It’s not Trump. It’s not Facebook and the Russians. It’s not even the latest “news” about the Kardashians.
No, for the world of institutional investing, the topic of our time is none other than fees.
Consultants draft tedious white papers on the optimal structures. Regulators condemn asset owners for fee waste. Asset managers pledge to lower fees. And asset owners explore new structures. The unifying theme: the admission that current fee structures for actively managed public market strategies unfairly favor asset managers and that this needs to change.
One idea that has received a disproportionate amount of attention is the 1-or-30 hedge fund fee structure introduced by Britt Harris and his team during his tenure at the Teacher Retirement System of Texas. While 1-or-30 represents a hard-fought improvement over the entrenched 2-and-20 model, it is hyperbolic to claim, as one well-known chief investment officer did, that the 1-or-30 structure “develops the revolutionary concept of ‘or’ — and this is where you find the real magic.”
Any magic is really just sleight-of-hand meant to distract us from realizing how low our expectations are for any meaningful improvement in the existing misaligned fee structures.
Another oft-cited improvement is Japan’s ¥162.7 trillion ($1.5 trillion) Government Pension Investment Fund’s adoption of a new fee structure for its traditional active managers (approximately 20 percent of its portfolio). GPIF told the Financial Times that it sought to “build a win-win relationship between GPIF and external asset managers” by offering managers a passive-level base fee, no cap on performance fees with a carryover mechanism, and multiyear investment contracts.
As I would fully expect, not all asset managers seek a win-win.
Pensions & Investments reports that “an executive with one GPIF manager, who declined to be identified, said as Japan's marquee client, the GPIF can set its own rules — but his firm would have preferred to continue enjoying the certainty fixed fees provide.”
Unnoticed and undiscussed are alternative fee structures developed by a handful of managers, some that refund performance fees to clients during periods of underperformance or offer a rational management fee derived from the company’s operating budget. (These managers are certainly progressive, but they are not pioneers. In 1956, Warren Buffett offered investors in his fund a 25 percent performance fee, a 6 percent hurdle, and his personal guarantee to absorb at least some of any losses.)
These and other structures incrementally improve the inequity of existing fee structures — but they do not recalibrate the alignment in favor of the allocators.
In fact, traditional fee structures prohibit such recalibration because by they are unfair and misaligned by design: If the fundamental relationship between asset owners and managers is that an asset owner gives capital to an asset manager and pays that manager some fee (management and/or incentive) for an uncertain return, then no amount of tinkering with the absolute values will shift the asymmetry in favor of the allocator.
Regardless of the “compromises,” traditional fee structures (including 1-or-30 and GPIF) require the allocator to bear performance risk, even though it cannot affect this risk (other than by redeeming). Moreover, the manager, in spite of its stated conviction and perhaps co-investment, offers no assurance that the return target will be attained, yet is generally assured of receiving some payment for services. (Even 0-and/or-30 does not eliminate the asymmetry of performance risk.)
To achieve fee structures aligned with the objectives and competencies of most allocators, we must expand the window of possible choices to include those that will be seen as utterly unthinkable by today’s standards.
What does such an unthinkable choice look like?
An asset owner gives capital to the manager, who in turn pays the allocator “rent” for the use of its capital. The amount of this rent is not based on the future, uncertain performance of the underlying strategy but rather is derived from an agreed-upon statistical metric related to the manager’s investment strategy (e.g., expected return).
The allocator no longer pays fees to the manager for the use of its own capital and is assured of receiving the investment outcome it seeks (i.e., the negotiated rent). The manager gets the capital and potential revenue it needs to run its business. (I first came upon this rent idea when speaking with Britt Harris at a conference in 2016. I’ve been researching it since then. I found one source that offered a similar idea. The author replied to my initial query but went silent thereafter. Perhaps his thought leadership proved too progressive, although his article remains on the Mercer website.)
Unlike current fee models, this model starts from a position of alignment because it does what current models fail to do: It transforms the manager from agent to principal. (Of course, the strength or weakness of this alignment is negotiable.)
As principal, the manager bears the full risk of performance (after all, it is the party best suited to manage this risk) and receives no predetermined payment for its efforts. These conditions should fully incentivize the manager to prudently manage its business by assessing with a high degree of confidence the strategy’s expected return or other benchmark metric used to derive the rent and carefully managing its asset gathering to ensure assets under management do not adversely impact performance. Additionally, the manager also strives to understand the market environments in which its strategy might excel/struggle and in those times asks clients for additional capital/returns capital to clients (with a call on these funds for future investment). And without question, this realignment forces managers to tightly control their costs, especially those painful but often unspoken pass-through fees, where funds bury such costs as travel, research, and bonuses.
Under this structure, a manager is risking not just a portion of its wealth but its entire business — the ultimate expression of skin in the game and alignment.
I put this rent idea to some of the best and brightest asset allocators in our industry. Once they put aside the behavioral issues (“no manager would accept this”), they all supported it in some fashion.
They did raise several objections. For example, this structure shifts an allocator’s risk from investment performance to the creditworthiness of the manager. True, but allocators already bear this risk with the current fee structure; this rent model makes it explicit. Moreover, this counterparty risk could be managed through prudent manager selection: Just as in current structures, asset owners should select managers they believe will meet or exceed their investment targets and prudently manage their business so it can weather the inevitable periods of underperformance. The use of an escrow account and collateralization were also mentioned.
Others pointed out that the use of this structure might be restricted to larger, established firms. Perhaps, but this is a business decision I leave to the manager, and implementing this idea will certainly result in adaptations currently unforeseen. Others wondered whether this structure produces tax complications or requires a change in regulatory oversight (e.g., would the Financial Industry Regulatory Authority and the Securities and Exchange Commission continue to be the appropriate regulatory bodies for U.S. managers?). I’ll leave these and other details to the appropriate experts.
Over all, my engagement with these asset allocators fortified my hope that this rental model might actually be adopted. Several admitted having thought about such a structure. A U.S. public fund, independent of my research, is currently asking prospective managers to explain why this rental model should not be used. And one nonprofit has this model in place with a current manager.
We need to realize that fees are the means by which an allocator can influence the manager to provide the desired risk-return profile. Our existing fee models structurally limit an allocator’s influence and allow the perpetuation of current inequality and misalignment.
Other structures do exist. Of course, they may appear radical or unthinkable. But remember: At one point, any change from 2-and-20 was radical and unthinkable.
And how did that turn out?