How to Review What a Scheme Discloses About Governance

The review runs in a fixed order. The first step settles that the object is accountability rather than a verdict. Then the parties are sorted by role, each one located in what the scheme publishes, the categories that touch the holdings read, every date and issue number checked, and the document set against the one before it. The Securities and Exchange Board of India (SEBI) sets what must be published and how, at sebi.gov.in.

The order is the whole method, and the order is not decorative. Each step exists because the step after it becomes unreadable without it. A role that has not been named cannot be located. A paragraph about how holdings are priced cannot be weighed until the party that wrote it and the party that checks it are both known. And two documents cannot be compared until the top of each one has been read and the date found. A governance review is a reading order before it is a reading, and most of what goes wrong goes wrong in the first step rather than the last.

One scheme carries every example that follows. The Girnar Large Cap Equity Fund, an invented open ended equity scheme, is run by Girnar Asset Management Limited; its net assets stand at Rs 4,200 crore, the units in issue at 120.00 crore, the separate holder accounts behind them at 3,80,000, and the expense ratio at 1.65 per cent of net assets a year. The portfolio sits with Kalyani Bhagat as fund manager; operations sit with Sohail Merchant.

Three matters are settled elsewhere and no part of any of them is redone below. The vehicle itself, and what each party inside it does, belongs to how a scheme is assembled. The running charge, and the striking of a value per unit each day out of the books of the scheme, belongs to pricing. The contents of each published document belong to scheme documents. All three are taken as read. The order below is the one in which a person holding units can actually read the accountability out of what is on the table.

What is a governance review actually for?

The purpose of the review is settled before anything is opened. Most readers arrive wanting an output the material does not produce. A governance disclosureAnything a scheme puts on the record about who carries which duty inside it, and how that duty was carried out. answers two questions and only two: who is accountable to whom, and what has been put on the record about it. A governance disclosure does not answer whether a scheme is well run. No document published by the scheme, by the manager, by an industry body or by a regulator answers that, and hunting for one turns a short exercise into a long and unsatisfying one.

The same thing in everyday terms. A school puts a notice board up in the corridor listing who is responsible for the bus route, who runs the canteen, who sits at the fees counter and who to write to if any of the three goes wrong. A careful reading of that board settles exactly who to ask about the bus. The board settles nothing about whether the school is a good one. The board was never built to say, and a parent who stands in front of it waiting to find out will stand there a long time. Starting with the right question is what stops the whole exercise turning into a search for reassurance.

So the output is fixed first. When the last document is closed, two short lists should be writable: who answers for what, and what has changed since the previous issue of each document. Everything in the steps below exists to produce those two lists. Anything else the reading leaves behind, particularly a feeling about the manager, came from the reader rather than from the documents.

Which parties are involved, and what is each one answerable for?

The parties are sorted by role before a word of content is read. Each role answers for something different, and the differences are the point. The asset manager runs the scheme from one day to the next. The trustee companyThe party that holds what a scheme has, on behalf of the people holding its units rather than on behalf of the manager. holds the assets of the scheme in trust for the people holding units. Holding them in trust is what turns its accountability towards those holders rather than towards the manager. The custodianThe party that keeps safe the securities a scheme has bought, and settles what the scheme trades. keeps the securities the scheme has bought. The registrar and transfer agentThe party that keeps the record of who holds how many units, and processes what those holders instruct. maintains the record of who holds how many units. The auditorThe party that examines the accounts of a scheme and reports on what it found there. examines the accounts and reports on them.

The shape of that list is worth noticing. Five parties, five different duties, and only one of the five is the party whose name is on the scheme. The imbalance matters. The manager is the party a reader can name without opening anything, and the manager is also the party whose account of itself carries the least information, for the same reason that a shopkeeper is not the person to ask whether the shop is honest. A reader who can only name the asset manager has identified the party whose account of itself is the least informative one available to them.

Two habits make this step fast. The roles go down in a column before a single document is opened. The reading is then a search for something rather than an intake of whatever turns up. And against each role goes the direction its accountability runs in. One word of direction separates the trustee company from every other party on the list.

Where does each role appear in what the scheme publishes?

Go looking for each one by name of role, and treat the search itself as part of the reading. A role that appears repeatedly, in documents of different kinds, with a duty described the same way each time, has been placed on the record properly. A role that appears once, in passing, in a sentence that does not say what it answers for, has been mentioned rather than disclosed. The difference is visible from the outside and it costs nothing to check.

Now the case that readers handle worst. Everything the scheme publishes has been worked through and one of the five roles appears nowhere at all. The instinct is to assume the search went to the wrong place, close it and move on. The instinct is wrong. The search has just produced the most concrete result of the whole exercise. A role that cannot be located in anything published is a finding rather than a failure of the reading, and the response is to write down which role and which documents were searched, and then to ask.

