Insights

Investmentmasteryhub Request a briefing

A multiple-choice item can tell you whether a person recognises a phrase they saw last Tuesday. It is a weak guide to whether that person will notice a missing constraint in a live paper. Investment work inside an organisation is mostly the second problem. The first problem is what vendors of “knowledge checks” still sell, because it is cheap to mark.

Recognition and application part company as soon as the answer is not on the page. A participant can select the correct definition of tracking error and still sign a report that treats a changed benchmark as a performance event. The item did not ask them to do the work they will actually be asked to do. It asked them to point at a sentence.

A case task looks different. You give a short file: a mandate page, a one-line risk appetite, a table with two missing footnotes. You ask the participant to mark the sentence that would not survive a later audit, or to write the question they would put to the manager before the meeting starts. There is still a right-enough answer. It just is not hiding among four options that were written to look similar.

Building that kind of assessment is slower. You have to decide in advance which mistake would actually matter in the role. You have to accept that two competent answers can differ in wording. You have to brief the people who will mark the practicum so that they are scoring the decision path, not elegance. The cost is the point. If the organisation wanted a certificate of attendance, it did not need an assessment at all.

The useful output for the commissioning organisation is not a league table of staff. It is a map of the themes that attracted the same error. If a cohort consistently misses liquidity in a report that looks “fine” on return, the next cycle should spend time there. A ninety-percent score on definitions would have hidden that. The score would have been comforting. Comfort is not the purpose of the file.

We still use a short multiple-choice gate at the start of some modules, as a check that the reading brief was opened. We do not file that gate as evidence of competence. The evidence is the marked case and the practicum memorandum. If a vendor cannot show you those artefacts, they are selling recognition and calling it training.

Educational content. Not a recommendation.

A single risk number is attractive because it ends a conversation. Committees like conversations that end. The difficulty is that the number ended the wrong conversation. It answered “how large is this on the scale we already chose” and left untouched “what did the scale leave out”.

Value-at-Risk is the usual example. It is a statement about a chosen horizon, a chosen confidence, and a chosen window of history. Those choices are the content. The printed figure is the residue. When a dashboard leads with the residue, a non-specialist reader is invited to treat a stable VaR as a stable world. Liquidity can have left the building while the figure barely moved. The silence is in the tail, which the number was designed not to discuss.

Explaining distributions to a non-financial audience does not require a lecture on moments. It requires a picture of a range, a picture of a worse range, and a sentence about what would have to be true for the worse range to arrive. People follow that. They do not follow a Greek letter and a colour. The work is to keep the worse range from being a cartoon, and the base range from being sold as a promise.

Some visualisations help. A set of dated scenarios, each with a one-line mechanism, is more honest than a heat map that turns every risk into a traffic light. A simple table of “what would change our mind” is more honest than a cone that looks like a forecast. A cone is a drawing of uncertainty that most readers will treat as a path. If you draw a path, you are teaching a path, whatever the footnote says.

Visualisations that mislead tend to share a habit: they hide the choice. A single needle on a dial hides the window. A green tile hides the scenario that was never run. A smooth area chart hides the days on which nothing could have been sold at the mark. If a picture cannot show the choice, it should not be the picture a committee sees first.

None of this is an argument against measurement. It is an argument against letting measurement collapse into a mascot. A committee that can say “this is a 99 percent, ten-day figure on a two-year window, and here is the stress we ran instead” is doing the work. A committee that can only say “risk is 4.2” has been trained into a number. That is a literacy failure, even when the number was calculated correctly.

Educational content. Not a recommendation.

A year after a decision, three things have usually happened. The people in the room have changed, or at least their attention has. The market that made the decision feel urgent has been replaced by another. The paper that was tabled has been filed in a way that preserves the recommendation and loses the argument. Reconstruction then becomes archaeology.

A committee that expects to be asked “why did we do that” should be able to recover four elements without calling the original presenter. The assumptions: what had to be true for the preferred option to be acceptable. The alternatives: the options that were on the table, including doing nothing. The limits: the appetite, mandate or liquidity constraint that bounded the choice. The dissent: who objected, on what grounds, and whether the objection was answered or merely recorded.

Those four elements are not minutes as social courtesy. They are the minimum that lets a later reader see a decision rather than a result. A result (“we increased the allocation”) is easy to store. A decision is a path. If the path is missing, the organisation cannot tell a supervisor, an incoming chair, or itself whether the next similar paper is a repeat or a contradiction.

Training helps only if it is aimed at those artefacts. Teaching people the names of asset classes does not produce a better minute. Teaching them to ask “what would change this recommendation” and “what is the alternative we are rejecting” does. The second lesson is shorter and less impressive on a brochure. It is the one that still matters when the pack is late and the room is tired.

Dissent is the element organisations are most tempted to tidy away. A clean minute feels like good governance. It is often the opposite. A later reader who cannot find the objection cannot tell whether the room was aligned or merely quiet. Recording an unanswered concern is not disloyalty. It is how a committee remains a committee rather than a ratification ritual.

If you want a practical test, take a decision from last year and try the reconstruction with the current papers only. If you cannot recover assumptions, alternatives, limits and dissent, the failure is in the file, not in anyone’s memory. The next programme you commission should practise writing those four lines, not reciting a market calendar.

Educational content. Not a recommendation.

Most people who need to read a supervisory text are not lawyers and should not pretend to be. They still have to find the duty that applies to their process, separate it from the recital that explains why the duty exists, and know when the document has stopped being a rule and started being a speech. That is a reading skill. It can be taught without turning the reader into counsel.

Start with structure, not with the first paragraph. European texts usually tell you, in a predictable order, who they are for, which terms they define, which obligations they create, which exceptions they allow, and which earlier texts they repeal or sit beside. If you cannot see those parts, you are not yet reading the instrument. You are reading prose. Skim for the architecture before you highlight a sentence that happens to contain your product name.

Definitions are not decoration. If “client”, “professional” or “financial instrument” is defined, that definition governs every later use, including the uses that feel like ordinary English. The ordinary English meaning is the trap. A practical habit is to copy the defined terms that appear in your process onto a single page and to refuse to paraphrase them in the first pass. Paraphrase is where duties quietly shrink.

Obligations and explanations live in different rooms. Recitals, consultation papers and Q&A documents can be useful, but they are not a substitute for the operative article. A reader who cites a recital as if it were a permission is doing the reverse of diligence. A useful annotation in the margin is “must / may / recites”. If you cannot mark the line, you do not yet know what kind of sentence you have.

Then stop. The point of literacy is to know when to hand the file to someone who is allowed to advise. A learning programme that leaves people confident they have “interpreted the regulation” has failed in the same way as a programme that leaves people confident they have selected a security. The competent outcome is narrower: the reader can find the duty, quote the definition, list the open questions, and write a short brief for counsel or compliance that does not already contain the answer they hoped for.

That is also why we do not teach “regulatory updates” as a stream of news. News decays. A method for opening a text does not. If your organisation’s next cohort needs a horizon scan, that can be a briefing. It should not replace the quieter skill of knowing which sentence is doing the legal work.

Educational content. Not a recommendation.

If one of these notes describes a gap you already see in a committee pack or a training file, the next step is a scoping conversation, not a mailing list. We do not run a public newsletter from this site.

Request a briefing Read the quality framework

Why multiple-choice tests fail investment training

Talking about risk without a single number

What an investment committee should be able to reconstruct

Reading a regulatory text without a legal team

Discuss these themes inside a programme