Paid for the Work, Not the Number

Fund administration is priced in basis points because fund management is priced in basis points, and the convention transferred without anybody re-deriving it. It is worth re-deriving, because the two are not the same activity and the pricing that suits one misprices the other in a direction that is easy to demonstrate.

The commitment here is concrete and checkable: fees are based on the extent of the administration performed, not on the value of the assets in the fund. Under the time-based model, they do not automatically increase with growth in the fund's assets under management. An alternative monthly minimum is agreed against stated assumptions — the number of shareholders, the frequency of subscriptions and redemptions — with additional fees where the administration exceeds that minimum.

What administration actually scales with

Take the work seriously for a moment and the mismatch becomes arithmetic rather than argument.

Administration burden scales with the number of legal entities in the structure, because each one has its own filings, its own records and its own calendar. It scales with jurisdictional count, because each territory adds a regulator, a reporting format and a set of deadlines that do not align with anybody else's. It scales with transaction volume, since subscriptions, redemptions, transfers and capital calls are each a discrete piece of processing. It scales with reporting obligations, which are set by the vehicle's classification, its investor base and its markets. And it scales with complexity — share classes, currencies, side letters, waterfalls, in-kind arrangements.

Not one of those correlates reliably with asset value.

A fund that doubles in value has not doubled its shareholders, its jurisdictions, its filings or its transactions. Frequently it has changed none of them: the same positions are worth more. Meanwhile a small structure with four entities across three territories, monthly dealing and an intricate waterfall can require considerably more administration than a large single-jurisdiction vehicle with quarterly dealing and one share class. Ad valorem pricing bills the second more than the first, which is precisely backwards.

The evidence for the general property

This is not merely an assertion about administration. The regulator's own cost work on the adjacent activity points the same way.

The Financial Conduct Authority's Asset Management Market Study, published as MS15/2.3 in June 2017 after analysis covering more than 20,000 share classes and 30,000 strategies, confirmed weak price competition across a number of areas of the industry. Its profitability analysis found that cost per pound of assets under management falls as assets under management rise — economies of scale — while both absolute profitability and profit margin appear correlated with assets under management. It also found that customers rarely engage with charges at all: on visits to look at funds, fewer than nine per cent looked at charges.

That study concerned management rather than administration, and the distinction should be kept. But the structural point it evidences is general to ad valorem pricing: where unit cost falls as the asset base grows, a fee expressed as a fixed percentage of that base transfers the scale gain to the provider by default. Nothing in the mechanism decides who should receive it. The pricing convention decides, silently, and it decides the same way every time.

Where the convention came from

The basis-point default is not arbitrary, and understanding its origin explains why it persists where it does not fit.

In management, a percentage of assets has a defensible logic: the manager is paid to affect the value of the base, so a fee expressed against that base aligns the two, however imperfectly. Whatever the objections, the variable in the fee is a variable the provider is engaged to influence.

Administration has no equivalent claim. The administrator does not affect asset value and is not engaged to. The fee variable and the service have no causal relationship in either direction, so the alignment argument that justifies ad valorem in the first case is simply absent in the second. What carried across was the unit of account, not the reasoning behind it — and once a convention is the market norm it no longer requires reasoning, because the question a buyer asks becomes how many basis points rather than why basis points.

What each model does to the incentive

The honest treatment is to state what each arrangement rewards, including the one used here, rather than to declare a winner.

Ad valorem rewards growth in the asset base, which the administrator does not cause. It is administratively simple, requires no timekeeping and no scope negotiation, and it is genuinely easy for a client to compute. Its defect is that the fee moves with a variable the work does not track, in both directions — an administrator on this model takes a cut in a falling market for work that has not reduced, which is the same mispricing running the other way and is rarely mentioned by its critics.

Fixed fee rewards precision in scoping and penalises scope creep. It gives the client certainty, which has real value in budgeting. Its defect is that everything turns on where the scope line was drawn: work outside it becomes a negotiation at the moment it is needed, and the incentive on the provider is to define the line narrowly at the outset, when the client is least equipped to judge what falls outside it.

Time-based rewards accuracy about what was actually done. Its defect is the obvious one, and it should be said plainly rather than defended: it requires the client to trust the record. The provider knows how long something took and the client does not, and no amount of granularity in an invoice fully closes that gap. What it offers in exchange is that the fee tracks the work and only the work, and that a period in which little happened costs little.

There is no arrangement without a defect. There are arrangements whose defect is visible to the client and arrangements whose defect is not, and that is the distinction worth making.

Nothing here is advice on any structure or arrangement, and terms should be assessed alongside independent counsel. The reader's test is narrower than a comparison of rates. Ask what the fee is a function of, and whether that variable is one you can observe and verify. A fee model a client can check belongs to a different category from one a client must accept, and the difference between those two categories does not appear anywhere in the headline number.