Using Optimization to find Synthetic Equity Universes that minimize Survivorship / Selection Biases

When testing quantitative models on single equity level, it is often not possible (or computationally too expensive) to use the full index constituents’ history in order to avoid significant distorting biases.

A more pragmatic approach is to find a representative universe by selecting a subset of the index constituents instead. However, this approach is sensitive to survivorship and/or selection biases.

In this paper we present an innovative approach based on optimization techniques, which avoids the typical pitfalls and is very efficient.

You can download the paper here: