An alternative approach to asset manager fees

In a previous post, I argued for a new metric – cumulative dollars earned ‘CDE’ (and for this to be compared to ‘cumulative fees earned’ by the manager). Secondly, my colleagues at WTW have long argued that a fair asset manager fee would be no more than one-third of the gross value created. This balances the need to compensate the agent for their skill with the recognition that the principal is supplying all the capital at risk. It is time to combine these ideas.

Sticking with the status quo ad valorem rate arrangements for the time being, how do we approach the principle of the fee should be no more than 33% of the value added? There are two choices: (1) predict the manager’s future gross alpha and agree to an annual fee representing one-third of that amount, or (2) use a performance fee mechanism to calculate payments after the event. Clearly, with the first option, actual experience in an individual case is likely to differ from initial expectation – for better or for worse. In aggregate, however, given that alpha is a zero-sum game then we know that this approach will mean the asset managers take more than 100% of the value created, which is not the intention.

Does this mean we have to go down the performance fee route if we are to solve the macro issue? Regrettably, because I dislike the complexities necessary to correct for the unwelcome side effects of traditional performance fees, I think the answer is ‘yes’.

Therefore we need a less-complex solution, and I believe that paying a share of cumulative dollars earned offers a fair and transparent alternative. In principle, we measure the CDE and the asset manager is entitled to 33% of that amount. In practice there is a little more complexity but, I would argue, nothing like the complexity needed for current performance fees. I suggest the necessary elements are agreement on:

– The value sharing (say 33%, but could be different)

– Any base fee element (in extremis this could be zero). An obvious reference point would be the appropriate index-tracking (or perhaps the appropriate smart beta) fee rate. The opportunity could be taken to move away from the basis point structure within the industry and set a dollar payment rate (possibly indexed to wage inflation)

– A withholding mechanism. Changing the fee structure will not remove the noise from the performance results, and so there will still be a requirement to protect against cumulative overpayment. One option would be a symmetrical clawback system, where in a subsequent year the manager returns money to bring the cumulative fees paid back to the agreed share of CDE. On the assumption that this would be too painful for the asset manager, a withholding rate (say, 50%) could be agreed. The earned-but-not-paid part of the fee would be carried forward to the next calculation date. I am aware that there are some (but not many) performance fee structures with such mechanisms already in place for long-only equity mandates, but it is different from the current arrangements in the alternatives field and so there may be implications (such as tax crystallising) which could make this unworkable. Of course in private equity there are 100% withholding mechanisms – the problem there is that fees are paid on total return rather than alpha.

The mechanics of calculating the fee are then (fairly) straightforward. At the end of the first year the value of the benchmark portfolio is calculated – this is a notional portfolio that starts at the same size as the real portfolio and changes in value in line with the benchmark or index (and is adjusted to mirror the cash flows into and out of the real portfolio). The difference in the dollar value of the actual and benchmark portfolios is the dollar value created (or detracted) by the manager (the CDE). The share accruing to the asset manager is then calculated, say 0.33 x CDE. From this, the dollar value of the base fee paid over the year is then subtracted, leaving the dollar value of the performance fee element. As suggested above, a proportion would be paid immediately and the remainder withheld until the next calculation.

The crucial aspect is that subsequent years are continually added so that the cumulative dollars earned are calculated over the whole life of the account. There are no rolling periods from which bad years can drop out, causing a fee boost, and there is no need for high water marks. If the asset manager adds considerable value over time, they pocket 33% of it (or as agreed). If they do not add any value at any stage, they only collect the low base fee. It is possible for a large fee to be earned in a single year, and for no value to be added after that. If the manager is terminated at that point, they may have earned more than the agreed 33% share, but the asset owner will have been partially protected by 50% of the pay-out being withheld.

Of course there is the complexity of how the accrued but unpaid performance fees are released on termination, but again this is relatively straightforward.

I think this is a fairer, better aligned mechanism. Anyone up for it?