How to Read a Fund Factsheet Before Placing an Order

The sheet is read in one fixed order, and the order starts with its date: nothing on it can check anything until the moment it describes is known. Then the value per unit, then which plan the holding sits in, then the ratio and its base, then each return with its period and basis, then the scheme's own load terms, then stop.

The situation is a particular one. An instruction is already on the table. Either the holder is about to give one, or one has been given and something in the file does not sit right, and somebody has sent over the scheme's sheet to settle it. The document is not being studied for its own sake. The sheet is being asked for the two or three things it can answer, and asked quickly which questions it was never holding. A sheet read out of order is not read faster, it is read twice. The second reading is the one that discovers the first was standing on nothing.

Every figure below belongs to a single invented scheme, an open ended equity one. The scheme is called the Girnar Large Cap Equity Fund. Standing behind it as asset manager is Girnar Asset Management Limited. Its net assetsWhat the scheme holds once what it owes has been taken off. A ratio quoted against the scheme is measured on this amount. come to Rs 4,200 crore; the count of units in issueThe count of units standing in every holder's name together at the instant the figures were struck. is 120.00 crore. Do that division yourself instead of accepting anybody's answer for it: Rs 4,200 crore shared over 120.00 crore units comes to Rs 35.00 a unit, with nothing left over. There are 3,80,000 foliosThe numbered account against which one holder's units, bank details and standing instructions are filed on the register. on its register. Kalyani Bhagat is the manager of the portfolio; operations sit with Sohail Merchant.

A good deal is already settled before the reading order begins. The striking of a unit's value out of the scheme's own books, one figure standing for one day, is settled elsewhere. What a unit amounts to as a legal interest is settled elsewhere too. Money becoming units, units becoming money, and the rule deciding which day of value an instruction is considered against are all settled elsewhere. A scheme sheet as a document, and what an asset manager puts out and how often, is covered separately. The order below takes all of that as given and does exactly one thing with it: it walks the lines on the sheet in the order a person with an instruction in hand should walk them.

What is the sheet doing when an instruction is already in hand?

The sheet is narrowing the search, and narrowing is a smaller and more useful job than it sounds. Somebody with an instruction in front of them has a handful of live questions, and the sheet can settle perhaps three of them outright, can partly settle two more, and holds nothing whatsoever on the rest. Knowing which is which before the reading starts is what turns a five minute read into a thirty second one. The reading order below exists because each line checked makes the next line usable, and skipping ahead only means arriving at a number that cannot yet be interpreted.

There is a tariff card taped to the counter of a courier office. The card states what the office charges and on what basis it charges it. The card does not state where any particular parcel is this morning. The card was never a tracking record, everybody knows it, and nobody stands there arguing with it about that. A scheme sheet sits in the same relation to an instruction. The sheet describes the pool and the terms; the holder's own record describes what was done. Reading one as though it were the other is where almost all of the wasted time on this goes.

So the order runs like this, and every step after the first is standing on the one before it. First comes the moment the sheet belongs to. Then the value per unit, with a check that the moment sitting beside it is the moment the question needs. Then a settling of which planOne of the versions the same scheme is offered in. The holdings underneath are the same; the running charge is not, and so the value per unit is not either. the figures describe. Then the expense ratio, with its base set down beside it in plain words. Then each return figure in turn, none of them read until the stretch it covers and the basis it stands on are both noted next to it. The benchmark figure takes those same two notes, plus the one this sheet leaves out. Then the load terms, read as terms this scheme wrote for itself. Then the sheet's silences are named, and the sheet is closed.

Why does the moment the sheet belongs to have to be found first?

Because everything printed under it is a reading, and a reading only exists as the reading of one particular instant. The as atThe single instant every figure on a sheet is taken to belong to. Two figures drawn from instants that are not the same instant cannot be set against one another. line is where the sheet names which instant that was, and it is usually printed small, near a heading, in the sort of place the eye skates over on its way to the numbers. The record carries a date, and a figure whose moment cannot be named carries nothing. So that figure cannot be set against any record at all.

