Predicting Performance: How Persistent are Private Fund Returns?

Written By: Gerald O’Dwyer the PE Guru

“Past performance is not a guarantee of future results.” 

All investment professionals are familiar with the phrase, which is more than just legal fine print to cover any liability when discussing financial returns. The words are a simple warning stemming from lived experience; relying on a manager’s or an asset’s track record to drive investment decision-making has led many capital allocators astray. 

And yet, the temptation to chase returns is a recurring theme that takes shape in various formats across asset classes. In private markets specifically, the belief in “top quartile” managers is pervasive, resulting from past research that suggests the top-performing general partners tend to keep outperforming. 

Further, the empirical data on performance dispersion between top and bottom quartiles (up to 20%+ deltas in IRR depending on the vintage and strategy) signals that GP selection is incredibly important. Plus, a commitment can be locked up for 10 or more years, which means getting the selection wrong offers a painful lesson in opportunity costs. 

Enter the theory of performance persistency.
 

Its underpinnings seem valid enough. The opportunity set for private markets is vast, and a GP’s access to deals is not uniform like it is, say, in US public equities, where investors have the same list of a few thousand stocks to choose from. Add the fact that private market GPs offer operational expertise and hands-on support, and it’s easy to imagine that a GP’s alpha can be persistent. 

Our latest research puts that theory to the test. We look to the returns generated across closed-end funds in GPs’ fund families (a sequence of funds in a similar strategy), leveraging components of our newly released Performance Score framework. Our results for PE, VC, real estate, and funds of funds suggest that performance persistency does in fact exist in private markets, but the practical implications for capital allocators are limited. 

Overall, we found that strong performance in a family’s prior funds was correlated with strong performance in successor funds to some degree, but that’s when looking at the most recent available returns data. To be useful for the allocator’s decision to invest, persistency must be predictive at the time a new commitment can be made. 

The issue―which we and others have found―is that early IRRs are notoriously unreliable indicators of where a fund’s performance eventually ends up. At the time the next sequence in the family is fundraising, the predecessor fund is on average only 3.5 years old, leading to a high degree of drift from that time to its final IRR. The vertical distance between each dot and the 45-degree line in the scatter plot below corresponds to the amount of drift each fund has experienced.
 

Relationship between IRR at time of successor’s fundraise and latest available IRR 

We find that, after controlling for information on prior fund family returns that was available at the time a successor was fundraising, the predictive power of a GP’s track record is near zero. 

In our latest Allocator Solutions report, we explore this data and more. While we don’t recommend using past performance to predict future results, there are other practical takeaways for allocators gauging a manager’s track record.