Bent Flyvbjerg

Who they are

Flyvbjerg is Emeritus Professor at the Saïd Business School, University of Oxford. He was the first BT Professor of Major Programme Management and a Professorial Fellow of St Anne's College until September 2021, and he chairs Oxford Global Projects. He was academic director of Oxford's MSc in Major Programme Management. His university profile gives his research focus as megaprojects and cities, and a methodology he calls phronetic social science, together with behavioural questions that include the influence of power on rationality, strategic misrepresentation, optimism bias, the planning fallacy and reference class forecasting. It says he has authored or edited ten books and more than 200 papers, and lists a knighthood and two Fulbright scholarships. Three of his books named there are Making Social Science Matter, Real Social Science: Applied Phronesis and Rationality and Power. With Dan Gardner he later wrote How Big Things Get Done (2023).

Key papers

The 2002 paper in the Journal of the American Planning Association, written with Mette Skamris Holm and Søren Buhl, is the foundation. It analysed 258 transport projects worth US$90 billion and found costs underestimated in 86 per cent of them, by 27.6 per cent on average, with no sign of improvement over 70 years. Its title asks whether the cause is error or lie.

A 2003 comment in the Harvard Business Review answered an article by Dan Lovallo and Daniel Kahneman. Flyvbjerg praised their account of overoptimism, anchoring, competitor neglect and the outside view and argued that it underweights forecasts deliberately manipulated to get projects started.

The 2006 paper in the Project Management Journal, "From Nobel Prize to Project Management: Getting Risks Right", proposed reference class forecasting. By its account, the United Kingdom Department for Transport commissioned him and the consultancy Cowi to develop uplifts from reference class data, and from August 2004 local authorities seeking transport funding had to apply them.

Key ideas

Two ideas run through the work. The first is the empirical one: forecasts for large public projects are biased in one direction and the bias does not shrink as experience accumulates. In the 2002 sample, rail averaged 44.7 per cent cost escalation, bridges and tunnels 33.8 per cent and roads 20.4 per cent. The second is explanatory. Flyvbjerg rejects a purely technical account, and rejects optimism as the primary cause, in favour of political and economic explanations in which interests and power shape the number. He argues that optimism calculated on the basis of incentives is not optimism. The 2002 paper hedges this: it says the data support the lying thesis and that proving intent needs interviews or surveys.

His remedies follow from the two ideas. For the bias in how planners imagine their own projects, he proposes starting from the record of comparable cases. For the motive to understate, he proposes accountability: transparency, performance specifications, regulatory regimes and private risk capital, and penalties for consistently foreseeable errors.

Experimental psychologists read the same record differently. Roger Buehler and colleagues showed in 1994 that students with no stake in a thesis deadline still ran well over their own estimates, which suggests that the bias he sets aside exists without an incentive. The two accounts are compatible, and the article on why the planning fallacy persists treats them that way.

---