Survivorship bias is a type of sample selection bias that occurs when a data set only considers “surviving” or existing observations and fails to consider observations that already ceased to exist.
In finance, an example of survivorship bias is when studies on mutual fund returns only use databases that contain data about mutual funds that currently exist, and fail to include data about funds that are no longer existing.
The mutual funds could’ve been closed for a variety of reasons such as mergers and acquisitions, restructuring, or poor financial performance. Another example of survivorship bias could occur when studies on the profitability of certain industries fail to include financial information about acquired or bankrupt companies. In such a case, financial analysts base their evaluation solely on companies that currently exist in the market.
Impact of Survivorship Bias
Generally speaking, survivorship bias tends to create conclusions that are overly optimistic, and that may not be representative of real-life environments. The bias occurs because the “surviving” observations often tend to have survived due to their stronger-than-average resilience to difficult conditions, and leaves out other observations that have ceased to exist as a result of such conditions.
Going back to the mutual fund returns example above, a study that exhibits survivorship bias can skew the returns positively as it only considers mutual funds that are currently in existence. The mutual funds have survived difficult economic conditions such as a recession through either their composite breakdown or the ability of the fund’s management to react to the market and adjust their investment strategy.
However, during the recession that the funds have survived, other mutual funds were likely forced to shut down due to poor performance. The net effect would be positively-skewed results that would fail to accurately depict the actual returns realized by all mutual funds.
Therefore, when evaluating the returns of mutual funds (regardless of the time horizon), it is important to consider all mutual funds that meet the criteria for the study. If a study specifically aims to measure the performance of the top or “surviving” mutual funds, it should be disclosed in the research paper’s methodology.
Example of Survivorship Bias
Consider the following information about mutual fund returns:
Assume that all funds meet the criteria set out by researchers.
If we only considered the funds that are still active, the average return calculated would be 9%. By contrast, if the study included all possible observations that met its criteria, the calculated average returns would be only 3% – two-thirds less than the return calculated under survivorship bias.
Thus, it is very important for researchers to carefully vet the information that they choose to use in their studies. However, this is easier said than done, as some researchers may fall into the trap of survivorship bias despite their best attempts at mitigating such a risk.
For example, when dealing with large databases that contain thousands or millions of data points, the omission of observations may be more challenging to keep track of. Thus, the onus is also on database managers to ensure that their data sets do not contain survivorship bias. This can be done by implementing new rules and procedures, complying with strict standards, or educating staff about best practices when it comes to logging data.
How to Prevent Survivorship Bias
In order to prevent survivorship bias, researchers must be very selective with their data sources. Researchers must ensure that the data sources that they have selected do not omit observations that are no longer in existence in order to reduce the risk of survivorship bias.
CFI is the official provider of the global Financial Modeling & Valuation Analyst (FMVA)™ certification program, designed to help anyone become a world-class financial analyst. To keep learning and advancing your career, the additional CFI resources below will be useful: