Revision, and its imitation#
Back of the Envelope · Week 7
Nobody will ask you to lie#
The request is never for a different answer — it is for a second look at one assumption, and the reason given is usually fair.
The pattern only shows up in aggregate.
Are you sure about that growth rate?#
- Are you sure about that growth rate?
Asked once, this deserves an honest answer — and sometimes the honest answer is yes.
Twelve requests, all pointing one way#
- Are you sure about that growth rate?
- Isn’t that contingency conservative?
- Shouldn’t the comparison include the benefits we always count elsewhere?
- …and nine more, every one aimed the same direction.
No individual exchange contains a lie. The estimate drifts anyway — and it drifts in a direction somebody chose.
20% and 30% against 10% and 15%#
Base case (“do nothing”) — 20% (P50) and 30% (P90) contingency.
Project case — 10% (P50) and 15% (P90).
Each figure is arguable alone. Set side by side, they push the same direction — the direction that favours the project.
−$579.5 million, funded in advance#
NPV of −$579.5 million. Benefit–cost ratio 0.57 — 0.68 with wider economic benefits.
$2,030 million and $2,500 million committed, before the evaluation was even published.
Nobody hid anything. The analysis was published, and the money had already been committed.
Did something about the world change, or something about the room?#
New evidence, a corrected error, a genuine factual find — revise, and say what you learned.
Who is asking, how senior, how many times they’ve raised it — not reasons.
A third move: widen the range, leave the centre. Honest when a challenge reduced your confidence without saying which way to move.
Defences that work by being in place beforehand#
Assumptions written down and visible. Comparison against projects that actually happened. Review by someone with no stake in the outcome.
None of these requires courage in the moment — that is exactly their virtue.