Disposition effect

Definition

The disposition effect is the tendency to sell winners too early and hold losers too long — realizing gains eagerly and deferring losses indefinitely. Odean (1998) measured it directly in account data: investors realized gains at ~1.5x the rate they realized losses, and the winners they sold went on to OUTPERFORM the losers they kept — the behavior was costly in both directions. It is the single most direct behavioral assault on exit discipline, which is why the platform’s management strategies are its structural antidote.

How it works / structure

• The mechanism: prospect-theory value curves (bias-loss-aversion) make realized losses psychologically expensive and realized gains cheap; selling a loser CONFIRMS the loss, holding it preserves the story that it’s temporary — reference-point accounting around the entry price drives the whole pattern.
• The measured costs: Odean’s sold-winners outperformed held-losers by ~3.4% over the following year; tax interaction compounds it (realizing losses has tax value — the behavior defers exactly the trades with fiscal benefit, acct-wash-sale mechanics); momentum’s existence (strategy-momentum) makes cutting winners and keeping losers a systematic wrong-way trade.
• Platform counters (engine-executable): pre-declared exits on BOTH sides at entry (mgmt-stop-loss, mgmt-profit-target — the decision made before the reference point forms), holding-period asymmetry diagnostics (average days-held for losing vs winning positions — the account-level fingerprint), and loss-realization rate tracking (Odean’s PGR/PLR ratio computed per account).
• The professional form: institutional accounts show it weaker but present (documented); models inherit it when trained on human-labeled exits.

When it applies

Every discretionary exit decision; account review (the PGR/PLR fingerprint is computable from any trade log); strategy design (exit rules that reference entry price inherit the bias’s geometry — R-multiple frameworks partially launder it into discipline).

Risk profile & failure modes

• Stop-widening in flight: the effect’s live signature — moving a stop away as price approaches it converts a planned loss into an unplanned catastrophe (mgmt-stop-loss no-widening rule exists for this).
• Breakeven magnetism: holding an underwater position “until it gets back to even” — the entry price has no market meaning; the behavior anchors risk to an irrelevant number.
• Tax-season inversion: the one period the bias reverses (December loss-harvesting is documented) — wash-sale mechanics await the unwary (acct-wash-sale).
• Overcorrected exits: mechanically cutting every winner at a fixed gain replicates the sell-winners half — asymmetric exit design (let profits run) is the momentum- consistent alternative.

Evidence & limits

Shefrin-Statman (1985) named and framed it; Odean (1998) measured it in 10,000 accounts; replications span countries, asset classes, and professionals (attenuated). The costs are sample-measured and interact with momentum regimes — in strong reversal regimes, the behavior accidentally profits, which is why the diagnosis is behavioral (rate asymmetry), not outcome-based.

Falsifiable-thesis examples

Illustrations only, not signals:

• “This account’s PGR/PLR ratio exceeds 1.3 (disposition fingerprint present)” — falsified by the trade-log computation.
• “Enforcing entry-declared two-sided exits will reduce this account’s average-days-held asymmetry by half within a quarter” — falsified by the before/after diagnostic.

Cross-references

• The engine underneath: bias-loss-aversion (prospect theory)
• The structural antidotes: mgmt-stop-loss, mgmt-profit-target, mgmt-time-based-exit
• The market pattern it fights: strategy-momentum
• The tax interaction: acct-wash-sale