Who Governs the Managers

There are two distinct questions in the management of deployed capital, and almost all attention goes to the first. What should be held is the first. Who decides what should be held, on what mandate, measured against what, and with what consequence for missing it — that is the second, and in most arrangements it has no owner at all.

The second question is a governance question rather than an investment one. It does not require a view on any asset. It requires a structure, and the structure is largely absent from the market because the parties best placed to build it are the parties it exists to constrain.

What a governance layer actually contains

Most readers have never seen one specified, so it is worth setting out rather than gesturing at.

It begins with the development of an investment strategy — not a portfolio, a statement of objectives and tolerances against which any portfolio can later be judged. It includes the selection of the external managers or commission banks that will execute it, which is a procurement decision with defined criteria rather than a relationship inherited from a previous decade. It establishes performance benchmarks and compliance requirements, so that a manager's results are assessed against something specified in advance rather than against a narrative supplied afterwards. It provides ongoing monitoring against that stated strategy and those benchmarks. It produces reporting. And it attaches accountability mechanisms with consequences, which is the component most often written down and least often exercised.

Notice what is not in that list. The governance body does not pick assets. It decides who picks assets, on what terms, measured how, and what happens when the mandate is missed. Those are different jobs requiring different information, and the persistent error is to treat the second as a subordinate function of the first — as something the manager can reasonably be asked to supply about itself.

Why the conflict is structural

The usual objection is that a competent, honest manager will report candidly and act in the client's interest, which is often true and beside the point. The problem is not that firms selling products are dishonest. It is that a firm which sells a product will find reasons the product fits, and it will find them sincerely, because the alternatives it knows best are its own and the questions it is practised at answering are the ones its products answer well.

Disclosure does not repair that. The regulatory record is the useful evidence here, because disclosure was tried first and then superseded. Under MiFID II a firm must tell a client whether its advice is provided on an independent basis. But where advice is independent, the Directive does not stop at telling: Article 24(7) requires the firm to assess a sufficient range of different product providers' products, with the range not restricted to instruments issued by entities having close links or other close legal or economic relationships with the firm — and it prohibits the firm from accepting and retaining fees, commissions or any monetary or non-monetary benefits from third parties, requiring that such payments be returned to the client in full as soon as possible, with only minor non-monetary benefits permitted and only where disclosed.

Read the shape of that remedy. The legislature did not conclude that inducements were acceptable provided they were declared. It concluded that an incentive survives its own disclosure, and it removed the incentive. That is a structural judgment about how advice degrades, arrived at by people with access to a great deal of evidence about it, and it is the same judgment underneath a governance layer held separately from execution.

The same logic applies to how the governance layer itself is paid. A fee expressed as a percentage of assets is an incentive with a direction: it rewards the accumulation and retention of assets under the arrangement, which is not always what a principal needs and is occasionally the opposite. Charging instead on the extent of the administration performed — as this firm does, rather than on the value of the assets in the fund — removes that particular gradient. It does not make the adviser virtuous. It removes one specific reason to be wrong.

Selection is a procurement decision

Manager selection is where the absence of a governance layer shows first, because in its absence the decision is made by adjacency. The manager selected is the one already banking the family, or the one introduced by a trusted party, or the one appointed by a predecessor for reasons nobody now present can reconstruct. Each of those is a real signal and none is a criterion.

A procurement decision looks different. It states in advance what the mandate is for, which characteristics are required and which are merely desirable, what would constitute evidence of each, and how many providers will be assessed against that specification before an appointment is made. It also records why the unsuccessful candidates were unsuccessful, which is the document that matters three years later when performance disappoints and the question becomes whether the original reasoning still holds or was never written down.

The discipline is not that adjacency is disqualifying. It is that adjacency should have to compete.

The consequence mechanism

The part of a governance framework that determines whether it is real is the last one, and it is the part most frameworks leave undefined.

Monitoring without consequence produces a reporting pack. Everyone reads it, everyone notes the variance, and nothing follows, because nothing was specified to follow. A mandate that can be missed indefinitely is not a mandate; it is a description. What makes the difference is deciding in advance — while relations are cordial and no one has underperformed — what a breach of the stated strategy triggers, at what threshold, on whose determination, and over what period. Those decisions are easy to make in year one and nearly impossible to make in the quarter when they are needed, which is the whole argument for making them early.

The corollary is uncomfortable and worth stating: a governance body that has never exercised a consequence has not been tested, and its principal does not yet know whether it works.

None of this is investment advice, and nothing here recommends any structure, manager, benchmark or asset. Arrangements of this kind are designed alongside independent counsel in each relevant territory, and a principal considering one should take advice from an adviser unrelated to this firm. The argument is narrower than it may appear, and it is not about performance: it is that the question of who decides has an answer in every arrangement, that the answer is frequently nobody, and that nobody is a decision which was made by default rather than on purpose.