This is an interactive narrative map. Expand, drag and select.
This is an interactive narrative map. Expand, drag and select.
IN A RECENT Financial Analyst Jour-nal article on “The future of invest-ment management”, Gary Brinson made the following observation: It’s only a matter of time before investors would revolt against the current lev-els of asset management fees being charged in the industry.
As he pointed out: “One need only measure the aggregate fees that are paid by investors with the aggregate value added by their managers to see that something is amiss... because investing is effectively a zero sum game, aggre-gate fees for most managers should be associated with passive management. Yet these fees are significantly larger than fees associated with passive man-agement. This illogical conundrum will ultimately have to end.”1Nobody disputes that asset manag-ers need to be paid for their services. The debate, rather, is over what quan-tum is fair.
How do you approach the problem of ‘fair’? One thing that’s clear is that nobody has developed a single model approach to addressing the problem. What follows is a description of three methodologies: each has important limitations but, all told, do advance our understanding of the problem.
The first addresses the question posed by Brinson. Can the performance contributions of asset managers really justify the fees charged? The other two assessment approaches address the question of “did I pay too much?” given what was delivered.
Let’s start by turning the question around. Just how much skill does an asset manager have to have to jus-tify the kinds of fees currently being charged? Richard Ennis2 provides an excellent methodology for determining that.
The Ennis plausibility model is pred-icated on the insight that the more expensive a product, the harder it is to deliver a product that will satisfy an investor’s expectations. Fees are only be plausible if they are consistent with an acceptable probability of success.
We then translated Ennis’s research into a tool that’s flexible enough to test the reasonability of fees with different skills levels and different acceptable levels of success.
As can be seen in the table, to main-tain the same level of investor suc-cess probability, the higher the fee charged for a prod-uct, the higher the fund manager skill required to justify the higher fee.
The results were a source of considerable con-cern. A manager had to be more than 61% cor-rect to justify costs as low as 30bps (0,3%). Interna-tional research points out that only top quar-tile managers exhibit that level of invest-ment skill. Worse, a sur-vey of current asset manage-ment fees in South Africa suggests that 30bps is actually on the low side, with 45bps being closer to the norm for active management mandates. Move into the unit trust space, where fees can vary between 100 and 150 bps (1% to 1,50%), and the Ennis plausibility model suggests that the manager would need to be 91% correct to merely have a 50/50 chance of adding value. Should a manager’s hit rate be as low as 30% (this actually approximates the skill level of above average active managers) there is only a 21% probability of earning a positive alpha at the 30bps. Now assume your manager skill is above average – or around 30%. The model shows that even at a management fee of 5bps (0,05%), there’s only a 28% chance of the fund manager adding positive alpha. At fees of 3%, there’s 0% probability that there will be positive alpha.
By Ennis’s calculations, most active fund management fees are simply too high given the low probability that managers’ skills are high enough to deliver positive alpha.
Let’s attack Brinson’s observation from a different direction. If, as he suggests, the base fee for active management is closer to what you’d expect to pay for a passive strategy, then perhaps a more useful exercise would be to break down a manager’s performance into its alpha (excess return not related to the market’s movement) and beta (marketrelated return) components to calculate just how much you are really paying for alpha.
Effectively, by disaggregating performance into its alpha and beta components, and assigning a passive fee to the beta component, you could back out the “true” cost (carry) of the alpha component of return from the manager’s active management fee. That was exactly the approach proposed in 2004 by the Hewitt Investment Group (HIG), a British-based consulting group. They termed this concept “alpha carry”. HIG believed the concept of “alpha carry” could be used as a way to assess the relative attractiveness of one investment product against another after the fact (ex-post).
We applied the HIG measure against a representative sample of South African unit trust value managers. In this exercise we’ve assumed that investors could obtain a passive equity (beta) exposure at 10bps. We’ve used the actual fees charged by these unit trusts in our calculation of the alpha carry. Strictly applying the “alpha carry” calculation methodology leads us to conclude that Fund A delivered the most alpha over that period and charged the least for that alpha.
But while the HIG approach to calculating alpha carry may have intuitive appeal, disaggregating alpha and beta strategies may be more complex than anticipated.
For many managers their beta bets (ie, the degree to which the manager exposes the fund to the market’s systematic risk) are also viable strategies for them to earn excess return (alpha) and, as such, that strategy deserves to be rewarded with more than purely passive fees.
The other limitation of the “alpha carry” methodology is that it can only be used on an ex-post basis. As such, it provides little insight into future alpha carry. Knowing that past performance is no guide to future performance we certainly don’t want a framework that merely rewards asset managers for last year’s outperformance – particularly as reversion to the mean in manager performance appears to be more the norm than the exception.
The AVAdd index The real limitation of the HIG methodology is that it fails to recognise that success in long-term investing owes as much to lowering investment risk as it does to absolute outperformance. We believed that if wecould enhance the HIG alpha carry methodology to recognise both the excess return and the resultant change in risk, that this measure would be a better reflection of an investor’s true long-term priorities. Now the formule reads as follows: Value add = Fund performance – Benchmark performance
Volatility of portfolio/Volatility in benchmark
That means that the value add of an active manager could be higher than another fund manager’s if he:
Applying this new approach to our same value unit trust universe it’s apparent that now Fund G, not Fund A, added more value. By using the AVAdd index it was clear that in this specific instance it was risk reduction, more so than active performance, that dictated the final ranking outcome.
How can we take the debate forward? While our paper may not have been able to identify a single definitive way to assess manager fees, perhaps it has provided a useful basis for getting the discussion going between asset managers and their clients.
Fees matter. And there may well be some justification for believing that, as Brinson suggests, asset management fees as they’re currently structured may well be excessive for the type of returns that you can expect from active managers.
But if you’re jumping to the conclusion that the obvious way to address that issue is to simply resort to performance- based fees, think again. Studies show that the base fees for most performance-based fees don’t nearly approximate passive management fees. As such, at the total expense level, there are serious questions as to whether these types of fees – where you’re theoretically only paying if you outperform – don’t also exceed the cumulative value-add of a given strategy.
Please upgrade your internet browser to Edge or use Chrome for a better experience.
(NB: Your current browser is no longer supported.)