Sunk Costs in Portfolio Strategy: When to Cut Losses

The investor who holds a losing position because they've already lost money on it is making a decision based on the past, not the future.

This is the sunk cost fallacy in its purest form—and it's endemic in portfolio management. A fund manager watches a position decline 30%, then 40%, then 50%. The internal pressure mounts. Admitting the loss means confronting the original thesis as flawed. It means explaining to stakeholders why capital was deployed poorly. So the position lingers. Sometimes it recovers. Often it doesn't. But the decision to hold it was never really about expected returns. It was about the money already gone.

The behavioural economics here is straightforward but stubborn. Humans are loss-averse—we feel the pain of losses roughly twice as intensely as the pleasure of equivalent gains. This asymmetry makes us reluctant to crystallise losses. We'd rather hold and hope than sell and accept. The sunk cost becomes a psychological anchor, distorting every subsequent decision about what to do with the capital.

What makes this particularly damaging in portfolio strategy is that it compounds. A position that should have been exited at -15% is still held at -40% because the original loss is now larger, making the psychological pain of admission even worse. The decision-maker becomes increasingly committed to the original thesis, not because new evidence supports it, but because the cost of being wrong has risen. This is the escalation of commitment—a well-documented pattern where people invest more resources into failing courses of action to justify previous investments.

The portfolio impact is measurable. Capital trapped in a deteriorating position cannot be redeployed to better opportunities. Opportunity cost compounds silently. A manager might have identified a genuinely attractive entry point elsewhere, but the mental and emotional bandwidth consumed by the losing position—and the capital still tied up in it—prevents optimal reallocation. The portfolio becomes a museum of past mistakes rather than a dynamic allocation of future expectations.

There's also a subtle institutional dimension. Sell-side analysts rarely downgrade stocks aggressively enough, partly because they face pressure from the companies they cover. Buy-side managers face a different pressure: the need to justify their conviction. Cutting a loss early looks like a lack of conviction. Holding through the decline looks like patience and belief. The market rewards neither—it simply punishes capital inefficiency. But the social and professional incentives within investment organisations often reward the narrative of conviction over the mathematics of return.

The antidote isn't emotional detachment—that's unrealistic and probably undesirable. Instead, it's structural. Establish exit rules before entering positions. Define what evidence would falsify your thesis. Set a loss threshold that triggers a mandatory review, not a mandatory exit, but a genuine reassessment where the sunk cost is explicitly excluded from the analysis. Ask: if I had this cash today, would I buy this position at its current price? If the answer is no, the position should be sold, regardless of the loss already incurred.

This requires a different kind of discipline than most investors practice. It means accepting that some losses are the cost of learning, not evidence of bad luck. It means building a culture where cutting losses early is seen as good risk management, not failure. Some of the best portfolio decisions are invisible—the positions that never grow into catastrophic losses because they were exited when the thesis broke, not when the price had already fallen 60%.

The paradox is that acknowledging sunk costs—truly internalising that past money is gone and irrelevant to future decisions—is what separates disciplined capital allocation from hope-based investing. The future value of a portfolio depends entirely on what happens next. The past is already priced in. The only question that matters is whether the capital deployed today will generate returns tomorrow. Everything else is noise.