Workplace Investments

The Illiquidity Bargain

Private assets may belong in workplace retirement plans, if the price is right and the dose is correct.

Sometimes it’s the same content but in a different wrapper. (A still from a Hallmark’s how-to video)

Democratization of private assets is a politically charged topic, but apparently not too hot to touch. In the US, both parties are sympathetic to the cause, with some reservations about implementation. In Canada, private assets have been quietly introduced to nearly every target-date fund available off the shelf — no questions asked, no choice provided.

The topic deserves a nuanced discussion. Let’s tackle the confusion from three different angles: (1) whether private assets are inherently superior to public assets, (2) whether retirement savings plans have unique liquidity needs effectively limiting their exposure to private assets, and (3) whether this moment in time has any peculiar valuation or funds flow distortions that warrant additional caution.

Are private assets superior to public assets?

At the dawn of time, all assets were private. Today, many types of private assets have a twin on the public side: public and private equity, public and private debt, etc.

Any investment is a means of providing financing to businesses. At the bottom of the capital stack, senior secured debt provides the highest certainty of return but also the lowest level of expected return. At the top of the stack, there is common equity — the lowest certainty of return and the highest level of expected return. From that perspective, for the same place in the capital stack, private and public assets should have approximately the same risk and return.

At a more granular level, private assets have inherent risk/return features making them distinct from public assets:

• higher transaction costs (= lower liquidity)
• lower quality of information about the underlying business.

Other things being equal, investors in private assets should be compensated for bearing these inconveniences with superior returns.

* Illiquidity premium exists, but is the price right?

The choice between a public and a private asset is mostly a liquidity trade. An investor in private assets trades away a liquidity option for a price, which is the illiquidity premium. All options have positive values, so the illiquidity premium should be positive.

The illiquidity premium is often willed into existence on theoretical grounds, and 2% is quoted frequently, without much justification. Empirical studies show that it could be as low as 0.6%.

The illiquidity premium is not a free breakfast. The fact that the expected return is higher for private assets than for public assets tells us nothing about the fairness of the offered price for the liquidity option.

The illiquidity premium is cyclically dependent on fund flows. It can go negative if there is a local glut of capital in a particular sector. Just like oil futures briefly went negative at the onset of the Covid pandemic, or how the electricity prices occasionally dip below zero.

* Sharpe ratio can’t handle illiquidity

For many advocates, private assets have a better ratio of expected return to volatility of returns — both individually and as a component of a diversified portfolio.

Measurement issues are quite evident here. Price series for private assets appear less volatile due to infrequent measurement (smoothing) and because valuations are typically based on model pricing, not transactional pricing. Un-smoothing procedures are possible, but don’t fully address the problem: comparing frequent market prices with infrequent model prices creates a volatility wedge, which looks beneficial at the portfolio level.

Portfolio managers can talk about liquidity as a risk factor. However, liquidity is not a factor that can be plugged into a risk-variance optimizer. Comparing portfolios with and without private assets on the basis of the Sharpe ratio is like comparing apples and oranges, because liquidity is not a part of the equation.

* Customized = expensive

A flow of customized financing deals cannot be managed passively. It is inherently an active management strategy, and it is costly to sustain. Any investor in private assets must believe that these costs are justified. Not only that, but also the certainty of fees is generally higher than the certainty of active returns, which makes it a high bar.

Private asset managers usually operate on a performance fee scale, such as 2% of AUM + 20% of returns above 8% hurdle rate. Note that the performance component (a.k.a. “GP carry”) does not go negative. This is tantamount to the GP getting a free call option on 20% of the portfolio. Most investors don’t realize that asset-liability studies recommending an allocation to private assets can’t model that optionality.

* Asset-liability match is a big positive

With private assets, illiquidity is not a bug — it’s a feature. Invested capital that is subject to a lock-up period and potential liquidity gates is not at risk of suddenly evaporating. Mixing long-term investments with generous liquidity promises is a sure recipe for periodic bank runs.

Any good investment manager can have a streak of bad luck. Even institutional investors can be impatient, and losing a chunk of the investor base due to poor recent performance is damaging to the firm’s ability to compensate and motivate its staff. It is also damaging to the remaining investors in the fund, as the assets are sold when the liquidity call comes, not when the manager believes is the best time to sell the assets.

Investors agreeing to capital lock-ups can be compared to Odysseus, who ordered his crew to tie him to his ship’s mast, lest he fall under the spell of the Sirens. An investment manager with secure capital can focus on the mandate and not worry too much about losing clients due to market noise. That alignment is valuable to the investor. [1]

In addition, private asset returns are usually magnified by fund-level leverage. Manager’s access to efficient financing contributes to the return advantage of private assets. Efficient fixed-rate financing is only practically feasible for structures with liquidity restrictions. Liquidity promises undermine systematic use of the case for judicious use of leverage.

To our first question (“Are private assets superior to public assets?”), the answer is — unsurprisingly — “It depends.” There are no bad assets, only bad prices. And in a portfolio context, investors need to distinguish between actual risk, in all of its complexity, and measured or reported risk.[2]

Are private assets uniquely unfitting for the liquidity needs of retirement savings plans?

Let’s break this broad question into three parts: (1) What liquidity is required and expected in a workplace retirement savings plan? (2) Are standalone private asset investment funds useful in workplace plans? (3) Can target-date funds support illiquid allocations?

* What liquidity is required and expected in a workplace retirement savings plan?

Daily pricing and daily transactions are very common (and may be expected by plan participants), but they are not required, neither in the US nor in Canada.[3]

Plan contributions, including any employer match, are a part of the payroll process. The law generally treats retirement benefits as deferred compensation. The employer is required to pay any owed cash amount into the employee’s account within a specific time frame, such as 30 days. The same applies to plan contributions into the participant’s account with the recordkeeper.

Daily pricing is common because recordkeeper systems were built for mutual fund distribution. For them, it is more complicated (and expensive) to support custom liquidity restrictions than to provide the same unified liquidity standard. [4]

Significant liquidity requirements come from retirements and employee terminations. When plan participants leave employers, some may stay in the plan, but others will take their assets with them. Ideally, fund units should be transferred from one recordkeeper to another, but that is not always possible, so account liquidation is always a significant risk for the plan. Terminations may lump together, as happens with workforce reductions, or when the employer goes out of business.

For assets in the plan, the ability to switch from one permitted investment to another is expected. DC participants are responsible for choosing among offered alternatives to reflect their risk appetite, outside assets and liabilities etc. If these underlying factors change, participants should be able to change asset allocation.

In addition, plan liquidity needs to account for behavioural sensitivity. Some participants react to market news by making changes to their account allocations. Plan fiduciaries should study past events of this kind to measure behavioral sensitivity of their participant base.

Summarizing the above, retirement savings plans generally require the ability to liquidate an account within a few months. While that is more generous than full daily liquidity, it is also far from the usual lock-up periods for private assets (5+ years).

* Are standalone private asset funds useful in workplace plans?

A standalone investment fund that is primarily invested in private assets is a poor fit for a retirement savings plan. It simply cannot guarantee the required liquidity.

The conundrum is ‘solved’ by keeping a sizeable liquidity sleeve in the fund (10% to 20%). It’s effectively a risk-sharing mechanism: when liquidity is needed by participants of one plan, it is sourced from other plans’ participants who invest in the same fund.

This approach is problematic for several reasons. First of all, the liquidity sleeve is either a drag on performance (if it’s cash), or affects the fund’s risk (if it’s in liquid alternatives), or both. Further, liquidity needs of one plan can be correlated with liquidity needs of other plans, so the risk pooling can fail when it is needed the most.

Most importantly, draining the liquidity sleeve to satisfy one plan’s participants can disadvantage the remaining investors in the fund. Redeeming some investors at smoothed valuations is effectively a wealth transfer, and remaining investors subsequently absorb it through depressed future returns.

Plan fiduciaries justify offering standalone ‘semi-liquid’ funds on philosophical grounds, to offer access to all important (large) asset classes. That approach rarely works. It is instructive to review the plan’s investment choice hierarchy. Instead of asking “What asset classes will enable my participants to construct efficient portfolios?”, ask “What reasonable investment accommodations should I provide in addition to the default diversified fund?”

* Can target-date funds support illiquid allocations?

A target-date fund with a small illiquid allocation is very different from a standalone private asset fund with a small liquidity sleeve. That’s kind of obvious.

What is less appreciated is that the risk-sharing properties of the liquidity sleeve depend on the overall size of the fund and diversity of its investor base. For most plans, target-date funds represent the bulk of the plans’ assets. Importantly, most plans use off-the-shelf TDFs, so the liquidity risks are shared by many plans. There are many TDF families with a diversified base of participating plans, which reduces the risk of synchronized termination flows.

It may sound obvious, but investors in target-date funds typically default into the TDFs because these investors don’t want to make investment decisions. The investor base of the target-date funds is distinctly more inert than the rest of the plan population. That shields target-date funds from behavioral surges in liquidity demands.

In short, a large target-date fund with a diversified investor base is not a bad place for a reasonable allocation to illiquid assets. What is reasonable depends primarily on demographics and employee turnover.

