Fix City Finances 6: What to Do About Big Projects Going "Over Budget and Over Time"
The Economist's 2023 Business Book of the Year explains what every public official needs to understand about big capital projects.
In a year-end interview, Ottawa’s mayor discussed the city’s poor experience with its new light rail transit system. The discussion (starting around 20:35) turned to what went so wrong with that LRT project. The Mayor stated that “every major infrastructure project in the world arrives late and over budget”.
That’s not quite true.
Most do. In fact, 91.5% of big projects go over budget and/or over time.
But what can we learn from the 8.5% of projects that meet their costs and timelines?
A new book, How Big Things Get Done by Oxford professor Bent Flyvberg and former Ottawa Citizen writer Dan Gardner, explains why big projects fail to deliver on their promises at such a high rate. This book should be required reading for all public office holders deciding on any large infrastructure projects.
Rush to a decision
Flyvbjerg and Gardner identify the factors that make projects fail.
Big projects are inherently complex, and difficult to control. But the authors attribute the high failure rate to shallow upfront analysis and poor planning that leads to a project “lock in”.
Starting with cost, many big projects are approved based on flimsy cost estimates. Initial project budgets, it turns out, are often not really about being accurate, but more about selling an initiative.
The first budget should be thought of as a down payment. Proponents understand that once shovels are in the ground, everyone feels locked in. They’re just not prepared to walk away. No one looks good when the conclusion half way a project is that it turned out to be a bad idea all along.
Turning to production, projects fall behind schedule due to a lack of rigorous planning and an optimism bias that downplays potential risks. Humans are overconfident by nature and prone to snap judgements, that leads to uncritical analysis. This leads to a “failure by design” in which project proponents do not properly assess risks and think through the required contingencies.
The type of project matters
Flyvberg has compiled a database of 16,000 large projects. He found that different classes of projects failed at different rates. But he also found that when projects go over budget, there is a risk that they will go way over budget. He determines how often certain project types go over budget by more than 50% — and when they do, how big that overrun is.
Nuclear power plants, for example, experience a 120% cost overrun on average. But 76% of these projects go way over budget — and for those projects, the overrun is, on average, 200%.
At the other extreme is solar power projects. Because solar is highly modular and repeatable, they only go over budget on average by 1%. Furthermore, only 2% of solar projects go way over budget. For those 2%, their overrun is 50% on average.
Flyvberg’s table below shows the different overruns by different types of projects.
How to lower the failure rate
Flyvbjerg and Gardner identify the strategies and approaches that the 8.5% of successful projects get right. They suggest a number of actions to avoid overruns.
Their overarching guidance is to “think slow, act fast”.
That means having a long and deliberate planning process before construction starts, in which everything is scrutinized and tested. It involves giving people the time to experiment in order to understand what works. It involves simulations, iterations and tinkering until you have a reliable plan.
Second, it means using experienced master builders who have done this sort of project before. Otherwise, your construction lead with be finding out what does and doesn’t work. Great learning for the next project, but not good for your project.
Third, it means using proven technologies and tested products. Again, your project should not be the guinea pig.
Fourth, it means using modular design as much as possible, so that constructors can learn once and build over and over again quickly. For a rail system, this might mean constructing stations based on a common design. Solar power projects are reliable on cost and delivery because solar is a highly modular and repeatable technology.
Fifth, it means recognizing that when it comes to budgets, your project is not as unique as you think it is. A more reliable method for costing is to “take the outside view”, i.e., look at the cost of similar projects and use that as a basis for projecting your costs. Any initial bottom up costing exercise will have great difficulty modelling risk, and accordingly, unable to accurately project costs.
Not easy … but not inevitable
Big infrastructure projects are unimaginably complex. It is not surprising that they are prone to delays and cost overruns. But it is not inevitable that they will always come in late and over budget.
Flyvberg and Gardiner point to the factors that determine success or failure. Every public official involved in infrastructure decision making should be asking if the projects being put forward for approval include the success factors consistent with on time and on budget delivery.