Comparing Mutual Fund Screening Methods Across Short and Long Samples
Summary
This research summary compares seven ways to identify active mutual funds with positive, negative, or zero risk-adjusted performance. It explains a factor-model benchmark and methods that adjust for multiple testing, including bootstrap and false discovery rate procedures, alongside a structured mixture model. Since these approaches target different quantities, some seek exceptional individual funds while others estimate the share of funds in each performance category.
Monte Carlo simulations calibrated to US equity funds compare short and long observation windows under continuous and nearly discrete manager-ability distributions. In short samples, two bootstrap approaches and one FDR approach tend to have lower bias and root mean squared error, with better coverage, than the alternatives described. Performance rankings vary for longer samples, so results from long histories may not carry over to short rolling windows. The article also reports that applying the methods to historical US fund data leads to sharply different estimates of how many funds are skilled, unskilled, or average. Results depend on model assumptions and the selected sample.
Key ideas
- Risk-adjusted alpha is used to distinguish skilled, unskilled, and zero-alpha fund managers.
- Screening procedures differ in whether they seek standout funds or estimate category proportions.
- Short-sample simulations favor certain bootstrap and FDR methods on estimation accuracy measures.
- The methods' relative performance shifts between short and long samples.
- Historical fund classifications vary substantially depending on the screening procedure.
Tags
This summary was written by Stratmill's research agent from the original; it is not a copy of the source.