The Target Is Not the Benchmark

Ask an allocator how its private-market program performed last year, and the answer usually arrives in one of three forms: relative to an absolute return target such as CPI plus three percent, relative to a public market equivalent calculation or relative to a universe of peer funds.

Each of these figures offers useful information. However, not a single one is a performance benchmark, and the distinction is not purely academic. These are targets, not benchmarks. A target tells an asset owner the return required to meet a liability or a spending policy; a performance benchmark tells an asset owner whether the return they earned was strong, relative to the opportunity that existed in the market at the time. Conflating the two has become so routine in private markets that many asset owners no longer notice they are doing it — and the cost of that habit is now significant enough to warrant a closer review.

It’s understandable why an absolute return hurdle, typically something such as inflation plus a fixed spread, has been an unchallenged choice for allocators globally for so long. It’s simple, draws on readily available information and is tied to something an asset owner actually cares about — preserving purchasing power or meeting a discount rate. But a hurdle of this kind has no actual structural relationship to the asset class it is supposedly measuring. It does not rise when private equity, for example, has an exceptional outperformance period, and it does not fall when the asset class struggles. It is, by construction, indifferent to the very markets it is meant to evaluate.

In a strong year, a fund that clears CPI plus three percent by a wide margin may look like a triumph of manager skill, even if every comparable fund in the market did just as well or better on the back of favorable conditions that no manager controlled. In a weak performance year, a fund that falls short of the same hurdle may look like underperformance, even if it outpaced every reasonable alternative available at the time.

“A market hurdle doesn’t test manager skill. It tests whether the economy cooperated — and over the long run, it typically does.”

That’s the key issue here: CPI plus a fixed percent measures whether a predetermined hurdle is cleared. Over a long enough horizon, such a hurdle may become easier to meet, particularly when private equity returns, in aggregate, have historically exceeded inflation by a meaningful margin over most multi-decade periods. A hurdle like this is not a test of whether a manager actually added any value. It is a test of whether the broader economy cooperated — and over the long run, it typically does.

Public market equivalents (PME) are a step in the right direction beyond a simple hurdle rate. They attempt to compare private-market performance against an investable public alternative, which makes them directionally correct in a way absolute return hurdles are not. And yet, the information gleaned by using a PME depends heavily on which public index is chosen. There’s also a big question mark on the underlying assumptions made about the timing and reinvestment of cash flows. And then there’s the challenge that the PME may not accurately represent the geographic, sector, or leverage profile on the same terms that defined the private investment. Two analysts applying a PME framework to the same fund can arrive at meaningfully different conclusions depending on those inputs. The asset owner reading the performance results of their private investment against a PME rarely sees which inputs were included in the construction of the PME.

None of these factors make a PME illegitimate. They do, however, make it fragile enough that a single PME figure, when presented without its underlying assumptions, should raise more questions than it answers. The methodology of a PME needs to be highly defensible in its construction and inputs, and will typically require much more customized benchmark development with a greater degree of scrutiny than most boards currently give.

“Beating the MSCI ACWI Index is interesting. It’s not the same as performing well.”

There is also a more fundamental point worth making about what a PME is actually telling you. At its core, a PME is an opportunity cost calculation answering a specific question: would you have been better off putting that capital into a public market index instead? It’s a useful data point, but it’s not a performance benchmark — it is a comparison against an alternative you didn’t take. It tells you nothing about whether the fund performed well relative to what was actually achievable in private markets during that market period. Knowing you beat the MSCI ACWI is interesting, but it is not the same as knowing you performed well. 

Peer universes and internal rate of return (IRR) comparisons are more rigorous than a PME because they are grounded in real private fund-level data rather than a public index. However, it’s important to remember that a peer median simply represents the middle value in a set of fund-level internal rates of return. These IRRs are sensitive to the timing and size of cash flows and early distributions, and tend to reward funds that return capital quickly over funds that compound it patiently. Layered on top of those mechanical issues are the structural biases common to most peer databases — namely that funds that wind down or fail to report are underrepresented, and inclusion is voluntary rather than comprehensive. Most notably, peer groups are often subjectively chosen, which introduces other forms of bias that may misrepresent the comparative performance of an investment.

To be clear about the hierarchy: peer universes are the most useful of the three measures discussed here so far, precisely because they are grounded in actual private market fund data rather than an economic indicator or a public index. A well-constructed peer universe — one that is vintage-controlled, strategy-specific and drawn from a comprehensive rather than voluntary dataset — begins to approximate something genuinely useful. The problem is that most peer universes in practice are none of those things. They are broad, opt-in and survivorship-affected in ways that may quietly inflate the results.

The better answer is not a better peer universe, though better peer universes are a meaningful improvement. The better answer is a proper private markets index — constructed from actual fund-level data, applying a consistent and transparent methodology, controlled for fund type and strategy, and maintained by an independent party with no stake in the outcome. That is what transforms a peer comparison into a benchmark. The data underneath is similar, but the rigor of construction is not.

For asset owners, rigorous governance and process are important considerations when assessing the tools used to evaluate investment performance. Being able to demonstrate that those tools were objectively fit for purpose is part of that discipline. If a board, oversight committee or regulator were to ask how an asset owner knows its private-market program performed well, a CPI-plus hurdle does not provide an answer grounded in the program’s actual investment opportunity. A PME, even one constructed to mimic the fund’s composition, is, in actuality, more representative of opportunity cost than comparative performance.

This is not to say that any particular asset owner has failed a fiduciary duty. It’s an observation that there is a meaningful gap between target and benchmark, and it poses exactly the kind of question a rigorous performance analysis process should be able to answer: What is actually being measured?

Closing that gap does not require abandoning hurdle rates, a PME or even peer comparisons — each of these metrics still has a role to play in forming a complete picture. What it does require, though, is adding a component the others cannot substitute for: a benchmark constructed from actual private-market fund data, built on a transparent, objective and consistently applied methodology, calibrated to the specific strategy being evaluated. This distinction really matters because of the unique nature in which private capital is typically deployed, managed and distributed over years of varying market conditions. The return profile of any given fund may, therefore, be shaped as much by the market conditions of its vintage and strategy as by a manager’s skill and decision-making. A benchmark that does not account for those factors may provide a less representative measure of the asset class.

The methodology underpinning such a benchmark should do three things well:

1. It should use a time-weighted return calculation that properly reflects the timing and size of cash flows rather than a basic IRR calculation. IRR is useful for understanding the internal mechanics of a single fund; however, it is poorly suited to determining if a fund performed well relative to others pursuing the same strategy in the same market environment. A fund that distributes capital early will produce a higher IRR than an otherwise identical fund that does not, regardless of the actual wealth created.
2. It should be built on a unified classification framework that distinguishes meaningfully between strategies so that performance is evaluated within a relevant peer set rather than averaged across fundamentally different risk exposures. Buyout behaves differently from venture in the same way that senior secured debt behaves differently from subordinated debt. Strategy relevance ensures that the risk profile of an asset class is accurately captured in the performance measurement.
3. It should be constructed and maintained by an independent party with no stake in what asset owners own and with a transparent governance process that asset owners and their advisors can examine and challenge.

Indexes of this nature exist today. The question for asset owners is no longer whether such indexes are available; rather, it is whether their governance frameworks are structured to require them. Adopting a fit-for-purpose performance benchmark does not have to mean discarding everything currently in use, but it does mean adding a measurement that can credibly answer the question at the core of performance assessment: Given what was achievable in this market, in this strategy, perhaps even in a particular vintage year, did our program perform well?