Megaprojects and Risk: An Anatomy of Ambition (2003) by Bent Flyvbjerg is a dry but interesting look at how expensive projects are chronically over budget, under utilised and delivered late.
The book gives the strong impression that the big budget government and private projects like the Channel Tunnel, the Øresund Bridge and many rail projects are chronically misrepresented in financial modelling that precedes them.
Flyvberg and his team have analysed many big public infrastructure projects and have found that 9 out 10 cost more than planned, 9 out of 10 are late and 9 out of 10 are not utilised as much as projected. Roads projects, are, however, quite often used more than expected It’s a terrible record. The record of rail projects is the worst, Average traffic on rail projects was 61% of the forecast level and sometimes as low as 25%.
After looking at these projects the authors found that those involved in making plans had lots of pressure put on them to cook forecasts. Essentially the promoters of the projects are dishonest and are either after political gain or are set to build the project and personally benefit with the risk generally taken up by taxpayers.
The book suggests a framework for better assessment with a range of forecasts made and investigation of how close the forecasts are to what happens after the fact. They also suggest that private money be solicited and made to take some of the risk. If it can’t be obtained it is a good suggestion that the project was not really viable.
It could be argued that the projects that are not viable but that are huge achievements should be undertaken anyway. This is a legitimate point. However, deception in the initial phases should still be reduced as much as possible. Part of the problem is described by the authors as being that projects are sometimes chosen on the basis of who has the most dishonest figures.
It’s a really interesting book that anyone who believes in infrastructure deficits and that more public spending in infrastructure is required should definitely read.