Let’s consider a hypothetical example. The average employment life of 40 years, split among 5 jobs, and full account liquidation occurs at each termination and also at retirement. This translates into a steady-state liquidity requirement of 12.5% of AUM. Most of that is typically netted against contributions. We’ll ignore salary progression.

Now, let’s consider a bad recession with a mass 10% termination across all plans invested in a given TDF and no new hiring in that year. That brings the nominal liquidity requirement to 22.5%. In addition, the value of the TDF may falls perhaps by 10%. When assets are redeemed, and a portion of the NAV is from stale priced, there is an implicit wealth transfers from remaining investors to redeeming investors.

In the table below, these implicit wealth transfers from remaining to redeeming investors in the TDF are estimated for different allocations to illiquid assets for this hypothetical scenario.

Implicit wealth transfer from remaining TDF investors to redeeming investors for a 22.5% redemption.

From the last column, we can conclude that only a low allocation to illiquids can result in a reasonably small wealth transfer effect (and minimize the risk of potential litigation). Wealth transfers must always be considered a portion of the redemption value comes from stale prices.

Is this the right time to add illiquids?

Changing the TDF glidepath is a difficult problem from a governance perspective. The result of changes will be evident in the distant future, long after the current generation of portfolio managers and plan fiduciaries leave their jobs.

Private assets, like many public assets, are cyclical: there are times when funds flow in like a flooding river, and there are times when it’s more like a small drying creek. Depending on that flow, the value of giving up liquidity may or may not be acceptable.

This year, there is a specific political push to ‘democratize’ private assets. It comes in two main flavours. One is the idea that many important companies stay private for a long time (partly because being public becomes costlier every year due to new layers of regulation). Think of Stripe, OpenAI and SpaceX. The other is the private credit industry, which mushroomed after the GFC when banks were penalized for asset-liability mismatching. In both cases, it’s a narrative about fair access to investment opportunities.

The behavioral impulse to invest like the rich do has always been strong. There are hundreds of books, YouTube and TikTok videos about this, and it is a subject of conversations in many aspiring circles, friends and families alike. There is a common view that the rich have better access to better tools, monetize that through higher returns and continue to pull ahead of the average investor.

The regulators bar the general public from buying certain investment products on the basis that disclosure is insufficient, or not provided in a standardized format, or else the nature of the strategy is too sophisticated for a typical investor to properly evaluate its risks and potential return. Hence, only qualified investors can access them.

The point of regulatory investor qualification restrictions on these non-prospectus strategies is not to make them inaccessible, but to make them inaccessible directly for investors who likely don’t have the necessary understanding and/or likely can’t afford a total loss. Advised access is still available, but it is not cheap. And so the rich can use non-prospectus strategies because they can afford the necessary advice. And if they invest without advice, they are deemed to be able to afford the loss of the investment.

The advice, from a regulator’s viewpoint, comes from someone who is competent and responsible — i.e. legally liable — either directly (as a registered investment advisor) or indirectly (as a plan sponsor with resources to hire competent consultants for making decisions). The advocates of democratization argue for shifting a part or all of the legal liability from advisor/sponsor to individual investors. After all, it’s their money, why can’t they invest as they please?

Politics, however, is never clean. This year is the year of three massive IPOs (SpaceX, OpenAI and possibly Anthropic), continued boom in data centre financing by private credit, as well as some redemption restrictions in ‘semi-liquid’ private credit funds. A cynical view is that democratization is simply a strategy to offload overpriced assets to unsophisticated investors.

The good news is that most TDF managers take careful steps. If the illiquid allocation is 1% of the fund, the specific valuation entry point is not particularly important. In TDFs with active equity, portfolio managers can make their own judgment about valuations of these large IPOs, even though there is indirect pressure to keep tracking error against the index in a reasonable range.

The change of index rules to facilitate early admission of SpaceX is a great opportunity for plan fiduciaries to contemplate their investment beliefs. What is the market? Is tracking error important? Are measurements realistic? Has indexing grown too large? Does the tail wag the dog now?

It’s certainly an exciting time, and one of the most important times in the career of a plan fiduciary.

Notes:

[1] Cliff Asness advocates that it is all backwards: institutional investors pay the illiquidity discount (in a form of lower return or higher fees or both) to get nice smooth infrequently measured returns, because it makes them better investors and prevents bad decisions based on market noise.

[2] Funding regulations for pension funds have valuable implicit real options based on short-term differences between actual risk and measured/reported risk.

[3] Similarly, there is no regulatory requirement to offer a choice of investments in a defined contribution plan, even though choice is common and expected by the workforce.

[4] Modern workplace retirement plans can also be built around ETFs held in a brokerage account with intra-day liquidity (with or without restricted list of investments).