Free course · Module 9
Survivorship Bias in Investing
Look beyond the winners still in view. Count who started, find who is missing and judge a track record with the full group.
Written by the StockEducation Editorial Team · Reviewed by Associate Professor Charles Lo, CPA, PhD · Updated 15 Oct 2026 · How we check lessons
Pick a topic to learn
Read in order, or return to the idea you need.
Idea 1 of 4
Find the group that disappeared.
A sample can look stronger when unsuccessful members disappear. Survivorship bias occurs when you study the members still visible and treat them as the original whole group. A current fund list, a collection of winning stocks or interviews with successful investors may leave out important experiences.
Consider four hypothetical funds, each starting with $100. Their one-period total returns are +20%, +10%, −10% and −20%, including any closure proceeds. The two positive-return funds remain listed; the other two close after that period. A survivor-only average is +15%, but all four together average 0%.
What should you ask before trusting an average?
Count the starting group before interpreting the surviving group.
Idea 2 of 4
Read the fund-survival statistic correctly.
Survival is a different measure from performance. SPIVA’s Mid-Year 2026 Report 2 lists 2,190 active US domestic equity funds at the start of its 20-year window ending June 30, 2026. Of those, 37.85% remained active. The rounded figures imply about 829 survivors and 1,361 funds no longer active. Source 1.
How does SPIVA keep the missing funds in view?
Do not confuse “still active” with “a successful investment.”
Idea 3 of 4
Look at the selection process.
What you can inspect depends on what made it into the sample. Damage on returning aircraft shows what those aircraft survived. It does not directly show damage on aircraft that failed to return. The same question applies to funds still trading and companies still in an index: what conditions selected them?
Wald’s wartime work estimated aircraft vulnerability from survivor damage using a statistical model. It illustrates the need to account for selection and assumptions, rather than treating visible observations as a complete record. Source 2.
Ask how survival affected which observations you received.
What is documented, and what belongs to the popular story?
How does the same trap enter a stock backtest?
Ask how an observation earned its place in your sample.
Idea 4 of 4
Test the claim and record the missing data.
A small difference in a reported annual rate can accumulate. Use the hypothetical model below to compare 8% with 7% over 20 years on $10,000. The one-point difference is an assumption for this exercise, not a universal correction for survivorship bias.
Compare two annual-rate assumptions
Hypothetical · no tradeScenario A minus B: $7,913
Annual compounding, no contributions, withdrawals, fees or taxes. These rates are assumptions, not forecasts.
- Reproduce the default comparison. Change the years to 10, then 30, keeping everything else fixed.
- Set both rates equal. Explain why the gap disappears without proving either rate is realistic.
- Choose a performance claim and record its starting group, missing members, dates, weighting and closure treatment.
Formula: starting amount × (1 + annual rate)years. Displayed dollars are rounded from unrounded calculations. Local practice and reading ticks do not record LMS completion.
Is one percentage point a universal bias adjustment?
Correct the dataset and method, not the answer with a guessed discount.
Quick check
Three quick checks. Feedback straight away.
A database drops closed funds, which tended to have weaker returns, and reports a survivor-only average. Relative to the same-period average including those funds, the result is most likely:
Choose an answer to see the explanation.
According to SPIVA U.S. Mid Year 2026, what share of active U.S. domestic stock funds survived the 20 years to June 30, 2026?
Choose an answer to see the explanation.
You backtest a strategy from 2006 to 2026 using only the companies in the S&P 500 now. What is the main flaw?
Choose an answer to see the explanation.
Why keep a closed fund in the original opportunity set?
Review your takeaway card and the learning path
- Start with the full group. Survivors alone can tell a different story.
- Keep measures distinct. Fund survival is not investment return or benchmark success.
- Check the selection process. Missing observations and hindsight can distort a claim.
- Use the real method. A guessed return adjustment does not repair missing data.
Reading markers and these practice checks stay on this device. They do not mark the LMS lesson complete.
Sources and assumptions
- S&P Dow Jones Indices, SPIVA U.S. Mid-Year 2026. Report 2, 20-year All Domestic Funds survival; Appendix B definitions of survivorship, underperformance and peer-return construction.
- Abraham Wald, A Method of Estimating Plane Vulnerability Based on Damage of Survivors, CNA reprint, 1980. Statistical work on aircraft survivors; Concluding example discusses protective armor. Illustrations on this page are conceptual, not historical damage records.
- Carhart, Carpenter, Lynch and Musto, Mutual Fund Survivorship, Review of Financial Studies, 2002. Abstract reports sample-length-dependent historical estimates; not a universal return adjustment.
Sources checked 15 October 2026. SPIVA figures end June 30, 2026. Four-fund returns, aircraft diagrams and compounding rates are teaching illustrations. Compounding assumes annual reinvestment and no contributions, withdrawals, tax or fees. Native LMS completion remains separate from local practice.
StockEducation provides general education, not personal financial, tax or legal advice. Nothing here recommends a specific security. Investing involves risk, including loss of money you invest.
