The planning fallacy
Almost everyone has had the experience of starting a task with a confident estimate of how long it will take, only to find themselves working late into the night as the deadline approaches. Psychologists have a name for this: the planning fallacy. The term was introduced in 1979 by Daniel Kahneman and Amos Tversky to describe our tendency to underestimate how long future tasks will take, even when we know that similar tasks have taken longer in the past. Later research showed that people often underestimate costs and risks in the same way.
The effect is remarkably consistent. In one well-known study, university students were asked to predict when they would complete their final-year thesis. On average, they finished almost three weeks later than their own estimate, and fewer than a third finished by the date they had predicted. Strikingly, many students also failed to meet the 'worst-case' estimate they had been asked to give. The pattern is not limited to students. Software projects, home renovations and even simple household jobs show the same tendency, and large public projects are notorious for overrunning; the Sydney Opera House, to take a famous example, opened about a decade later than planned and cost many times its original budget.
Why do intelligent people keep making the same mistake? One explanation is that when we plan, we focus on the specific task in front of us and imagine the steps going smoothly. We construct a mental scenario in which nothing goes wrong, and we rarely think about interruptions, illness or the simple fact that some parts of a task are harder than they look. Optimism also plays a part: believing that a task will be quick makes it easier to start, while admitting that it will take months can be discouraging. Past experience should correct this, but we tend to treat earlier delays as unusual events caused by bad luck rather than as a reliable guide to the future.
Interestingly, people are much better at predicting how long other people will take. Observers who are given the same information as the planner usually produce longer, and often more realistic, estimates, perhaps because they are not emotionally invested in finishing quickly and pay more attention to how long such tasks usually take. Asked when a colleague will finish a report, most of us automatically allow for the delays that we ignore in our own plans. This difference between the 'inside view' of the planner and the 'outside view' of the observer has inspired one of the most effective remedies.
Instead of building an estimate from the details of the current task, planners can start with a 'reference class': a group of similar tasks whose outcomes are already known. If comparable projects have typically taken twice as long as expected, that figure becomes the starting point, adjusted only slightly for the specifics of the case. Some governments now use this approach when planning major roads and railways, where an optimistic estimate can cost billions. A second technique is to break a task into its smallest parts before estimating. Listing every step makes forgotten tasks visible and tends to produce longer, more realistic forecasts. Neither method eliminates optimism entirely, but both shift the estimate in the right direction. The aim is not to become pessimistic, but to plan for the delays that past projects show are almost certain to occur.