Sunk Costs Are Irrelevant and Almost Nobody Behaves That Way
Money already spent cannot be recovered by any future decision, so it should not influence any future decision. This is simple to state, genuinely difficult to practice, and expensive to ignore.
The Principle
A sunk cost is an expenditure already made that cannot be recovered. The rule from economics is that sunk costs should not influence decisions, because a decision can only affect the future, and the money is gone under every available option.
The only relevant comparison is between incremental future costs and incremental future benefits from this point forward. What was already spent appears in neither column.
Why It Feels Wrong
The rule conflicts with a strong intuition about consistency and waste. Abandoning a project after spending heavily feels like admitting the spending was wasted, and people are averse to that admission.
The uncomfortable truth is that the money was wasted at the moment it was spent on something that will not pay off. Continuing does not recover it, it adds to it. The choice is between losing what was spent and losing what was spent plus whatever is spent from here.
Continuing a failing project does not recover the money already spent. It only increases the total amount lost.
Where It Appears in Corporate Finance
The most common form is a capital project running over budget. A factory approved at 100 million has consumed 80 million and now requires another 60 million to finish, with an expected value of 90 million on completion.
The instinct is to note that total spending would reach 140 million against a 90 million asset and cancel. That reasoning is wrong. The 80 million is gone whether the project is completed or abandoned. The actual question is whether spending 60 million more to obtain a 90 million asset is worthwhile, and it plainly is.
The reverse error is equally common. A project that has consumed enormous sums but requires more than the remaining value should be stopped immediately, regardless of how much has been invested and how painful the write off appears.
Research and Development
Pharmaceutical development is the clearest recurring case. A compound may consume hundreds of millions across years of trials. When interim results are discouraging, the decision to continue should depend only on the probability adjusted future value against remaining trial costs.
Organizations frequently continue because teams have invested careers, because cancellation is a visible failure, and because momentum is easier than reversal. Companies that build formal stage gate reviews with predefined criteria do so specifically to force the decision to be made on forward looking evidence rather than on accumulated commitment.
The Personal Version
The same error appears in investing as the disposition effect, holding losing positions to avoid realizing a loss. The purchase price is a sunk cost. The only question is whether the asset is attractive at today's price given today's information.
A useful test is to ask whether you would buy the position today at the current price if you did not already own it. If the answer is no, the reason for holding is psychological rather than analytical.
Why Organizations Are Worse Than Individuals
Escalation of commitment is amplified in organizations because the person deciding whether to continue is frequently the person who approved the original spending. Cancelling means publicly acknowledging a bad call, and career incentives argue for continuing and hoping.
The structural remedy is to separate the approval decision from the continuation review, so that the person judging whether to proceed did not authorize the original commitment.
The Bottom Line
Only future costs and future benefits belong in a decision. Everything already spent is context, not input, and the organizations that build that discipline into their review process are the ones that stop bad projects early.