Survivorship Bias

Survivorship bias refers to the bias that is introduced when returns are calculated from a pool of live investment funds only. Funds that die usually do so with poor returns. Since a cohort of live funds includes funds that would eventually die, it is more realistic to calculate historical returns from a pool that includes both live and dead funds. Using only live funds—namely, surviving funds—would produce historical returns that are artificially high.

There are many ways to calculate survivorship bias, the simplest being the difference between the returns of live and dead funds. It is sometimes preferable, however, to calculate survivorship bias using three different portfolios:

1. the surviving portfolio,
2. the observable portfolio, and
3. the complete portfolio.

Returns can be raw returns, returns in excess of a benchmark, risk-adjusted returns, or excess returns from a factor model. Failing to adjust for survivorship bias can lead to returns that are unduly inflated. Most studies of hedge fund survivorship bias, estimate the bias at two to four percent per year.

Survivorship Bias