Portfolio Construction: Assembling the Building Blocks

Executive Summary:

  • Estimating private asset returns in one of the key building blocks of portfolio construction. The nature of private asset behavior means there are competing return metrics. LPs should understand the advantages, shortcomings, and use cases of each calculation to accurately assess returns.
  • Private markets funds like to market lofty return targets. Actual return data paints a different (but very attractive!) picture, highlighting the need for robust data to calibrate return assumptions.
  • Data suggests that returns vary meaningfully by strategy and time period, providing optionality for LPs to construct portfolios that fit their return goals.

In Volume 1 of our series on portfolio construction we introduced what we believe are the three key building blocks of portfolio construction: return, risk and liquidity. These are the three critical attributes that must be estimated for each private markets strategy, fund or asset in order to thoughtfully develop and implement a private markets portfolio. They are arguably the most iconic trio, with apologies to Kim, Kylie and Khloe.

We frequently refer to risk, return and liquidity as the building blocks of private markets portfolio construction. While literal volumes of research have been conducted in an effort to quantify these attributes, their behavior is still not widely understood. We aim to contribute to that growing body of literature by dissecting each of those building blocks using what we believe to be among the most comprehensive private asset datasets in the world in conjunction with our 30+ years of experience managing private markets portfolios.

First block up:

Return Calculation 101

Before we delve into what the data suggests about private markets returns, here’s a brief technical interlude on measuring private markets returns. Those who have spent their days studying the various flavors of public market equivalents (PMEs) and can differentiate between Simple Dietz and Modified Dietz can skip this section like it’s a university class at 8AM on a Friday (which, of course, no one at Hamilton Lane ever did).

Traditional private market investments have the inconvenient feature of irregular cash flows, both in terms of timing and size. This means the default rate-of-return measurement tool for private markets investments is the internal rate of return (IRR). That is a different measurement tool than the time-weighted return (TWR) methodology which is typically used to measure listed asset returns. In a TWR calculation, we calculate the change in total value (adjusted for inflows and outflows) for a single period, which is typically quarterly for private markets given that this is the frequency at which valuations are struck. A series of single-period returns is then compounded and annualized to determine the average annual return over the desired time horizon.

TWR vs. IRR Definitions

There are cases where IRR and TWR will give similar answers. For example, a portfolio with a relatively constant net asset value (NAV), plus cash flows that are small in relation to that NAV will yield a similar (but not the same!) TWR and IRR over a reasonable time horizon. In other cases, IRR and TWR can yield very different answers. For example, the performance for a single fund commitment often looks very different in IRR terms than TWR terms since the capital at work can fluctuate substantially.

TWR & IRR Comparison: Single Fund Return

We should introduce one more comparative rate-of-return metric that is commonly used by private markets practitioners and is favored by academics studying the asset class: The public market equivalent, or PME. In simple, approximate terms, the goal of a PME is to simulate an investment in a publicly listed index or security.

Some readers may question where multiples such as total value to paid-in capital (TVPI) or distributions to paid-in capital (DPI) fit in our performance evaluation equation. After all, those are figures that many managers proudly tout on the fundraising trail and can be more insulated from certain manipulation techniques than IRRs. While those figures are useful for manager diligence and benchmarking, they are less useful in the context of asset allocation. Multiples are cumulative return calculations that don’t – on their own – account for the duration of the investment, putting shorter duration investments at an inherent disadvantage. Multiples also do not lend themselves well to time horizon calculations for portfolios that are constantly making new investments. The prevailing annual rate of return is preferable in the context of asset allocation decision making.

So, which metric is best for evaluating private markets performance? As we said in Volume 1 of our series, the best approach is to look at all of them to piece together as much of the return picture as possible. It’s a bit like video assistant referee (VAR) at a football match – though hopefully we do a bit better than the VAR that worked the Liverpool v. Tottenham match last October.

Calibrating Return Expectations

If you survey a room full of investors about their return expectations for private markets, you’re likely to get a wide spread of answers. That’s understandable: Return expectations are often colored by early experiences with the asset class and initial meetings with general partners. And general partners are perennially (and necessarily!) an optimistic bunch. They have high expectations for the returns that they can deliver to their investors and there is a marketing incentive to guide their return expectations higher.  

Over the first two months of 2024, Hamilton Lane screened over 200 private markets funds. Here’s what the stated expected returns – in IRR and TVPI terms – for those opportunities look like:

GP Stated Return Expectations

Average of Opportunities Screened by Hamilton Lane in 2024

Pretty appetizing! But remember, these are numbers in placement memorandums and marketing materials. Is there data to back up those claims? Let’s focus on private equity, which, on average, lays claim to the loftiest target returns.

