Quartiles vs. Mean-based Benchmarks

By: Avi Turetsky and Alinah Shahid

What is the best way to measure private equity fund performance? This question stands at the core of the fund selection process. In our experience, industry participants traditionally use a simple method, which is to calculate the internal rate of return (IRR) or total-value-to-paid-in capital (TVPI) of a fund, and then to benchmark this figure against a peer set of funds, using quartiles. 

The obvious intent of this method is to evaluate whether a given private equity fund outperformed or underperformed its competitive set. However, methods such as IRR and TVPI quartile benchmarking have important limitations. These limitations include, but are not limited to, the fact that a fund’s quartile performance may not provide a good indication of its contribution to the return of a real investor portfolio, the fact that benchmarking against a peer set may not be the most effective way to strip out market conditions and other ‘luck’ factors that can provide non-repeatable benefits to a fund, and the fact that quartiles ignore the often disproportionate impact of the very best and worst funds and companies in any portfolio or peer set. In other words, IRR and TVPI quartile rankings may not provide all the information that investors think they do, and in some (or perhaps many) cases, they can even be misleading.

This new white paper series is intended to provide private equity investors with a set of practical tools to overcome these limitations. In putting this series together, we draw from market-leading work by practitioners and academics in both public and private markets, in order to bring best practices from these sources into the real world of private equity investment.

The above is a brief synopsis of the white paper referenced in the title. It is not itself a complete record of that paper and cannot be relied upon in isolation. A copy of the full white paper is available upon request.

Enter your email to receive this white paper and other research.