The sunk cost fallacy is a logical fallacy or flawed argument for decision-making. In the sunk cost fallacy, prior investment is cited as a reason for pursuing a course of action. The term is usually used to describe persisting in a bad investment, on the grounds that otherwise, the time, money or effort invested in the project will be wasted anyway. The expression "throwing good money after bad" is sometimes used to describe this behavior.
In economics, a "sunk cost" is any cost that has already been paid and is impossible to recover. In purely rational decision-making, sunk costs should not have any influence on decisions, since they cannot be recovered. For instance, consider the situation of a man who, having already purchased a train ticket, is offered a quicker ride to his destination. The fact that he has already paid for a train ticket should have no bearing on the decision to take or refuse the faster ride, since the money spent on the first ticket is lost whether he chooses to accept the ride or not.
The term "sunk cost fallacy" describes a common situation in human behavior in which sunk costs influence decision-making despite the fact that they have already been spent and are not recoverable. For instance, a person who buys a book and begins reading it may find out that she does not enjoy it. She may, however, continue to read on, citing the fact that she has paid for it. This is irrational, since she will have paid for the book whether she reads it or not, and in fact is merely wasting time on an unpleasant activity instead of doing something she enjoys. This is an example of the sunk cost fallacy in action.
In some cases, the sunk cost fallacy can lead to escalation of commitment on a large scale. For example, British and French government investment in the Concorde super-sonic transport actually increased once it became clear that the project was likely to lose money. From a rational perspective, abandoning the project would have been superior to continuing to invest.
The sunk cost fallacy may sometimes be a misinterpretation of decision-making processes. In many cases, there are consequences to actions which may not be taken into account by a purely economic analysis. For example, consider the case of a government investing money into a project. There may come a point when, as with the Concorde, from a financial perspective, the best course of action is to abandon the project and invest in an alternative. Abandoning a project, however, might have negative political consequences such as damaging voter confidence. Persisting therefore brings benefits not accounted for in a purely rational model.