Private Equity Fund Net IRR Cumulative Distribution

Vintage Years 2010-2020

The data doesn’t support those target returns as strongly as some may hope. While there is a substantial segment of the industry delivering the >20% net IRRs they promised, nearly two thirds of the funds from 2010 – 2020 vintages fell short of that 20% hurdle. Keep in mind that this is a fairly successful period for private equity. A significant chunk of the > 20% cohort is unrealized and still subject to market risk. It’s also worth noting that these net IRR figures may be impacted by lines of credit (LoCs) and other performance enhancers that have the potential to boost reported net IRR without meaningfully increasing the wealth created for LPs.

But 20% is a very high hurdle! Consider this: The average public market equivalent for these funds (using MSCI World) is around 8%. A fund returning 20% would be generating a 1,200 bps premium to that hurdle. That’s massive! The average funds in this cohort are generating returns in the mid-teens, representing hundreds of basis points of outperformance.

Vintage Year Returns

Let’s move away from individual funds since most LPs are not building portfolios of single funds. LPs commonly look at IRR on a vintage year basis since it represents the total since inception return from the time of the investment decision. That is relevant for closed-end fund investors since the primary allocation decision is made at the time of commitment and rebalancing an exposure incurs significant frictions.

Pooled IRRs by Vintage Year & Strategy

The table above highlights since-inception vintage year IRRs since vintage year 2000 by strategy with summary information at the foot of the table. We define “Good Year” and “Bad Year” to be the average vintage year IRR plus or minus half of a standard deviation.

The vintage year IRR data confirms some commonly held beliefs and offers a few surprising ones:

  • On average, private equity vintage year IRRs tended to be higher than private real asset or private credit returns. Private equity returns averaged net IRRs in the mid-teens, while real assets and credit were a couple hundred basis points lower.
  • Venture capital, surprisingly, does not have a higher maximum return or “good year” IRR compared to its private equity peers. The strategy seems to come with more variance in vintage year returns, which is perhaps less of a surprise given the perception of venture capital as a “high risk” strategy.  The strategy has posted a run of strong vintage years post-GFC, where vintage year IRRs have averaged 20%, though some of those years have substantial unrealized value remaining.
  • Some might also be surprised that the average vintage year IRR for large cap buyout funds is above small- and mid-cap (SMID) buyout funds. There is a widespread perception that small fund performance dominates large fund performance. In aggregate, that is not always the case .
  • Private credit has historically exhibited the lowest average returns and least variance across vintage years, consistent with the perception that credit is a less risky strategy. Hold this thought for when we get to the investment pacing discussion.
  • Distressed debt vintage year returns have underwhelmed relative to private equity returns, despite distressed debt often being marketed with “equity-like” returns. LPs should carefully consider whether distressed debt should occupy a private equity-oriented portfolio.   In our view, distressed debt is more at home in a private credit-oriented portfolio.
  • Real estate has shown that it can be sensitive to economic bubbles. Note that the funds from the 2005 – 2007 era, which were investing substantial capital at the height of the real estate bubble, just barely returned their principal to investors. Strip out those vintage years and the average vintage year returns for real estate rise to about 13%, which approaches private equity territory.  
  • Infrastructure is an interesting case. That dataset is younger and the pre-2010 sample contains funds focused on emerging markets which were energy infrastructure-focused funds. The composition of the infrastructure fund market is much different today. This is a segment where the oldest historical aggregate data may not be as reflective of what to expect going forward.

The bottom line: There is meaningful return variation both by strategy and by vintage year. On the latter point, it is sometimes helpful to look at relative returns to provide context. A vintage year with private equity posting a 15% IRR may seem attractive at first glance. But if an equivalently timed and sized investment in public equities also gained 15% annually over the same time frame, you may start to wonder why you’re paying all those fees. 

So, let’s compare the vintage year IRRs we just showed to an MSCI World PME. We can quibble over which index makes for the most appropriate benchmark. The MSCI World, which is widely used by investors as a private equity benchmark, is reflective of the global public equity portfolio that most investors hold (making it a good measure of opportunity cost) and it has a similar geographic composition to an all-private markets composite.

IRR vs. PME Spread by Vintage Year and Strategy

While a complete adjudication of whether private equity generates superior performance to public equities (and how private equity generates those premium returns) is beyond the scope of this paper, the table above suggests private equity strategies have historically and consistently bested global listed equities by hundreds of basis points per annum. The dotcom era and immediate follow-up vintage years for venture capital are the glaring exception. 