Back to everyday terms. A repair bill arrives with an amount, a date and no name of anybody who did the work. Nothing obliges the householder to conclude that nobody did the work, and nothing obliges the conclusion that the bill was misread. The bill does not name the person. Failing to name the person is a fact about the bill, and the next move is to ask. Documents are allowed to be incomplete, and noticing is not an accusation.

Which categories of disclosure touch how the holdings are handled?

Three of them, named here as categories without what has to be inside any of them. The first is the valuation policyThe written method a scheme follows to put a price on each thing it holds, applied the same way each time., the scheme stating in writing how it prices what it holds. The second is how the votes attaching to those holdings are exercised, the scheme stating what it did with the say that comes with the shares in the pool. The third is connected party dealingA dealing done with somebody related to the scheme or to its manager, which has to be shown on the record as such., the scheme reporting dealings done with parties related to it or to its manager.

The three categories are worth separating from everything else. Each is the point at which an arrangement about people becomes an effect on the pool. A pricing method decides what a unit is worth. A vote decides what the scheme did with the influence it holds. A dealing with a connected party is the classic place where an interest could sit on both sides of one transaction. Sitting on both sides is exactly why such a dealing is reported as a category of its own rather than folded into everything else.

SEBI sets what must appear in each one, how often it appears and in what form, and the position in force is read at sebi.gov.in. Requirements of that kind get revised, so the position that governs is the one in force at sebi.gov.in on the day it is needed.

Why does a document without a date and an issue number tell so little?

Because a disclosure is a statement about a moment, and a statement about a moment that does not say which moment is barely a statement at all. The top of the document is read before the body of it. A date is wanted there, along with the versionWhich issue of a repeating document is in hand, without which nothing in it can be set against anything. marker that identifies which issue of a repeating document is in hand. Neither is a formality. Both are what make the next step possible.

A price list taped to a shop wall with no date on it is the everyday case. Every line of it can be read and it still settles nothing about whether the price of rice went up. Only the list that was on the wall before would settle that. The paper is perfectly legible and completely uninformative about movement, and no amount of careful reading fixes that. A document carrying neither a date nor an issue number can be set against nothing, and the most informative reading available is closed off before it starts.

So the discipline is small and strict: date and issue number first, content second. Both go down next to the name of the document. If either is missing, that goes on the list of things to ask about, alongside any role that could not be located.

Why does a change tell more than a level?

The comparison step produces the actual information, and the comparison step is the reason the previous one insisted on the date. Set this issue of a disclosure against the one before it and read the difference. Read what appeared that was not there before, what went, and what stayed but was reworded. The three readings are the product of the whole exercise, and none of them exists in either document on its own.

The distinction underneath the step is worth naming plainly. The step turns on the difference between a levelWhat a document states as it stands today, read on its own without anything to set it against. and a change. A level states what the scheme discloses today. Every scheme discloses something in every category, so the mere presence of a paragraph separates nothing from anything and carries almost no information. A change states what moved, and something moves because somebody decided it should. A level is what a document says and a change is what somebody did, and only the second is a thing a reader can sensibly ask about.

Two practical notes. The comparison is slow and unrewarding right up until the moment something has moved, and that moment is exactly what the comparison exists to catch, so the unrewarding stretches are not wasted. And a change is not by itself good or bad. A change is a question, and the honest output of this step is a list of questions rather than a list of conclusions.

What can a disclosure never say?

The limits of a disclosure are worth naming before the reading finishes. Naming them stops a reader hunting for something that is structurally not there. A disclosure records what was done and reported. A disclosure does not record what was decided, or why it was decided that way, or what was considered and turned down. Those decisions live in rooms nobody publishes. And no disclosure is a grade, a score, a rating or a ranking of any kind, in any category, on any scale.

A disclosure existing is evidence that something was reported, and it is not evidence that what was reported was sound. The gap is easy to lose sight of. The documents are formal, they are dated, they are signed, and they carry every visible feature of assurance without being it. Nothing inside them says so. A reader has to say it to themselves.

There is a second and larger limit, and it is the one most worth stating in full. Reading every governance disclosure a scheme publishes, and reading them well, says nothing about what a holding will be worth. Oversight of how a scheme is run is not a check on what the holdings do, and the two questions do not touch. A scheme can have every role placed clearly on the record, every category disclosed on time and every change explained, and the value of a unit can still fall. The value of a unit moves with what the holdings are worth, and with nothing that appears in these documents at all. A governance review identifies who to ask. The review says nothing about what will happen to the money, and it does not protect the money.

Where do the content, the frequency and the format get decided?

Not by the scheme. SEBI decides what must be disclosed in each of the categories above, how often each disclosure has to appear, and in what form it has to be presented. All of that sits at sebi.gov.in, to be checked there whenever it matters. The Association of Mutual Funds in India (AMFI) sets out practice across the business as a whole at amfiindia.com, a place to read rather than a body that makes any of these rules. Such practice describes how things tend to be presented rather than what any scheme must do. The documents of the scheme itself then restate what applies to it, and that makes them the right place to start and the wrong place to stop.

Two further routes exist and are worth knowing about before they are needed. There is a route by which a person holding units raises a complaint, and there is a route by which an unresolved complaint is escalated beyond the manager. SEBI publishes what they are and how they work. Both routes exist, and the current description of each one is published at sebi.gov.in.

Everything in this area is revisable, so the description that governs is the one published at sebi.gov.in on the day it is needed.

What does the review look like run on one invented scheme?

Run the seven steps on Girnar Asset Management Limited and the Girnar Large Cap Equity Fund, and let the gaps in the record do half the teaching. Start with the roles, and notice at once what has been done deliberately in this record: it names Girnar Asset Management Limited, it names Kalyani Bhagat who manages the equity scheme, and it names Sohail Merchant who heads operations. The record gives no name at all to the trustee company, to the custodian, to the registrar and transfer agent, to the auditor or to the distributor. Every one of those is referred to by role and nothing else.

The silence is the first lesson of the review rather than a hole in the example. The named parties here are the ones whose names tell a reader least, and the unnamed ones are the ones whose duties tell a reader most. A governance reading has to invert exactly that shape. A review that consists of recognising the name of the manager has read the least informative line on the table.

Now anchor the one place a governance question turns into a rupee figure. Against the Rs 4,200 crore of net assets the equity scheme carries, there are 120.00 crore units in issue, so one unit divides out at Rs 35.00 exactly. The net assets sit behind 3,80,000 foliosThe account number under which the units of one holder are recorded and their instructions processed., so about Rs 1,10,526/- stands behind the average one. The expense ratio is 1.65 per cent of net assets a year, and 1.65 per cent of Rs 4,200 crore is Rs 69.30 crore for the year, or Rs 69,30,00,000/- written out in full.

The figure is worth stating plainly for what it is and what it is not. The charge is the size of the arrangement over a year: what the whole set of parties, duties and reporting costs the pool. No disclosure resolves whether the arrangement is worth that, and the charge is not a verdict on it. How the same charge is spread across the days in a year, and what it comes to on a single day, is covered under pricing.

Check A and Check B are not two checks of equal worth, and saying so is the honest thing to do: Check A is row One and row Two rearranged, so it draws on the same two figures and cannot fail. Multiplying a unit value by a rate and then by the units in issue, or striking the rate on the total and then dividing by the units in issue, is one equation written two ways. Check A confirms the arithmetic was typed correctly and confirms nothing else.

Check B can fail, and that is why it is on the table. Check B goes round by the folio count, a row no governance disclosure ever writes, and lands Rs 1,980/- short of the direct figure. The shortfall is not an error to be tidied away. The shortfall is exactly 1.65 per cent of the Rs 1,20,000/- of net assets that the rounding of the average folio to whole rupees left over, and it reconciles to the paisa. Note also that the rounding here is one sided and does not cancel. Rounding the average folio down leaves Rs 1,20,000/- unallocated. Rounding it up to Rs 1,10,527/- would overshoot the net assets by Rs 2,60,000/-, more than twice as far out in the other direction.

Which readings does this record simply not support?

A method is only honest when its limits are named alongside its steps. Some steps above can be run on this record and some cannot, and a reader shown only the runnable half learns a method that quietly assumes complete information.

Three steps can be run: the roles can be listed, the direction each one answers in can be written down, and the parties this record names can be separated from the parties it refuses to name. The rupee anchor can be computed and checked twice. Three others cannot be run. There are no portfolio holdings here and no weights, so nothing about how the pool is composed can be read; there is no commission figure for any distributor, so nothing about that arrangement can be read; there is no assets figure for the manager across its whole range, so nothing about scale can be read.

And the largest absence is the one that lands on step six. One moment is fixed here for one scheme, so no previous issue of any disclosure exists to set the current one against. The comparison step is therefore described as a method and never run. The whole force of step six is that the second document has to be real.

What does a finished review actually produce?

Two lists and a set of questions, written down rather than remembered. The checklist below is the shape of the output. Every period, format and figure of the kind it might carry belongs to SEBI at sebi.gov.in; the checklist is simply the set of things written down as the reading goes, and the column on the right is the guard against the reading drifting into a verdict.

Notice what is not a column here. There is no score, no weighting, no total and no scale, and none of those was left out for space. A checklist that added them up would manufacture exactly the verdict the documents cannot support, and a number invented at the end of an honest reading is worse than no number. The invented number looks like a result.

The Girnar Large Cap Equity Fund, Girnar Asset Management Limited, Kalyani Bhagat and Sohail Merchant are invented, along with the net assets, the units in issue, the folio count and the expense ratio.
Educational material. Not advice on any investment, tax, budget or market position.