Cognitive Biases · CB-21
People reliably underestimate how long a task will take and how much it will cost — even when they know, from direct personal experience, that past similar projects ran over.
The tendency to underestimate the time, costs, and risks of future actions while overestimating the benefits, even when a person has extensive direct experience of similar past projects running over budget and behind schedule — the bias persists despite explicit awareness of one's own track record.
Introduced by Daniel Kahneman and Amos Tversky in 1979, with the underlying pattern extensively documented across large infrastructure and IT projects, notably in economist Bent Flyvbjerg's research finding large public infrastructure projects overrun their budget by an average of tens of percent, with some categories running far higher.
The Mechanism
Best-case, most-likely, and actual outcomes reliably diverge
Even the 'most likely' estimate — not just the best case — is reliably beaten by the actual outcome, which is what makes the planning fallacy distinct from simple optimism about a single best-case scenario: people systematically underestimate their realistic, most-probable timeline too.
01 · THE INSIDE VIEW SYSTEMATICALLY UNDERESTIMATES; THE OUTSIDE VIEW DOESN'T
Kahneman's proposed fix targets the actual mechanism
People planning a project naturally focus on the specific details of this project (the 'inside view'), which feels unique and tends to omit unknown obstacles — while an 'outside view,' looking at how similar past projects of this general type actually turned out, produces far more accurate estimates by anchoring on real base rates rather than optimistic specifics.
02 · IT PERSISTS DESPITE DIRECT PERSONAL EXPERIENCE OF PAST OVERRUNS
Knowing your own track record doesn't automatically fix the bias
Even project managers with years of direct experience seeing their own past estimates blown through continue to underestimate future project timelines — the bias isn't primarily a lack of relevant data, it's a failure to actually apply that data (the outside view) rather than reasoning from the specifics of the current case alone.
03 · IT'S DOCUMENTED AT MASSIVE SCALE IN PUBLIC INFRASTRUCTURE
Flyvbjerg's research quantifies the pattern across thousands of projects
Bent Flyvbjerg's extensive analysis of large infrastructure projects worldwide found average cost overruns in the range of 20-45% depending on project type, with rail projects among the worst — a pattern so consistent and well-documented that 'reference class forecasting' (a formal outside-view technique) is now a recommended practice in major infrastructure planning.
Where It Fails / Inversion
Where it fails / inversion
Not every schedule or budget overrun reflects the planning fallacy — sometimes a project genuinely encounters unforeseeable, non-recurring circumstances that wouldn't be captured by any reasonable reference class of similar past projects, and treating every overrun as evidence of poor planning discipline ignores genuine uncertainty that no amount of outside-view correction could have anticipated.
How To Use It
Worked example · estimating a project timeline using reference-class forecasting
Before committing to a project timeline, explicitly identify a reference class of similar past projects (your own or industry data), note how long those actually took relative to their own initial estimates, and adjust your new estimate using that historical overrun ratio — rather than estimating purely from the specific details of the current project, which is exactly the inside-view approach that reliably produces the planning fallacy.
How to use it
Before finalizing a timeline or budget estimate, explicitly look up how similar past projects actually performed relative to their own original estimates, and adjust your new estimate by that same historical ratio — rather than trusting your detailed, inside-view reasoning about why this particular project will go smoothly.
See Also