Credit strategies have outperformed global equities in some vintage years but not others, and by a smaller margin. Absent a few outlier years in the early aughts, private credit performance is only about 50 bps above listed equities.  And that’s OK! The return target for most credit managers is not usually tied to listed equities – it’s more likely to be tied to interest rates. We see some evidence of a similar phenomenon in real estate and infrastructure, which have lower correlations to traditional assets. 

All that is to say: The benchmark assigned to the private markets portfolio will be a key determinant of the private markets portfolio construction. A portfolio striving to best a return hurdle of MSCI World plus 500 bps will necessarily demand a different portfolio construction than a portfolio that assigns separate benchmarks to each private markets strategy.

Horizon Returns

So far, we’ve looked only at individual fund and vintage year returns. These are useful datapoints for understanding potential outcomes of new investments from the time of inception. But most investors build portfolios consisting of dozens of funds investing across many vintage years. The asset allocation decision should consider total portfolio returns over intermediate and long-term time horizons as much as the returns of single investments or single vintage years. Especially since portfolio performance will be a mixture of the performance of assets already in the ground (investment decisions made years ago) and the performance of new investments.

Below we show rolling one-year time weighted returns for private equity, private credit and private real asset strategies.

Rolling One-Year Returns by Strategy

This viewing angle seems to confirm and expand on some of the takeaways from the vintage year IRR data:

  • This angle shows more of the high potential of VC returns. Look at how well they did during the halcyon days of 2021! But, like with the vintage year IRRs, average one-year returns for venture capital are similar to the rest of their private equity peer strategies.
  • We split the real asset returns summary table into “all periods” and “post 2010.” The “post-2010” sample reduces the impact of the GFC on real estate returns and the impact of the thin sample of infrastructure funds in the aughts. The inclusion of pre-2010 periods has a material impact on the distribution of returns for both strategies. The post-2010 sample yields average returns closer to expectations.
  • Credit had some periods of strong single-year returns in the aughts, especially for distressed debt. But those returns have been more muted in the post-GFC era. For origination credit, this could partially reflect the expansion of credit into “safer,” more senior strategies that have lower target returns.

But these are only single-year returns. They are helpful for understanding the range of potential short-term outcomes, but we are most concerned with what to expect over some sufficiently long-time horizon. We select 10 years since it seems to be a common “long-term” horizon used by LPs. Ten years also aligns with the legally defined fund term of many private markets funds (although whether most private markets funds actually wind down by year 10 is topic for another paper…).

Rolling 10-Year Returns by Strategy

The 10-year pictures present a little more cleanly, since long-horizon returns exhibit less variance. This view exposes a few more relevant data points for information portfolio construction:

  • Long-term returns for buyouts have most frequently ranged from 11% - 14%. Those figures are a bit lower than the vintage year IRRs suggest, though that isn’t surprising. These horizon returns are less susceptible to some of the financial engineering games that vintage year IRRs (especially younger ones) are sensitive to.
  • Growth equity exhibits seemingly attractive characteristics, since it appears to have much less downside than venture capital, while still retaining some of the upside. Growth equity posted better 10-year returns than buyout in the early aughts and since 2014. Food for thought: The post-2014 period coincided with a general preference among investors for growth factor exposure over value factor exposure.
  • The effects of GFC-era funds weigh on horizon returns for real estate until 2017, when those funds began to liquidate and fundraising re-accelerated (so newer funds began to occupy a larger share of the index). Since that period, real estate returns have bounced back to low-to-mid-teens levels. Keep in mind this is a mix of risk (core-plus, value-add, opportunistic) profiles.
  • The upward slope of infrastructure returns is consistent with our earlier comments around the changing composition of that market and the resulting impact on returns. We think that the recent infrastructure returns are more representative of what to expect going forward.
  • While they don’t approach the peaks of private equity returns, the returns among credit funds originating loans have remained remarkably consistent, or at least more consistent than other private markets strategies. Distressed debt funds have not consistently generated equity-like returns over the last decade, disappointing investors that expected private equity-like returns from the strategy.

Next Up

There is a lot more we could say about private markets returns (i.e., how do they respond during public market drawdowns? What about during periods of high inflation or high interest rates? Do small funds do better than large funds?) but for now we’ll leave you with a (very obvious) clue for the next volume in this series:

Related: The Tradeoff Between Risk & Return