Bent Flyvbjerg: Why Megaprojects Fail Despite a Rational Plan

Megaprojects fail despite a rational plan because the plan is never only rational: a drawing arrives at the meeting with figures, deadlines and arrows and looks neutral, and then power enters. Bent Flyvbjerg, a professor at the University of Oxford and the IT University of Copenhagen, has devoted his career to that question, from Rationality and Power (1998), his study of planning in the Danish city of Aalborg, to Megaprojects and Risk (2003) and How Big Things Get Done (2023). His thesis is that decisions presented as technical are shot through with interests, conflicts and actors of unequal capacity, and that the iron law of megaprojects, over budget, over time, under benefits, over and over again, is explained by that mix of power and deception more than by errors of calculation.

Aalborg: how power dismantled a rational plan

Aalborg was the first case. Flyvbjerg followed for years a city-centre plan designed to reduce traffic and improve public space, technically sound and adopted by consensus, and watched the chamber of commerce, the police and car interests dismantle it piece by piece in informal negotiations until what remained produced the opposite of what was intended. His conclusion, formulated as propositions on the relation between rationality and power, is that power defines what counts as knowledge, that rationality is the strategy of the weak and that in open conflict the more powerful wins, not the better argument. Planning is technical and political practice at once.

Phronetic social science: values, power and consequences

His approach, phronetic social science, takes from Aristotle the notion of phronesis, practical wisdom, and studies concrete decisions by asking what is at stake, who has capacity to influence, which arguments gain authority and who wins and loses from the outcome. Against a social science that imitates physics in search of general laws, Flyvbjerg defends the case study and the question of values and power, and he has applied that method to the world's largest projects: bridges, tunnels, dams, airports, railways and Olympic Games.

The iron law: optimism bias and strategic misrepresentation

The data confirm the iron law. His database of more than sixteen thousand projects in 136 countries shows that nine out of ten exceed their budget, that average overruns run from 20 percent for roads to more than 90 percent for dams and railways, and that the Sydney Opera House cost fourteen times its estimate. Flyvbjerg identifies two causes: optimism bias, the human tendency to underestimate costs and time, and strategic misrepresentation, the deliberate lying of promoters and politicians who present low estimates to get a project approved, knowing that once started nobody will stop it. Berlin's airport, nine years late, and Boston's Big Dig are the examples.

Remedies: reference class, accountability, think slow and modularise

His remedies are concrete. Reference class forecasting, which estimates costs and time from what happened in similar projects rather than from what the promoter promises, and which the United Kingdom and Denmark have built into their appraisals. Accountability: whoever estimates must answer for the estimate. And, in How Big Things Get Done, think slow and act fast, planning thoroughly with models and prototypes before building, and modularity, breaking projects into repeatable pieces as solar farms or prefabricated pavilions do. Critics say the method may discourage necessary projects; he replies that what it discourages is lying about them.

Flyvbjerg's lesson is that a megaproject fails not when the calculations go wrong but when power decides before the calculation and uses technical language to hide it. Asking who defines the problem, who gains from the answer and which risks were minimised is as necessary as checking the figures, and planning that does so produces fewer cathedrals in the desert and more works that deliver what they promised. Rationality, in his reading, is not enough to govern; but without it power governs alone.