Survivorship Bias Not Significant In Equities

May 9, 2005 (—Lipper research on survivorship bias in US mutual funds shows that survivorship bias matters little or not at all for many fund types - yet varies each year for others.

The influence of survivorship bias seems not to have a significant impact for global equity funds, mid-cap core funds, and small-cap core funds on both a one-year and three-year basis – but for other categories, such as large-cap growth funds, the importance of the bias changes each year, according to the report.

However, among multi-cap funds, the research shows the bias matters “in all years and in all styles.”

When examining survivorship bias, Lipper says it is necessary to examine both style and capitalization, and that previous studies have not examined capitalization in depth. Survivorship rates differed within styles, among different capitalizations slightly over a one-year period, but significantly over a three-year period for value funds, less so for core and growth funds.

Time also has a significant impact on whether or not bias matters. For any one-year period, the bias has what researchers called “a small to middling impact;” basically nothing.   However, as the test period lengthens towards three years in Lipper’s research (and further in other studies, Lipper said), the bias matters more.

“Die” Caste

Lipper’s research shows that it is not necessarily the poorly performing funds that “die.” Funds disappear also as a result of market performance, and because fund management companies have “thrown in the towel.”   Many more funds died during the period of 2000-2002 than in 2003 and 2004, timeframes that Lipper associates with the bear market being too difficult for management to handle. According to the data, there have been years where funds put out to pasture had better average returns that those that survived.

Mergers and acquisitions can also have an impact on fund survivorship because in the process of transitions, sometimes good or better performing funds get removed. “Given the number of mergers that have occurred in recent years,” researchers wrote, “we wonder if the survivorship bias impact from these recent mergers has been a net zero on survivorship bias.

The study concluded that survivorship bias is most important for asset management firms and academic studies, but not so for the retail investors, who are most concerned with what funds are available for their purchase at a particular moment.  T he funds were tested both examining style purity and without style purity, which showed no real difference in the conclusions, Lipper said.