What it means
Imagine looking at the performance of all mutual funds that exist today over the last ten years. The list will show impressive average returns, but many funds that did badly in that time were closed or merged into others, so they are no longer on the list.
The remaining funds are the survivors, and their average is higher than the average of all the funds that started. This matters in finance because decisions often rest on historical data.
A study of the companies currently in a stock index will overstate past returns, because companies that collapsed or were removed are no longer in the index. A start-up founder reading only success stories will underestimate the odds of failure.
The effect can be large. In a fund database, closed funds are often the weakest, and leaving them out can raise the apparent average return by a noticeable amount.
Similar problems arise in studies of business strategies, hedge funds and entrepreneurs, where the failures rarely write books or publish track records. The remedy is to use data sets that include failures, sometimes called survivorship-bias-free data.
Analysts should ask what happened to the entries that are missing and whether they performed differently from the ones that remain. When reading performance claims, it is wise to ask how many products were launched and how many are still available.
The nuance is that the bias is not always deliberate, and it can arise from the way data is collected. A fund company may close weak funds quietly, and a database may drop delisted companies.
Whoever uses the data must therefore check how it was built, rather than assume it is complete.
In practice
Real-world examples.
Example
A financial adviser shows a client a list of the funds that have beaten the market over ten years. The client asks how many funds started the period, and learns that dozens were closed or merged, which makes the list look better than the full picture. She decides to ask for the complete record, including closed funds, before choosing.
Example
An analyst studies the long-term return of companies in a stock index today. She realises that firms that went bankrupt years ago are not included, so she switches to a data set that includes delisted companies.
Example
A business author writes that successful entrepreneurs all took big risks. A reader points out that countless entrepreneurs also took big risks and failed, but they are not in the book. The reader concludes that risk-taking alone does not explain success, because the same trait appears among the failures.
Formula
Calculation
Survivorship bias = average return of survivors - average return of all funds that started
Suppose 10 funds started a ten-year period. Six survived, with an average annual return of 8%, and four closed after average annual returns of -2%. The average of survivors is 8%. The average of all ten is (6 x 8% + 4 x -2%) / 10 = (48% - 8%) / 10 = 4.0%. The bias is 8% - 4% = 4 percentage points, so a table showing only survivors would double the apparent average return.Case study
Seen in the real world.
Greystone Capital is an illustrative, fictional investment consultancy that was asked to recommend an equity fund to a charity. The first screen of funds showed an average ten-year return of 9.4%, which seemed strong.
An analyst asked the data provider how many funds were available at the start of the period. The answer was 120, but only 85 still existed, and the 35 that had closed had delivered an average of 3.1%.
Including the closed funds, the average fell to about 7.6%, which is (85 x 9.4% + 35 x 3.1%) / 120. In this illustrative case, the consultancy told the charity to compare funds against the full group, and built a checklist for every future search that asks what happened to the failures. The checklist also asks for the starting number of funds and the date range, so that gaps are easy to spot.
Watch out
Common mistakes.
- Judging a strategy or a fund category only by the products that are still around, when the failed ones would lower the average.
- Assuming that a data provider has included closed or delisted entries, without checking how the data was built.
- Treating stories of successful people or companies as typical, when the failures are rarely written about.
Questions
People also ask.
How can I reduce the risk?
Use data that includes closed and delisted entries, and ask how many started the period compared with how many remain.
Does it only affect investing?
No, it appears in business studies, hiring decisions, medical research and any case where failures drop out of the sample.
Is survivorship bias the same as selection bias?
It is one type of selection bias, in which the way a sample is chosen leaves out those that did not survive.
From the founder's library
Take it further with the book.
Build your financial confidence beyond this definition. Shihan's full-length guide, Accounting Fundamentals, takes the same plain-English approach and turns it into a complete, practical playbook for non-finance managers, business owners and students - with chapter-end quiz answers and presentation slides included.
25% off with code MMHQ25, applied at checkout. Priced in USD - checkout may show the equivalent in your local currency.
View the book and save 25%Related