Sunk Cost Bias in Trading

The sunk cost bias in trading is the tendency of investors to hold losing positions longer than rational economics justifies, because they want to recoup their initial investment. A trader who bought a stock at $50 but sees it fall to $30 may refuse to sell, reasoning “I will only sell when I break even”—a decision based on past loss (sunk cost) rather than forward-looking value. This bias combines loss aversion with mental accounting fallacy and is a major driver of portfolio underperformance and disposition effect behavior.

Why sunk-cost bias persists despite being irrational

The rational decision rule for a trader is simple: compare the forward-looking expected return of a position to alternative uses of capital. If a stock is worth $25 today and you hold it at $30, your entry price is irrelevant; the fact that you are down $5 per share is a sunk cost. Selling the loser and redeploying the $30 into a stock with better prospects (higher expected return per unit of risk) is the correct move. However, most traders violate this: they check the position daily, feel pain from the loss, and think “if I wait long enough, it will come back to $50 and I will be even.” This magical thinking springs from:

1. Loss aversion: The pain of a realized loss is roughly 2–2.5x the pleasure of an equivalent gain (in behavioral finance terms, losses have higher absolute value than gains). Holding the loser delays realizing the pain.
2. Mental accounting: The trader sees the position as a separate “account” with its own gain/loss balance, not as part of a unified portfolio. Breaking even on the position becomes the goal, not maximizing total portfolio return.
3. Regret aversion: The trader fears that if the position recovers to $50 after they sell at $30, they will experience acute regret (“I should have held”). This prospective regret is stronger than forward-looking decision-making.

The disposition effect: empirical proof of sunk-cost bias

The disposition effect is the empirical observation that investors are more likely to sell winners than losers—the opposite of tax-efficient behavior and risk management. Studies of retail trading data show that investors hold losing positions roughly 2x longer than winning positions before selling (Odean, 1998). The mechanism: realizing a loss triggers deep loss aversion, so traders hold losers to avoid the pain; realizing a gain is painful in a different way (it closes the “account,” removing upside optionality), but the fear of regret if the winner continues rising makes sellers accept the realized gain. The net result: portfolios accumulate losers and are short winners—a structure that guarantees underperformance.

Market evidence: holding losers extends drawdowns

Portfolios weighted toward held losers underperform because the stocks remain loser precisely because they have weakened fundamentals, sector headwinds, or management dysfunction. A trader who bought a bank stock at $50 before credit quality deteriorated and it fell to $30 is now holding a deteriorated asset. The rational response is to sell and avoid further downside; the sunk-cost response is to hold and wait for recovery that may never come. Data on post-loss holding periods shows that stocks held by investors experiencing large unrealized losses underperform the market by ~5–10% annualized over the holding period—the cost of the bias compounds over time. Conversely, stocks that investors sell at gains (avoiding regret) often continue outperforming, leaving alpha on the table.

The endowment effect and reference points

The endowment effect is related: people overvalue assets they own relative to equivalent assets they don’t own. If you own a stock at $50 (entry price as reference point), you might reject a $45 bid because you feel entitled to $50 despite the stock’s fair value being $40. The reference point is arbitrary (what if your entry was at $60, not $50?), yet it shapes decisions. Breaking a loss-making position requires shifting the reference point from entry price to current price and then to fair value—a mental effort most investors skip. Instead, the entry price becomes the “aspiration level,” and selling below it triggers loss aversion.

Loss aversion and risk-taking: the flip side of sunk cost

Sunk-cost bias is not the only loss-aversion manifestation. A trader down $100k on a position might start taking excessive risks (“double down” or add leverage) to recover the loss in a single trade—a form of gambler’s fallacy or desperation gambling. Conversely, a trader might exit a trade too early (taking a small loss) because the pain of watching a position move against them is intolerable. Both extremes (holding too long, exiting too fast) are driven by loss aversion but with opposite symptoms. The optimal behavior is to ignore sunk costs entirely and evaluate each trade on forward-looking merits; few investors achieve this.