The same habit runs without thinking in other parts of life. Nobody argues with an electricity bill about the amount before finding the meter reading it was raised on and the day that reading was taken. If the reading turns out to be from before the tenant moved in, the argument about the amount stops instantly. The amount was never the problem. The moment was. The numbers on a scheme sheet look self-contained in a way a meter reading does not, so a scheme sheet works the same way and gets treated worse.

There is a second reason to do this first rather than last, and it is about time. If the moment on the sheet turns out to be a different one from the moment the question is about, the reading is finished in fifteen seconds and the search moves to the right record. Reading the whole sheet first and reaching the moment at the end costs five minutes of annotating figures that were never going to help. The step that costs least comes first.

Is the value per unit on the sheet the one the instruction will use?

Almost certainly not, and that is the ordinary case rather than a fault in anything. The figure on the sheet belongs to the sheet's own moment. The instruction attached to a day of its own, settled by conditions that are not printed on the sheet and are covered separately. Two correct figures that answer two different questions will differ, and treating that difference as a discrepancy is how a query gets raised about something nobody did wrong.

The sheet's value per unit is genuinely good for one thing: checking the pool it came out of. The two rows above it carry the division. Rs 4,200 crore of net assets set against 120.00 crore units in issue yields Rs 35.00 for one unit, and on this specimen the division leaves no remainder at all. How that figure gets struck out of the scheme's books, and which day of it an instruction is considered against, are both covered separately. The step confirms that the three rows agree with each other.

One step further, checking an amount against it, turns up a residue in the arithmetic, and the residue has a sign. Suppose Rs 1,00,000/- against a value per unit of Rs 35.00. The division is 20000 over 7, or 2,857.142857 and then never stops. Units are carried to a fixed number of decimals and the remainder is rounded up, so what appears is 2,857.143. The rounding added exactly one seven thousandth of a unit. At Rs 35.00 that is exactly half a paisa, so the units allotted are worth Rs 1,00,000.005 and sit exactly halfway between two paise.

Say the awkward part out loud rather than tidying it away: that residue does not cancel. Rounding upward is one sided, so the sign is positive every time and there is no offsetting negative anywhere to net it off against. Half a paisa on a lakh is nothing anybody would chase. But a reconciliation that expects an exact match and finds five thousandths of a rupee has not found an error, and a reconciliation that quietly rounds it away has stopped being a reconciliation. The residue is written down with its sign, and the work moves on.

Which plan do the figures on the sheet actually belong to?

By reading the label rather than the number. The step sounds too obvious to count as one, and almost everybody skips it. The same scheme is offered in more than one version. The holdings underneath are identical, the portfolio is identical, Kalyani Bhagat manages the identical thing either way, and the running charge is not identical, so the value per unit is not identical either. On the specimen the two ratios read 1.65 per cent of net assets a year, then 0.85 per cent on that same base. The two sit 0.80 percentage points apart. The check takes a second. Unpicking what follows from skipping it takes an afternoon, and that is precisely the trade that keeps making people skip it.

Two things conspire here and neither of them is carelessness. The two figures belong together, so they are printed close together. And a document setting out one scheme has no room to keep explaining itself, so the labels are brief and set in small text. Meanwhile the person reading has an instruction in front of them and is hunting for a number, not for a label, so the eye goes to the digits and the label never registers. The same thing happens at a canteen counter where two plates look identical and the board above them carries two prices. Nobody misreads the board. The plates settled the question first, so nobody looks at the board at all.

There is a limit to what the specimen can show here, and it is worth naming rather than papering over. The plans carry different charges, so a value per unit belongs to a plan rather than to the scheme as a whole. The record behind this specimen fixes net assets and units in issue for the scheme as one pool and does not split either of them by plan, so a value per unit for each plan separately cannot be worked out from it. The specimen prints that row as a labelled refusal rather than a number, and names the two inputs that are missing: net assets by plan, and units in issue by plan.

What base is the expense ratio standing on, and where has it already gone?

