What’s essentially the most surprising thing about aggregate private markets performance calculations?
The widespread “tolerance” of mathematical errors, gross inaccuracy and representativeness amongst private market investors, consultants, enthusiasts, critics and even academics.
In traditional asset classes, investment professionals focus their attribution evaluation on every “micron” of performance difference. However, with regards to private market assets, over-convergence is commonplace.
The choppy waters of personal equity performance attribution
The variability of money flows makes performance attribution of personal market assets significantly tougher: returns usually are not generated by a stable underlying asset base, so there isn’t a possibility of reinvestment or compounding.
As I’ve written before, today’s performance attribution toolkit consists of metrics – internal rates of return (IRRs), total deposit values (TVPIs), public market equivalents (PMEs), and the varied alphas – that work in a single fell swoop on the asset level at best, but can can’t be generalized.
So what does generalization actually mean?
In non-mathematical terms, generalization allows for meaningful comparisons. We should give you the chance to say whether a given IRR or TVPI is objectively “better” than one other, whether it represents more return or less risk.
For two comparable investments, is an IRR of 15% higher than 10%? While the optical illusion implies this, in point of fact we cannot give an actual answer without more data. We need information in regards to the time and capital invested. This means time-weighted metrics and never the money-weighted approximations currently used.
This 10% IRR could also be preferable whether it is earned over an extended time period, say 4 years, versus two years for the 15%. This ends in a 1.4x invested capital multiple (MOIC) for the ten%, which beats the 1.3x MOIC for the 15%. But we still need a duration component to achieve an inexpensive conclusion.
According to the IRR narrative, money recovered earlier may very well be reinvested at the identical rate of return. But that is just an assumption. For fixed income securities, early redemption is often treated as reinvestment risk. Past returns are not any guarantee of future results.
But let’s stir the waters much more and solid one other stone.
Is a 1.4x MOIC higher than a 1.3x? Of course, right? In fact, all of it will depend on how much actual capital is deployed and what capital is to be deployed. If the 1.4x MOIC is generated by drawn capital that is simply 50% of a reference obligation and the 1.3x MOIC is predicated on an analogous obligation that’s 100% drawn, the latter outperforms the previous.
Based on this logic, all derived PME and alpha calculations are subject to the identical conceptual limitations. As a result, any money-weighted quartile information and rankings of and about private market investments can introduce significant data bias.
Mathematically, the generalization implies that additivity is a prerequisite for any robust statistical evaluation. The example above shows that without precise additivity we cannot determine a representative average.
The rules of economic mathematics dictate that rates of interest can only be averaged through compounding. However, the interior rate of return can’t be calculated appropriately over time. When IRRs are presented as annualized ratios or horizon measures, or worse for accuracy, as returns, this is feasible seriously misrepresent the actual returns.
But even when the IRR may very well be composed as in our MOIC example without further capital use information, the character of MOICs prevents us from properly averaging their performance.
The average IRR of our two hypothetical investments will not be 12.5%, neither is the common MOIC 1.35x the true average return. Again, we want a duration component and capital weighting data before we will make meaningful estimates.
The pooling trap
In aggregate private equity return calculations, the gross approximation is much more striking. Studies often pool money flows and treat the money flows of various funds as in the event that they got here from a single fund. This distorts the information even greater than in our previous examples.
Annualized differences in the worth of many basis points are treated without regard to mathematical accuracy or representativeness.
Pool money flows
The table above shows the money flows of three funds of various sizes and origins individually, pooled and pooled and weighted. This implies that the money flows are calculated pro forma, with the person money flows being weighted by the relative weight of the person funds.
The pooled IRR of 9.14% is different from each the (mathematically incorrect) individual fund weighted average IRR of 6.95% and the pooled weighted IRR of 8.13%. Nevertheless, the performance number should clearly reflect the worth created by the funds.
What’s even worse for accuracy purposes is that or from inception to the last reporting date. Even with the more conservative pooled weighted return, the idea because the starting suggests that the 800 pooled units of invested capital (1+8.13%) can be ^10=2.18x or 1,748 units.
Since their inception, pooled returns have created an obvious discrepancy. The 800 units of capital invested within the three funds produced “only” 1,160 units of capital, which is way below the “impression” implied by the pooled returns since inception.
Unwarranted trust is usually the results of a return from the initial horizon. As the instance shows, they create the illusion of increased wealth, on this case by an element of 1.5. This helps explain why marketing documents feature far too many 10x private market benchmarks.
The DaRC life jacket
One of the very best pieces of recommendation I’ve ever received is to never trust the currents from a pool or the ocean, or simply aggregate calculations. Always watch out.
To prevent accurate information from being lost within the PE pool, the duration-adjusted return on capital (DaRC) methodology ensures the needed everlasting framework. It first corrects the multiples by making an allowance for the timing of money flows after which leverages the additivity attributes of duration.
This keeps the pooled multiplier according to actual money flow balances: 1.45x. We then calculate a reputable average time-weighted DaRC net return of 8.39% with an inexpensive net maturity of 4.68 years.
To optimize allocation and risk management for a diversified portfolio, we want accurate performance metrics. But current private market metrics all too often fall wanting this benchmark. We can do higher.
If you enjoyed this post, remember to subscribe.
Photo credit: ©Getty Images/PBNJ Productions
