The sunk cost fallacy—giving undue weight to costs that have already been incurred and cannot be recovered—is responsible for an enormous amount of value destruction in investment portfolios. Its hold on investor behaviour is powerful and persistent, because sunk costs feel like real considerations even when they are formally irrelevant to any forward-looking decision. The money already lost in an investment has no bearing on whether the investment is worth holding going forward; the only question that matters is what the investment will do from here, which is determined by its current price and its fundamental characteristics, not by what it used to be worth.
The formal argument against sunk cost reasoning in investment decisions is simple and well-known. If an investor holds a position that has declined fifty percent and is deciding whether to sell, the relevant question is: at the current price, does this position offer an attractive prospective return? The prior value of the position, the amount already lost, the history of the decline—none of these is relevant to this question. The market does not know what the investor paid; the company's future earnings do not depend on what the investor paid; the appropriate allocation of the investor's capital going forward is not influenced by what she paid in the past. Only the current price and the current fundamental outlook are decision-relevant.
The reason sunk costs feel relevant despite being formally irrelevant is that they carry emotional weight that is independent of their financial significance. The investor who paid one hundred dollars for a position now worth fifty has lost fifty dollars that existed, that were a real part of her financial life. The loss is a fact about her history, and facts about one's history feel like they should inform one's decisions. The instinct to try to recover the loss—to hold until the position returns to one hundred—is an instinct to rewrite history in a way that investments do not allow. The position at fifty is a position at fifty; it will not return to one hundred because the investor needs it to, and waiting for it to do so delays the redeployment of capital to opportunities that are not anchored to a prior price that has no current relevance.
The practical application is difficult because the emotional weight of the sunk cost is real even when its decision relevance is zero. The investor who knows intellectually that the fifty dollars already lost is sunk still feels the pull of holding for recovery, because the knowledge and the feeling operate through different systems that do not fully coordinate. The corrective must therefore operate at the level of decision structure rather than at the level of in-the-moment reasoning. Pre-commitment devices—rules established in advance that govern sell decisions based on analytical criteria rather than price levels—bypass the in-the-moment sunk cost reasoning by removing the decision from the emotional context that generates it.
The investor who has established that she will exit a position when the investment thesis is clearly invalidated, regardless of the current price relative to her purchase price, has created a structure within which the sunk cost fallacy cannot operate. When the thesis is invalidated and the predetermined exit criterion is met, the decision to sell is the execution of a rule rather than a fresh emotional reckoning. The rule was established under better conditions; the execution requires no fresh evaluation of the sunk costs that make selling feel difficult. This is not a complete solution—the decision to establish the rule requires confronting the possibility of loss at a time when one hopes to avoid it—but it is substantially more reliable than relying on in-the-moment judgment to override the sunk cost reasoning that large unrealised losses generate.