Channel · confidence vs. evidence
Planning Fallacy
The planning fallacy is the tendency to underestimate the time, costs, and risks of future tasks while overestimating benefits, despite knowledge that similar past tasks ran over.
Free-run conditions
Kahneman and Tversky (1979) coined the term 'planning fallacy' to describe how planners ignore distributional information from past similar projects and instead take an 'inside view' focused on the unique features of the current plan. Buehler, Griffin, and Ross (1994) provided the first systematic experimental evidence, finding that only 30% of honors thesis students completed their work by their predicted deadline, with actual completion averaging 55.5 days against a 33.9-day prediction. The effect has proven robust across individuals, cultures, and project scales from household chores to multi-billion dollar infrastructure.
Channel readings
Honors thesis completion prediction. Only 30% of students finished by their predicted date. Average prediction was 33.9 days (best case) to 48.6 days (worst case), but actual average completion was 55.5 days. Even the 'worst case' estimates were optimistically biased.
Buehler, Griffin, & Ross, 1994 · n = 37 undergraduate honors thesis students
Christmas shopping and tax returns. Participants consistently underestimated completion times despite accurate knowledge of past performance. For Christmas shopping, 71% took longer than predicted. For tax returns, participants finished a week later than predicted on average.
Buehler, Griffin, & Ross, 1994 · n = 52 (Christmas shopping), n = 48 (tax returns)
Inside view vs. outside view manipulation. Taking the outside view by considering past similar tasks produced more accurate predictions. However, simply imagining obstacles was insufficient to eliminate the optimistic bias.
Buehler, Griffin, & Ross, 1994 · n = 62 undergraduates
Why the signal misleads
The planning fallacy arises from the 'inside view' where planners construct mental scenarios of how a task will proceed, focusing on the unique favorable features of the current plan. This scenario-based forecasting neglects the 'outside view' base rate of how long similar tasks typically take. Kahneman and Lovallo (1993) argued that people anchor on their inside-view scenario and insufficiently adjust to distributional reality. Additional mechanisms include focalism (focusing narrowly on the target task while neglecting background obstacles), the desire to create favorable impressions through optimistic forecasts, and motivation-driven discounting of past failures.
Some researchers argue the planning fallacy is partly strategic misrepresentation rather than pure cognitive bias: planners may deliberately underestimate to win approval for projects. However, experimental evidence showing the bias in anonymous predictions suggests cognitive factors are primary.
Calibration verdict
Signal confirmed — the trace settles to a stable, replicated level.
The planning fallacy has been replicated extensively across tasks, populations, and cultures. The bias appears for individual and group projects, small tasks and megaprojects. A 2010 review by Buehler, Griffin, and Peetz in Advances in Experimental Social Psychology concluded the effect is remarkably robust. Meta-analytic evidence from Flyvbjerg's megaproject database confirms massive cost overruns across transportation infrastructure projects worldwide, with rail projects averaging 45% cost overrun.
Recorded over-runs
Sydney Opera House construction · 1973
The Sydney Opera House was originally estimated at $7 million with completion in 1963. It was completed in 1973 at a cost of $102 million, a 1,400% cost overrun and a decade late. The government insisted on early construction before plans were finalized, leading to endless rework of the innovative roof design.
Edinburgh Trams Line 2 · 2004
The first instance of reference class forecasting in practice was applied to Edinburgh Tram Line 2 in 2004. Initial estimates were 255 million pounds with 25% contingency. Reference class forecasting using 46 comparable rail projects predicted 357 million pounds (50% risk) to 400 million pounds (20% risk). The project eventually went over budget, validating the outside view correction.
Boston Big Dig (Central Artery/Tunnel) · 2006
Boston's Central Artery/Tunnel project was 275% or $11 billion over budget in constant dollars when it opened. It stands as one of the most dramatic examples of the planning fallacy in U.S. infrastructure history.
Damping
Use reference class forecasting: identify a set of similar past projects, calculate how long they actually took, and use that distribution as your anchor rather than your specific plan. Also consider obstacles explicitly by asking 'what could go wrong?' before finalizing estimates.
Buehler et al. (1994) showed that taking the outside view by considering past similar tasks improved prediction accuracy. Flyvbjerg's reference class forecasting, applied to megaprojects, has successfully reduced cost overruns by forcing planners to anchor on historical data rather than optimistic scenarios.
Reading the trace in the wild
Watch for predictions based on detailed scenario-building about how a specific task will go, especially when past similar tasks ran over. Teams that say 'this time will be different' or 'we have a solid plan' while ignoring base rates from comparable projects are likely falling into the planning fallacy. The bias appears at all scales from household chores to billion-dollar infrastructure projects.
Adjacent channels
- Optimism biassiblingBoth involve overly positive predictions, but planning fallacy specifically concerns time and cost estimates while optimism bias is broader, encompassing predictions about personal risks and outcomes.
- Inside viewparentThe planning fallacy is a specific consequence of taking the inside view (focusing on case-specific details) rather than the outside view (using distributional data from similar cases).
- Reference class forecastingoppositeReference class forecasting is the deliberate countermeasure to the planning fallacy, using historical data from similar projects to make more accurate predictions.