# Flyvbjerg and Gardner's How Big Things Get Done explained: why megaprojects fail

> How Big Things Get Done (2023), by the megaproject economist Bent Flyvbjerg and the journalist Dan Gardner, draws on a database of more than 16,000 projects to show that only 8.5% finish on budget and on time, and just 0.5% also deliver the benefits promised. The book's remedy is to plan slowly and in detail, forecast from the real record of similar projects, and then build fast with modular, repeatable units and an experienced team. For radical centrism it gives the delivery discipline that an ambitious programme of housing, rail and grids needs, and Madrid's metro expansion of 1995–2003 is one of its best examples.

- URL: https://www.radicalcentre.eu/en/articles/flyvbjerg-gardner-how-big-things-get-done-explained
- Language: en
- Publisher: radical centre / centro radical
- Published: 2026-11-11
- Topic: Democracy & institutions
- Author: radical centre
- Licence: CC BY 4.0

## Key points

- Flyvbjerg's database covers more than 16,000 projects in 136 countries: 8.5% come in on budget and on time, and 0.5% on budget, on time and with the promised benefits.
- Overruns have fat tails: the Scottish Parliament ran 978% over budget and the Sydney Opera House 1,400%.
- The method: think slow, act fast; reference-class forecasting from similar past projects; modular 'Lego' designs; an experienced team.
- Madrid's metro built 56 km and 37 stations in 1995–99 and 75 km and 39 stations in 1999–2003, at about half the cost and twice the speed of industry averages.
- The UK has used reference-class forecasting in transport appraisal since 2004; Spain should build its own reference classes for rail, grids and housing.

Bent Flyvbjerg has spent his career collecting the results of big projects: bridges, dams, IT systems, Olympic Games, rail lines, office towers. His database now holds more than 16,000 of them from 136 countries, going back to 1910. In How Big Things Get Done, written with the journalist Dan Gardner and published by Crown Currency in February 2023, he turns that record into a book anyone can read. The numbers are brutal. About 48% of projects come in on budget. Only 8.5% come in on budget and on time. And 0.5%, one in two hundred, come in on budget, on time and with the benefits that were promised.

Flyvbjerg calls this the iron law of megaprojects. The typical project overruns a little; the dangerous ones overrun catastrophically, because the distribution has a fat tail. The Big Dig in Boston went 228% over budget, the Montreal Olympics 720%, the Scottish Parliament building 978% and the Sydney Opera House 1,400%, taking almost three times as long as planned. Europe has its own recent entry. Berlin Brandenburg airport opened on 31 October 2020, nine years late after six abandoned opening dates and problems with its wiring and fire safety system, at a cost of about €6 billion, roughly three times the original plan.

## Think slow, act fast

The book's central advice is a reversal of how politics usually works. Announcements come first, the ribbon is promised for before the next election and the detailed design is done on site. Flyvbjerg asks for the opposite: spend a long time planning, when changes are cheap, and then build as fast as possible, because the longer a project runs the more time the world has to throw floods, price spikes, lawsuits and changes of government at it.

Planning, for him, means more than a schedule. The authors tell readers to think from right to left, starting with the goal and working back to the means. They admire Pixar, which storyboards and re-storyboards a film many times before animating it, and Frank Gehry, who tests his buildings in detailed digital simulations before anyone pours concrete. Simulate and iterate while mistakes cost nothing.

The second tool is reference-class forecasting. Promoters estimate costs from the inside: their plan, their assumptions, their confidence. Flyvbjerg's data shows that inside views are systematically optimistic. The fix is to take the outside view: find a class of similar finished projects, look at how much they actually overran, and adjust the estimate accordingly. It sounds modest and it works. The UK introduced reference-class forecasting for transport projects in 2004, and the Department for Transport's 2020 update rests on evidence from 2,522 rail, road, bridge and tunnel projects. Denmark's transport ministry adds correction reserves of 50% at the early stage and 30% at the decision stage, part of them held centrally so that projects cannot quietly spend them.

The book also takes on a famous idea from Albert Hirschman, the Hiding Hand: the claim that underestimating difficulties is useful, because it gets projects started and people then rise to the challenge. Hirschman built his principle on eleven projects. In an earlier paper with Cass Sunstein, Flyvbjerg tested it against his much larger data and found the opposite far more common. Underestimation usually leads to overruns and disappointment, and the benevolent Hiding Hand is the exception.

## What's your Lego?

The third idea is modularity. Some project types almost never blow up: solar farms, wind farms, power lines, roads. The Independent Review's summary puts it plainly, solar and wind farms are highly modular, low-risk projects. They are built from a small unit repeated many times, and every repetition makes the next one faster and cheaper. Nuclear plants, big IT systems and Olympic Games sit at the other end, one-off projects that get no such learning. Flyvbjerg asks of every project: what is your Lego, the small piece you can build again and again?