Cost-basis addiction: the numerological trap

Many trading platforms display the entry price, unrealized loss, and percentage loss in the portfolio view. This design anchors the trader to the entry price, reinforcing sunk-cost thinking. A platform that showed only the current price, fair value estimate, and forward-looking recommendation would reduce bias, but it would also reduce user engagement (traders check positions less frequently if they don’t see the loss). Thus, platforms cater to sunk-cost thinking, and the interface itself becomes a behavioral bias amplifier. Some disciplined institutional traders avoid viewing their entry prices; they see only the current price and a rating (overvalued, fair, undervalued) to force forward-looking analysis.

Behavioral rules to override sunk-cost bias

Institutional traders combat sunk-cost bias with:

1. Stop losses: A pre-set rule: “If the position falls 20%, exit automatically.” This removes the emotional decision and honors the sunk-cost fact upfront.
2. Portfolio-level rebalancing: Rather than evaluating each position individually, rebalance the portfolio monthly or quarterly to target allocations, selling winners and losers mechanically. This depersonalizes decisions.
3. Reference-point shifting: Explicitly recompute the position’s fair value and set that as the new mental reference, discarding the entry price. This is harder than it sounds; humans naturally revert to entry prices.
4. Pre-mortems: Before entering a trade, ask: “If this loses 30%, how will I feel, and will I hold or fold?” Deciding in advance reduces in-the-moment emotion.
5. Win rate vs. risk-reward focus: Measure success not by “beating my entry price” but by risk-adjusted return and hit rate (% of trades profitable). This shifts focus from individual breakeven to portfolio alpha.

Sunk-cost bias in corporate finance: capital rationing

Sunk-cost bias is not unique to trading. Corporate managers often continue funding failing projects because they have already invested $50 million (“we are too deep to abandon it now”) rather than because the project’s prospective returns justify it. This is capital rationing bias applied to sunk costs. The correct decision rule—evaluate projects solely on forward-looking cash flows—is ignored in favor of “recovering the past investment.” This leads to escalating commitment and enormous write-downs when management finally admits the project was a mistake.

Tax-loss harvesting as a counter-strategy

One rational reason to sell a loser is tax-loss harvesting: a trader can realize the loss (and thus the pain), claim the tax deduction, and immediately repurchase the same asset or a similar one, capturing the loss on the tax return while staying in the market. This removes the emotional sting (“at least I got a tax deduction”) and allows the investor to move on. If a trader bought $50k in a tech stock and it is worth $35k, harvesting the $15k loss yields ~$4.5k in tax deductions (at 30% marginal rate). This trade creates a true benefit, offsetting the loss-aversion pain.

Mutual funds and institutional herding in sunk-cost losses

Mutual funds often exhibit sunk-cost bias at the institutional level. A fund manager who bought a stock when it was a hot idea and it has since underperformed may hold the position far longer than warranted, hoping to avoid admitting the mistake to investors and regulators. This can lead to concentrated losses in a fund’s portfolio and underperformance that persists for years. Index funds and passive strategies avoid this by construction: they hold every stock equally, not based on manager beliefs, so sunk costs cannot bias holding decisions.

Cognitive reframing to escape the trap

Psychologists suggest cognitive reframing to escape sunk-cost thinking:

• Instead of “I bought at $50, now it is $30, I am down $20,” reframe as “The stock is currently worth $30. Is buying at $30 a good idea with forward-looking fundamentals?” If the answer is no, sell.
• Instead of “I will hold until I break even,” reframe as “This capital is earning X% return here, versus Y% in the next-best use. Where is it better deployed?” Invariably, the sunk-cost position ranks lower.
• Instead of “I don’t want to regret selling before a recovery,” reframe as “Would I buy this stock today at the current price?” If not, holding is irrational.

These reframes are cognitively demanding and most investors lack the discipline to execute them consistently. This is why the most successful trading systems are mechanical: they remove human decision-making from loss-aversion triggers.