Vol. 16, No.3, 2018
Gregory C. Allen, Ivan S. Cliff and Walter J. Meerschaert
The authors propose a practical technique to correct for survivorship bias across return distributions for investment product universes. The technique is designed to work efficiently in a large-scale performance measurement environment. It uses all available data for
survivors and non-survivors, corrects for bias across the full distribution (from 1st to 99th percentile), and can be applied to other return-based statistics such as Sharpe ratio, standard deviation, and correlation. The technique is applied to a variety of product
universes over a 10-year period to highlight the practical ways that it can be used to improve the investment decision-making process.