Desirability Bias

Desirability bias is the psychological distortion of estimated probabilities in favour of outcomes the decision-maker wishes to occur. When desirability bias takes over, an investor forecasting the returns of a favoured stock subtly inflates the likelihood of success and deflates the likelihood of loss. A founder estimating the odds of a successful acquisition drifts optimistic. A portfolio manager biased toward growth funds over-weights the probability of sustained momentum. The bias is not outright delusion—the numbers may pass a surface plausibility check—but motivated reasoning silently tugs probabilities away from reality toward the outcome the brain prefers.

How desire distorts probability

Desirability bias operates through selective interpretation and unconscious pruning of evidence. A piece of good news for a stock you own gets absorbed as “meaningful signal”; equivalent bad news gets dismissed as “transient noise.” A forecast that a favorite sector will outperform feels intuitively right, so you underweight contrary research. The brain is not lying to you; it is unconsciously padding the case for the preferred outcome and discounting the case against it.

Neuroscientific research shows that when we contemplate a desired outcome, brain regions associated with reward light up. The desire creates a kind of affective pull on our reasoning. Mathematically, desirability bias tweaks probability distributions: the tails shift subtly, the mean drifts, and the entire distribution tilts. A forecast that previously peaked at a 50–50 chance now leans 60–40 or 65–35 toward the desired end—not dramatically, but enough to stack long-term decisions.

This bias is particularly powerful because it operates largely outside conscious awareness. A fund manager updating a discounted cash flow model on a company she hopes to outperform may unconsciously apply a slightly lower discount rate, a slightly higher terminal growth assumption, or a slightly rosier view of competitive advantages. None of these tweaks feel dishonest in the moment; they feel like “reasonable best estimates.” Aggregate them across hundreds of daily decisions, and the portfolio drifts systematically optimistic.

Desirability bias in equity valuation

Stock valuations are particularly vulnerable to desirability bias because they are forward-looking and rest on estimates rather than facts. A value investor who has already purchased a stock has a desire for it to be a good investment. This desire subtly raises her probability estimate of the firm’s long-term success, the stickiness of its competitive advantage, and its ability to maintain margins. A growth investor with a position in a high-flying tech stock estimates its odds of market-share gains more generously than an unbiased observer would.

The danger is cumulative. If desirability bias inflates the expected return by 200 basis points—from a true expectation of 5% annually to a biased estimate of 7%—compounding over 20 years translates that small shift into a vastly overstated terminal value. A £100,000 position that should grow to £260,000 in real terms (5% annually) gets mentally modeled as growing to £386,000 (7% annually). The bias thus distorts position sizing, rebalancing, and stop-loss discipline.

Sector and asset-class allocation

Desirability bias also tilts asset allocation. An investor who loves dividend-paying stocks and views them as “safer” may unconsciously overestimate the probability that dividend stocks will outperform in future bear markets, and underestimate the tail risk of dividend cuts during crises. A real estate enthusiast may overweight the probability of sustained rental yields and appreciation, underweighting the risk of foreclosure waves or construction booms that collapse valuations.

Sector rotation decisions are rife with desirability bias. A manager who believes energy stocks are undervalued and wishes to bet on a recovery may overestimate the probability of an oil-price rebound. The bias is not that she consciously manipulates numbers, but that she gravitates toward research and data that support the desired outcome and prunes counter-evidence. Over time, the portfolio drifts overweight toward energy, not because new information arrived, but because desire tilted her probability estimates.

Leverage and risk tolerance

Desirability bias interacts dangerously with leverage. A trader or fund manager who wishes to juice returns (a desired outcome) unconsciously lowers her estimate of the probability of a severe market drawdown. She may think: “A 30% bear market is unlikely; my track record and risk models suggest we can safely apply 1.5x leverage.” But desirability bias has already tilted her probability estimate of that drawdown downward. The outcome: she applies more leverage than the true risk profile warrants, believing the probabilities support it.

This pattern contributed to multiple financial crises. In the years leading to 2008, mortgage originators and bond investors desired high yields and low default outcomes. Desirability bias subtly shifted probability estimates: “Housing prices always go up” became less of an explicit claim and more of a background assumption; the probability of simultaneous defaults was underestimated; counterparty risk in securitized structures felt manageable. Each individual estimate passed a plausibility test; collectively, they were wildly optimistic.

Desirability bias in M&A and capital projects

Desirability bias is rampant in mergers and acquisitions. An acquiring firm often desires the deal to be strategically brilliant and financially accretive. This desire subtly inflates the probability estimates of synergies, revenue run-rates post-acquisition, and integration success, while deflating estimates of one-off costs and execution risk. Academic research shows that acquirers consistently overestimate synergy realization and underestimate deal costs, a pattern consistent with desirability bias.

Capital project evaluations suffer similarly. A manager proposing a £50 million expansion plant naturally desires it to succeed. Unconsciously, she may overestimate the probability of market demand materializing, underestimate competitive response, and lower the discount rate she applies to future cash flows. The project passes the NPV hurdle, partly because bias has tilted the numbers in its favor.

Detecting and countering the bias

Desirability bias is hard to spot because it operates through apparently reasonable probability estimates. The antidote requires structural discipline:

Pre-commitment to assumptions: Before entering a position, write down explicit probability forecasts for key outcomes (e.g., “70% chance of 10%+ annual returns, 20% chance of 0–10%, 10% chance of negative returns”). Later, resist the urge to casually upgrade the first number to 75% or downgrade the third to 5%.

Red-teaming and adversarial review: Assign someone to argue the case for the outcome you do not want. This person’s job is to inflate the probability of failure or underperformance and make the case compellingly. Force yourself to engage with it seriously rather than dismissing it as “pessimism.”

Backtesting subjective estimates: Track your probability forecasts over time and compare them to realized outcomes. If you consistently underestimate downside probabilities (i.e., bad outcomes happen more often than you predicted), you have evidence of desirability bias and can correct for it.

Using reference classes and base rates: Instead of reasoning from the specific case, ask: “Across similar deals, what is the actual success rate?” Base rates are harder to bias than bespoke forecasts because they force you to reckon with aggregate reality rather than the specific story you want to believe.

See also