Find the expense ratioThe running charge of a scheme for a year, quoted as a share of its net assets rather than as an amount of money. and, before it is used for anything, its base belongs beside it in plain words: a share of the scheme's net assets, running for a year. Not a share of the amount put in. Not a share of what the holding gained. Not a one off. The sheet states that base once, briefly, usually in a heading or a bracket, and it is the first thing to fall out of a reader's head. The same digits attached to a different base describe a completely different amount of money, so a ratio with no base written next to it is not yet a number.

Once the base is written down, the sheet's own rows do the rest of the work. On the specimen, taking 1.65 per cent of the Rs 4,200 crore of net assets leaves Rs 69.30 crore for the year, or Rs 69,30,00,000/- written out in whole rupees. Spread across a full year of days that is about Rs 18,98,630/- for one day, and shared over 120.00 crore units it is about Rs 0.001582 a unit for one day. The nature of the charge, the working roles it pays for, and the effect of a gap that size over a run of years are each covered separately. The step itself is mechanical: name the base, then let the sheet's rows produce the amount.

Now the trap that catches careful people. Having reached about Rs 0.001582 a unit for a day one way, the natural next move is to reach it a second way as a check: take the ratio, divide it by the days in a year to get a daily rate, and apply that rate to the value per unit of Rs 35.00. The same figure comes out. The second route is the first route rearranged and cannot disagree, so it proves nothing about the sheet. The first route computes the ratio times net assets, divided by the days in a year, divided by however many units are in issue. The second computes the ratio divided by the days in a year, times net assets over units outstanding. Since the value per unit is exactly net assets divided by units in issue, both are the same four quantities multiplied and divided in a different order, and multiplication does not care about the order. Getting the same answer twice there shows only that the arithmetic was done carefully. About whether the sheet in hand is the right sheet, it is silent.

There is a check that can genuinely fail, and it is worth more than both of those put together because it draws on a row the sheet never wrote. Take the unit balance for the holder's own plan, multiply it by the sheet's value per unit for that plan at the sheet's moment, and set the answer against the statement's valuation for the same moment. The unit balance came from the register, not from anything on the sheet, so the two sides are genuinely independent and the comparison has something at stake. When it disagrees, one of exactly three things is wrong: the plan is the wrong plan, the moment is the wrong moment, or the balance is not the balance it was taken to be. Three possibilities is a short list, and having a short list is the entire benefit of doing the check in the first place.

What has to be written beside a return before it can be used?

Two notes, and until both are written the figure stays unread. How long a stretch does it cover. And on what basis was it worked out, meaning had the running charge already been taken out before anybody computed it. On the specimen the scheme's figure is 13.4 per cent, over one year, and it is a net returnA figure computed from values the charge had already run against, so nothing more is deducted from it later on.. One year, one scheme, and that pairing is repeated every time the figure appears rather than letting it float free. A return with no period and no basis beside it is not a small figure or a large one. There is nothing yet for it to be small or large in comparison with.

The basis is the note people leave out, and it matters more than the period because the period at least tends to be printed. A scheme's published figure is computed from values the charge had already run against, day after day, before any of them was used. Nothing further comes off it afterwards. So in the 13.4 per cent on this specimen, the running charge of that plan is already inside the figure, and subtracting the ratio from the return a second time would be counting the same charge twice. The arithmetic above is the mechanical half of the job. Reading such a figure without letting it turn into an expectation about what comes next is covered under reading a return figure.

Then comes the row underneath, and it is where the sheet quietly stops helping. The benchmark returnThe figure for the measuring stick a scheme names for itself. Nobody can hold a measuring stick, so nothing is ever taken out of its figure. on the specimen is 12.1 per cent over the same one year. Both figures carry the same period, and both are printed in the same size of type. One of them had a charge running against it the whole way. The other never had anything taken out of it at all. Nobody holds an index, and nobody pays anything to hold one. The 1.3 points between them is therefore a difference between two different kinds of number, and nothing printed on the sheet points that out.

One more note belongs on the benchmark row and this platform's record cannot supply it, so the honest thing is to record the gap rather than fill it. A scheme's figure includes the income the scheme received. A measuring stick can be built from prices by themselves, or built with income added back, and the two builds of one stick do not land on the same figure. The invented record does not say which of the two builds the specimen's stated benchmark uses. So the check stays open: write it down as unresolved, and settle it from the scheme's own documents rather than assuming the convenient answer.