His favourite example is Spanish. Between 1995 and 2003 Metro de Madrid, led by the civil engineer Manuel Melis, built 56 km of line and 37 stations in a first four-year stage and 75 km and 39 stations in a second, at about half the cost and twice the speed of industry averages, according to Flyvbjerg's account in the Harvard Business Review. There was no signature architecture. Every station followed one standard design, built with proven cut-and-cover methods, and the project used no new techniques, designs or train cars. Tunnels were divided into stretches one boring machine and its team could finish, typically 3 to 6 km in 200 to 400 days, with up to six machines running at once. The teams competed with each other and shared what they learned. Residents were asked whether they preferred three years of works or eight, and accepted round-the-clock tunnelling. Fast was cheap.

The last ingredient is the team. Heathrow's Terminal 5 is the book's case of a client that turned its contractors into one real team. Madrid makes the same point from the other side: its tunnel teams got better with every stretch they finished.

## Why it matters for the radical centre

The left brings a real ambition here. Metros, grids, rail and housing are shared goods that ordinary people need, and the case for building far more of them in Spain and across Europe is strong. The market-liberal tradition brings the scepticism the data supports: promoters and politicians underestimate costs, money gets wasted, and every big project needs hard cost discipline, honest forecasts and competition between the people delivering it. Flyvbjerg's method fits neither camp, and it lets the two work together. You get to build big because you plan slowly, forecast from the record, design in modules and build fast.

That is why the book sits next to Jennifer Pahlka's [Recoding America](/en/articles/jennifer-pahlka-recoding-america-explained) in this programme. Democracia 21 put an agile state that delivers at the heart of its offer (see [our overview](/en/articles/democracia-21-miriam-gonzalez-proposals-radical-centre)), and [grids are the obvious test](/en/articles/grids-permits-why-renewables-wait-spain). For Spain a concrete step follows. The Ministry of Transport, the Treasury's evaluators and the regions could build shared reference classes for rail, grids, hospitals and social housing, publish the cost and schedule record of every finished project, and require every new one above a threshold to be estimated from that record. Madrid already showed the country how to build quickly and cheaply. Writing the method down and using it everywhere is the next job.

## FAQ

### What is How Big Things Get Done about?

It is a 2023 book by Bent Flyvbjerg and Dan Gardner on why big projects, from home renovations to Olympic Games, overrun so often and how the few successes do it. It rests on Flyvbjerg's database of more than 16,000 projects and argues for slow, detailed planning followed by fast, modular delivery.

### What is the iron law of megaprojects?

Flyvbjerg's name for the pattern in his data: over budget, over time, under benefits, over and over again. About 48% of projects come in on budget, 8.5% on budget and on time, and only 0.5% on budget, on time and with the expected benefits.

### What is reference-class forecasting?

A way of estimating a project's cost and schedule from the actual results of a group of similar finished projects, the outside view, which corrects the plan's own assumptions. It corrects the optimism bias of promoters. The UK introduced it for transport projects in 2004.

### Why is the Madrid Metro a model in Flyvbjerg's book?

Under Manuel Melis, Metro de Madrid built more than 130 km of line between 1995 and 2003 with one standard station design, proven cut-and-cover and tunnel-boring methods and up to six boring machines working at once, at about half the cost and twice the speed of industry averages, according to Flyvbjerg.

## Sources

- [How Big Things Get Done, by Bent Flyvbjerg and Dan Gardner](https://www.penguinrandomhouse.com/books/672118/how-big-things-get-done-by-bent-flyvbjerg-and-dan-gardner/), Penguin Random House (Crown Currency) (2023)
- [How Big Things Get Done: 5 key insights from the authors](https://nextbigideaclub.com/magazine/big-things-get-done-surprising-factors-determine-fate-every-project-home-renovations-space-exploration-everything-bookbite/40952/), Next Big Idea Club (2023)
- [How Big Things Get Done (review)](https://www.independent.org/tir/2023-fall/how-big-things-get-done/), The Independent Review (2023)
- [Holding down costs of megaprojects: the Madrid subway example](https://conversableeconomist.com/2021/11/24/holding-down-costs-of-megaprojects-the-madrid-subway-example/), Conversable Economist (2021)
- [Updating the evidence behind the optimism bias uplifts for transport appraisals](https://assets.publishing.service.gov.uk/media/6093d448e90e0726f52fc54c/updating-the-evidence-behind-the-optimism-bias-uplifts-for-transport-appraisals.pdf), UK Department for Transport (2020)
- [New Berlin airport opening 9 years late with little ceremony](https://www.news4jax.com/news/world/2020/10/31/new-berlin-airport-opening-9-years-late-with-little-ceremony/), Associated Press (2020)
