Thar she blows out!
Why do public infrastructure projects arrive late and cost more?
“Price, schedule and quality … choose any two!”
This old saying — sometimes called the Iron Triangle — refers to inescapable trade-offs between desirable project management goals.1
But when it comes to public infrastructure projects, it seems that choices are even more constrained. Projects almost invariably blow out on both price and schedule; and these blowouts are not necessarily offset by increased quality.
Blowouts are a worldwide phenomenon. A 2021 study of public investments by Bent Flyvbjerg and Dirk Bester reported average cost blowouts of 24%, 40% and 96% on road, rail and dam projects respectively.2 New Zealand is not exempt. Recent, stand-out examples include Wellington’s Transmission Gully motorway, Auckland’s City Rail Link and Christchurch’s Stadium.
By contrast, cost and schedule underruns are relatively rare. What might explain this asymmetry? What are the consequences of cost and schedule overruns? And do they matter?
Double then add 1
I worked for Hewlett-Packard as a software engineer in the 1980s. Mark Morwood, my boss, once asked how long it would take me to code a new feature for the company’s database product. He met my confident reply of “2 weeks” with “I’ll write down 5 weeks”, explaining that he always doubled engineer’s estimates then added one, irrespective of the units.
Morwood’s rule — double then add 1 — is an imprecise tool. But it points at a deeper truth. There are (or certainly should be) experienced people working on the development of significant-sized projects. Experienced people know that initial estimates are likely optimistic, and know of similar projects whose budgets blew out. But something must be stopping them from applying this knowledge to project estimates. Either that, or their wiser heads are overruled when estimates are publicised.
But this project was special, with unexpected challenges…
Every non-trivial project faces the unexpected. It snowed! Input prices rose (or fell). Supplies arrive early (or late). Components worked in isolation, but not together. Important things were overlooked. Design standards or regulations changed. And so on.
It will always be possible, after the fact, to find plausible project-specific reasons for a blowout. This post, however, is not about the sources of the unexpected or after-the-fact excuses.3 Rather, I want to explore why pretty much every project blows out over its initial estimates, when those proffering estimates must (or at least should) know the final costs and schedules of similar projects, and that each of those projects also faced unexpected challenges.
Proponents’ incentives lead them to downplay costs & schedules
Project development can takes years. Team members without a personal, professional, philosophical or political commitment to the project are less likely to stick around. This, and people’s tendencies to recruit others who share their views, skews the team composition towards supporters.
Proponents know that early realism may mean that the project never gets off the ground. Far better to start with a low-ball estimate and get further buy-in, then let ever-more realistic estimates dribble out over time. It is difficult for decision-makers to reverse their support, once given.4
Further, proponents know that their project might get pipped by a by dishonestly presented competing project. From their perspective, the “system” rewards dishonesty, encouraging them to act likewise.
The real costs of blowouts
But so what if infrastructure ends up more expensive, or arrives late? We needed it anyway…
But infrastructure is only worth building if the benefits outweigh the costs.5 Often the divergence between infrastructure project estimates and the final figures is so great that governments should, in hindsight, have spent the money on something else. The real cost of blowouts is distorted decision-making, and thus time and money spent on the wrong things.
A further cost is an erosion of trust. If we perceive that our institutions and politicians are systematically misleading us on these projects, then we are less inclined to believe their other statements.
What might reduce or limit blowouts?
These problems run deep, and there are no easy fixes. Transparency of assumptions and independent scrutiny of estimates could help, but aren’t particularly reliable without suitably informed and independent institutions to back them up.
Mechanisms to delay political commitment to projects would assist, as would processes to cancel projects that are blowing out:
Projects announced before governments are prepared to formally commit are also particularly risky. Only one third of projects are announced prematurely, but they account for more than three quarters of the cost overruns. Premature announcements would be no problem if Australia had a robust process for cancelling the duds, but most projects, once announced, are seen through to completion.6
In some cases, funders just need to get smarter about what they sign up to. For example, it is not uncommon in New Zealand for central government to jointly fund projects with the project’s proponent, often local government. Usually this involves both parties agreeing in advance to pay 50% of final costs.
But paying 50% of final costs overlooks an information asymmetry between the parties — the proponent knows more about the likely real costs than does central government, and can understate those costs to encourage agreement. A better strategy for funders is offer something like “we’ll pay exactly 60% of your estimate, and not a dollar more”, encouraging proponents to reveal honest estimates.
By Dave Heatley
The Iron Triangle is also known as the Project Management Triangle.
Flyvbjerg, Bent, & Bester, Dirk W. (2021). The Cost-Benefit Fallacy: Why Cost-Benefit Analysis Is Broken and How to Fix It, Journal of Benefit-Cost Analysis, October, pp. 1-25, doi:10.1017/bca.2021.9.
For a more comprehensive discussion of the sources of cost overruns, see Flyvbjerg, B., Garbuio, M., & Lovallo, D. (2009). Delusion and Deception in Large Infrastructure Projects: Two Models for Explaining and Preventing Executive Disaster. California Management Review, vol. 51, no. 2, Winter 2009, pp. 170-193. doi:10.1225/CMR423.
It’s rare, but not unknown, for cost-escalations to lead to an infrastructure project being cancelled. A recent example was the Auckland Harbour cycle bridge project. First mooted in 2011 at $40 million, estimates grew to $67 million in 2018, then $360 million in 2020. It received the final go-ahead at $785 million in 2021. The Government walked away from the project a few months later.
Flyvbjerg and Bester (2021) also report that the benefits of public investments tend to be systematically over-estimated. The combination of under-estimated costs and over-estimated benefits distorts decision-making even further.
Terrill, M., Emslie, O., & Moran, G. (2020). The rise of megaprojects: counting the costs. Grattan Institute.