Whose terms are the load terms on the sheet?

The terms belong to this one scheme and to nothing wider. The load termsThe conditions a scheme sets out in its own documents for a charge applied when money leaves it. printed on a sheet are conditions the scheme wrote into its own documents, and a different scheme run by the same asset manager may have written something else. Reading them as a rule of the market rather than as the terms of one scheme is what makes a reader carry an assumption from one holding to the next, where it is quietly wrong.

The everyday version is a shop's own returns counter. Every shop has a policy, the policies differ from shop to shop, and there is a body of law sitting above all of them setting out what a shop may and may not do. Nobody confuses the two. The notice at the counter gives this shop's policy, and the question whether the shop is allowed to have that policy at all is settled somewhere else entirely. A scheme's load terms sit in the same relation to what the Securities and Exchange Board of India (SEBI) settles above them.

On the specimen this row is a labelled blank, and that is deliberate. The invented record fixes no rate and no length of holding for the scheme, so there is nothing to print. Any figure appearing in that box would have been made up on the spot. A blank cannot be quoted back at anybody and a made up figure can, which is what makes the made up figure worse. Whether a scheme may set such terms at all, and what limits sit above them, is settled by SEBI, and sebi.gov.in carries it, whenever the question happens to be live. What an exit charge does to money leaving a scheme is covered under exit loads.

What was this sheet never going to answer?

Two questions arrive with almost everyone holding an instruction, and the sheet holds neither of them. Which day of value the instruction is considered against. And the smallest amount a scheme may accept at the outset. Both are settled by SEBI rather than by the asset manager, and both are read at sebi.gov.in, where the current position stands. Conditions like those move, and a figure copied out of them is wrong from the morning it changes. Learning that a document does not hold the question is a result rather than a dead end, and it is the cheapest result available anywhere in this reading.

The situation is the same as standing at a chemist's counter reading the price list to find out whether a medicine needs a prescription. The price list is accurate, complete and entirely unhelpful. Whether a medicine needs a prescription was never among the things a price list records. Four readings will not make it start holding the answer. The useful move is not to read harder; it is to recognise the category of the question and go to where questions of that category are actually settled.

Naming what the sheet does hold keeps the line clear rather than vague. The sheet holds the moment, the pool, the value per unit, the labels of the plans, the running charge of each and its base, the returns with their periods, the measuring stick, and the scheme's own terms about money leaving it. The list covers a real amount of ground and settles most of what a person with an instruction actually needs. The sheet does not hold anything decided above the scheme rather than inside it. Once a question can be sorted into those two piles in a second, the sheet stops being a document to search and becomes a document to read.

What does the whole reading look like run over one specimen sheet?

Here is a specimen extract assembled out of the invented record, with every row the record left empty shown as empty. Look at how briefly the labels are written. That brevity is the whole reason the plan step exists. Then run the eight steps across it. What makes the procedure work is the refusal to move to the next row until the current one is settled, and that is worth more demonstrated than described.

Now the walk, one row at a time, with what each step settles and what it deliberately leaves alone. Notice how often the useful outcome of a step is a boundary rather than a number.

Two things are worth saying about that walk before leaving it. The first is that three of the eight steps ended in a refusal rather than a figure, and that is a fair reflection of what reading a sheet against an instruction is actually like. The second is that the specimen's own rows check out against each other: Rs 4,200 crore set against 120.00 crore units lands on Rs 35.00 without a remainder, and taking 1.65 per cent of that same Rs 4,200 crore gives Rs 69.30 crore. Both are internal agreements, and an internal agreement cannot fail. The check that can fail is still the one that crosses out of the sheet and into the holder's own record, and it is the only one here with anything at stake.

Girnar Asset Management Limited, the Girnar Large Cap Equity Fund, Kalyani Bhagat and Sohail Merchant are invented.
Educational material. Not advice on any investment, tax, budget or market position.