Survivorship bias is the mistake in analysis of concentrating on processes or examples which succeeded and ignoring or downplaying those which failed. The most common result of this is drawing overly optimistic or positive conclusions. In a financial setting, this usually involves analysis which leaves out companies or funds that have failed and no longer exist.
Within finance, the most common form of survivorship bias involves keeping track of past investments, particularly mutual funds. For example, a company may launch 100 mutual funds. Five years later it may have dropped 25 of these funds completely or merged them with other funds, in both cases because of performance. This is normal behavior, as most financial firms see little point in keeping a consistently poorly performing fund open.
The problem occurs when the company produces figures showing its performance over the past five years. An average figure may well only include the remaining 75 funds because there is naturally no complete five-year data available for the 25 which have been dropped. This means the average is much more skewed towards funds which perform well.
This can be extremely misleading as an investor looking at the figure might expect a similar return on their investments over the next five years. In reality the chances are that the investments won't perform so well as the company will continue to launch some funds which perform badly. Some estimates suggest survivorship bias could mean performance estimates across the mutual fund industry are overstated by an average of almost one percentage point.
It's also arguable that some stock market indices are prone to survivorship bias. For example, an index may track the largest 100 companies in a particular market. From time to time this list will be revised to take account of changes in company size. In many cases, the companies dropping out of the list will have "shrunk" because their stock price has fallen.
This means that at any particular moment, the index will be less likely to reflect stocks which are performing particularly badly. This survivorship bias means the overall movement of the index is likely to be more positive than that of the market as a whole. The effect is not as pronounced as with mutual funds because some of the negative performance does show up in the index figure before the relevant stock is dropped. For this reason, some economists argue that stock market indices should not be classed as exhibiting survivorship bias.
