Dan Gardner is the New York Times best-selling author of Risk, Future Babble, and Superforecasting (with Philip E. Tetlock). His books have been published in 25 countries and 19 languages. Prior to becoming an author, Gardner was an award-winning investigative journalist. More >

Infrastructure: Think Slow, Act Fast

Originally published in The Boston Globe, April 1, 2021

Bent Flyvbjerg and Dan Gardner

As Biden administration officials prepare the big push for the president's $2 trillion infrastructure program, they are focused on politics. As they must be. It is politics President Joe Biden must overcome if there is to be a program at all.

But officials must look beyond politics, to planning and delivery — or the White House could succeed in getting its bill passed in Congress only to deliver a fiasco similar to California's high-speed rail project.

In 2008, California voters approved rail service from Los Angeles to San Francisco that would take only 150 minutes. The service could be built within 20 years at a cost of $40 billion. But from the start, construction ran into trouble. The timeline was repeatedly pushed back. Money hemorrhaged. The estimated total cost of the project ballooned to a staggering $100 billion.

In 2019, Governor Gavin Newsom bowed to the inevitable and killed the dream. Sort of. Rather than simply write off the enormous sunk costs, Newsom declared the track would still be completed between the small inland cities of Merced and Bakersfield, leaving it to future governors to figure out how to finish the line someday. This 170-mile bullet train to nowhere would cost an additional $20 billion. Now, even that radically scaled-back plan is over budget and behind schedule.

If the California high-speed rail project were an aberration, it wouldn't be of interest to anyone except California taxpayers. But it's no aberration. It is an extreme example of a general phenomenon.

Major projects of the sort envisioned by Biden fail with depressing regularity. A team at Oxford University has developed the world's largest database of such investments, drawing on thousands of projects in 22 fields. It shows that major projects routinely finish late, cost more than expected, and produce less benefit than anticipated.

It's true of railroads, highways, bridges, tunnels, and airports. It's true of buildings, dams, mines, oil and gas, aerospace, defense, nuclear power, and information technology. It's true in the United States and around the world. Globally, only 0.5 percent of projects deliver as promised.

But it's the magnitude and distribution of the failures that should most alarm administration officials.

Consider one category that is particularly important because it is embedded in most big projects today — IT. The average cost overrun on an IT project is 27 percent. That's bad enough. But it masks the full, scary reality: When IT projects are graphed, for example, the classic “bell curve" does not emerge from the data points. Instead, a “fat tail" appears. Fully 1 in 6 of the projects is a black swan, with an average cost overrun of 200 percent. And that's only the average. Overruns of 400 percent and more are a realistic possibility. These projects don't just fail. They melt down. And IT is in everything.

The causes of this dreadful record are complex, but the core of the story is rushed planning. It leads to problems not corrected and risks not mitigated. And it produces terrible forecasting that puts projects on a path to failure almost from the start. Rushed planning is how dreams become fiascos.

So what causes rushed planning? A major driver is politics. The desire to rally support and lock in projects so they can't be undone later often convinces politicians to skip elementary planning and get “shovels in the ground" as quickly as possible. That all but guarantees a proliferation of problems in the future. The price of a fast political win is too often a long, slow, budget-draining mess.

For projects to succeed, those responsible must think slow, act fast.

Rigorous planning — built around proven, scalable, modular technologies — accurately forecasts costs and timelines, spots problems, and mitigates risks before the first shovel is picked up. Delivery is then smooth and quick. There are great models available, from Tesla's Gigafactories to Ørsted's immense offshore wind farms and the enormous transportation projects in Madrid that were carried out with unprecedented speed and efficiency.

There's no question this approach can make the politics harder. Yet satisfying the demands of both politics and planning is the only way to truly succeed.

The United States needs the sort of ambitious investments that built the transcontinental railways of the 19th century and the interstate highways of the 20th century. But if the Biden administration is to “build back better" and fight climate change in the 21st century — while avoiding the fate of the bullet train to nowhere — it cannot let politics dominate the process.

Bent Flyvbjerg is the BT Professor of Major Programme Management at the University of Oxford, coauthor of “Megaprojects and Risk," and editor of the “Oxford Handbook of Megaproject Management."

Dan Gardner is the coauthor of “Superforecasting" and author of “The Science of Fear" and “Future Babble."