than anticipated for the ongoing construction of the Central Subway. A 2011 audit found a $90 million cost overrun across 29 different project, up from a projection of $800 million.
Why are government projects particularly prone to these kinds of overages? Back in 2009, Chris Edwards of the Cato Institute, a libertarian think tank, argued that "people tend not to spend other people's money as carefully as they spend their own. [...] One manifestation is the pattern of cost overruns on [...] large government projects." He goes on to give a greatest hits list of such overages, including the Virginia highway interchange budgeted at $241 million that cost $676 million, and the Denver International Airport—slated at $1.7 billion—that cost $4.8 billion. (Edwards mostly gives nominal, rather than inflation adjusted numbers, which exaggerates the magnitude if not the direction of the trend.) To the libertarian point of view, overruns like these can be blamed on "bungling by administrators, [...] political buck passing [and] a strong incentive for all the supporters of spending projects [...] to low-ball initial estimates of costs in order to receive legislative approval."
One possible solution they offer: dramatically shrink governmental spending of all kinds. After all, private firms can be expected to use the money more efficiently, right? Not so fast. The Boeing 787 ended up coming in at least 120% over budget. The cost of Apple's planned headquarters has grown from $3 billion to $5 billion in just two years. And within days of each other, Xcel Energy in Minnesota announced the cost to upgrade its nuclear plant would double and Georgia Power said its two new nuclear plants would have cost overages of $737 million.
So what gives? Why do we—all of us—fail so badly at estimating costs? Part of the blame has to be chalked up to what Flybjerg calls "strategic misrepresentation." That's the kind of overly rosy salesmanship that Willie Brown wrote about. Another part of the blame has to be given to what Cato calls bungling. But, regardless of these two factors, a large share of the blame has to be given to our brains. Humans happen to be awful at thinking in large numbers and over long spans of time. It's a macroeconomic version of why I always leave the grocery store having paid more than I thought I would.
But don't trust me. Take a look a the work of economic psychologist and Nobel Prize winner Daniel Kahneman, who explains how "optimism bias" makes us "tend to exaggerate our ability to forecast the future, [...] foster[ing] overconfidence." Human beings are bad at doing this kind of thinking—and we are bad in predictable ways. After all, when was the last time you ever heard of a big-ticket project coming in under budget?
The two big takeaways from the Transbay overage news? Well, first, it's hard to trust anybody's numbers until the project is actually completed. And second, it's hard to get too surprised when these things come in way over budget.