Survivorship Bias
Survivorship Bias is the logical error of evaluating only successful outcomes — funds, companies, or strategies that still exist — while ignoring the failures that have disappeared, producing overly optimistic assessments of average performance.
Survivorship bias is structurally embedded in how investment performance data is collected and presented. Databases of mutual fund returns typically include only funds that are currently operating. Funds that were liquidated, merged into other products, or shut down due to poor performance are removed from the historical record. Since poor-performing funds disproportionately disappear, average returns of 'surviving' funds systematically overstate what a typical investor would have experienced.
Research by Mark Carhart and others has quantified the survivorship bias in mutual fund data as adding approximately 1-2 percentage points per year to reported average returns. This is a substantial distortion: a strategy that appears to generate consistent alpha on survivorship-biased data may simply be capturing the statistical artifact of dead funds being excluded from the denominator.
The same logic applies to stocks. Index back-tests frequently use current index constituents as the historical universe, implicitly assuming investors would have owned only those companies that survived to be included today. Companies that went bankrupt, were acquired at distressed valuations, or were delisted due to failure are excluded, making the strategy appear far more reliable than live trading would have produced.
Survivorship bias extends to business narratives and case studies. Books about exceptional companies — 'Good to Great,' 'Built to Last' — study firms that achieved extraordinary long-run success. The management practices and cultural attributes of those firms appear causally linked to success because the equally similar firms that employed those same practices but failed are absent from the analysis.
For investors evaluating fund track records, back-tested strategies, or stock-screening methodologies, demanding survivorship-bias-free data is a basic due-diligence requirement. Point-in-time databases that include delisted securities and failed funds provide a more realistic picture of expected outcomes.