The Investment Policy: The Fence Before Any Choice

An investment policy is the written statement, fixed before any money arrives, of what a scheme is permitted to hold and what it is not. Three sets of limits apply at once: what the regulator allows, what the scheme document promises the people holding units, and what the manager's employer imposes internally. Whichever of the three is narrowest is the one that binds. The policy bounds the possible and chooses nothing.

Here is the shape of the thing before any of the detail. A scheme is a pool of money with a written contract wrapped around it, and one clause of that contract is a list of what the money may be put into and what it may not. The clause is settled and published before a single rupee arrives from anybody. The timing is what makes the clause a fence rather than a plan, and the gap between those two words matters more than anything else about it.

One scheme supplies every example that follows. Girnar Asset Management Limited, an invented asset manager, runs the Girnar Large Cap Equity Fund, an open ended equity scheme with net assets of Rs 4,200 crore and 120.00 crore units in issue, so the value of one unit divides out to Rs 35.00 exactly. Kalyani Bhagat manages that scheme and Sohail Merchant heads operations at Girnar Asset Management.

Now the honest starting point. Every proportion below is worked from the pool figure alone and illustrates a mechanism; not one of them is a holding of the Girnar Large Cap Equity Fund. The mechanism runs on ratios and written conditions rather than on any particular number, so what a scheme may hold can be set out in full without a single holding or limit value in front of you.

What is a scheme's investment policy, and where does it actually sit?

The policy sits inside the scheme document, in writing, under a heading of its own, and it was there before the scheme collected anything from anybody. An investment policyThe written statement, carried in a scheme's own document, of what that scheme may and may not hold. is a statement of permission and prohibition: these kinds of holding are open to this scheme, those are shut to it, and here are the conditions attached to the open ones. A written policy is not a forecast, not an intention and not a description of anything. It draws a set of edges.

Think about a field with a boundary wall around it. The wall settles where a person may walk and where they may not. The wall settles nothing whatever about where to plant, what to plant, or how much of the field to leave bare this season. Somebody standing at the gate can see the wall perfectly and still have no idea what the field contains. The wall stands to the crop exactly as a scheme's written policy stands to a scheme's actual holdings, and almost every misreading in this area comes from treating the wall as though it were a description of the crop.

The rental agreement on a flat behaves the same way. The agreement says the tenant may not knock down a wall, may not sublet and may not run a workshop out of the kitchen. The agreement never says where the sofa goes. A reader who wanted to know how the flat is arranged and read the agreement instead would come away knowing a great deal about what cannot happen and nothing at all about what has.

Three consequences follow from the policy being written down and settled early, and they matter more than they look. The promise was made to the people who bought units and not by Kalyani Bhagat personally, so the first is that she is bound by a document she may have had no hand in drafting. The second is that the policy is dated: it says what it says on the day it is read, and it can be altered, though altering it is not a private decision and there is a required process that exists precisely to protect the people who bought on the old terms. The third is that the policy can be read by anybody, before buying. A prospective holder therefore has one document setting out what is capable of happening to their money. The written policy is the only part of a scheme that sets out the full range of what may happen. Read it before the performance figures rather than after them.

Where do a scheme's limits actually come from?

From three places at once, and they are not alternatives to one another. The first is the regulator, setting what any scheme of a given kind is allowed to do. The second is the scheme document itself, promising the people who buy units something narrower than the regulator would tolerate. The third is the asset manager, adding internal limits of its own on top, for reasons of its own. All three apply to the Girnar Large Cap Equity Fund on the same morning, and none of them cancels another.

The regulator's layer is wide because of what it is for. A regulator's rule has to cover every scheme of that kind run by every asset manager in the market, and a rule written for all of them cannot be tailored to any one of them. So it sets outer edges that look startlingly permissive to anybody reading them as a description of one scheme. Reading them that way is a category error, in the same way that the speed limit on a highway settles nothing about how fast a particular school bus is driven.

The scheme document's layer is narrower, and it is narrower for a reason that is easy to state and easy to forget: that narrowness is what the holders bought. Somebody who put Rs 1,00,000/- into an equity scheme was buying the promise that the money would be run as an equity scheme, and the written policy is where that promise is recorded in enforceable words rather than in marketing ones. If the document says a thing will not be held, the scheme has given something up in order to be sellable, and the giving up is the product.

The asset manager's layer is the one most readers have never heard of, and it is often the tightest of the three. Girnar Asset Management may decide that no scheme it runs will go anywhere near the outer edge of what its own scheme documents allow, for reasons that are commercial, operational or simply cautious. Nothing obliges the asset manager to publish that layer in the same way as the other two. A reader who reasons only from public documents can therefore be surprised by how far inside its fence a scheme actually sits. Three sets of limits are live at once, they were written by three different parties for three different purposes, and a reader who has looked at only one of them has looked at the least useful one.

Which of the three layers actually binds the manager?

The narrowest one, always, and nothing else about the three matters as much as that. The binding layerWhichever of the three sets is tightest, because that is the one a manager meets before any of the others. is whichever of the three a manager would run into first if she kept pushing in that direction. Something tighter stopped the move earlier, so the other two never get tested. Both remain real and enforceable all the same. Take the school bus again. The road allows one speed, the school allows a lower one, and the driver has his own rule for the stretch past the market. Only the last of those three ever gets tested, and for predicting how fast the bus goes past the market, the road sign is the least informative thing available.

Two refinements make this usable rather than merely tidy. The first is that the binding layer is decided heading by heading and not once for the whole document. On one heading the tightest condition may be the regulator's. On the heading next to it the internal limit is far tighter. So there is no single answer to which layer binds this scheme; there is an answer for each condition in turn, and working it out means laying all three side by side on that one condition.

The second refinement is that the binding layer moves. An internal limit can be relaxed by the asset manager without any public event at all, at which point a condition that was never being tested becomes the tightest one and starts to bite. A regulator can tighten, at which point an outer edge that was pure background suddenly becomes the limit that stops a purchase. Which layer binds is a question with a different answer for every condition and a different answer over time, so it is looked up rather than remembered.

How does a written mandate become a list of what is eligible?

By subtraction, in a fixed order, and the order is not decorative. A mandate is a sentence or two of intent. On a Tuesday morning a manager needs a list rather than a sentence, and the list is produced by starting with everything the intent admits and then removing, in turn, whatever each written condition removes. Whatever survives that process is the permitted universeThe set of holdings still standing once every layer of limit has been applied to it., and it is a set rather than a plan: a manager may choose from it and may not choose outside it.

The word universe does a lot of quiet work there, so it is worth pinning down. The word does not mean everything in the market. The permitted set is everything still standing after the subtractions. Such a set is usually a great deal smaller than a reader imagines and a great deal larger than what the scheme actually holds. Two different gaps, in other words, and both of them matter: the gap between the market and the permitted set, and the gap between the permitted set and the portfolio.

How an Investment Mandate Shapes a Fund's Permitted Universe

Each of the four steps removes something different, so walk them once and slowly. Step one is the objective. The Girnar Large Cap Equity Fund says in writing what it is for, and that statement admits one kind of holding and quietly shuts out several others before any condition is even read. Step two is the written policy itself, removing part of what step one let in: the named exclusions, and the conditions hung on whatever survives them. Step three is the internal layer, where Girnar Asset Management cuts the set down further for its own reasons. Step four is not a step at all but a result: whatever is still standing is what Kalyani Bhagat may choose from.

Being eligiblePermitted by the written policy, which is a wholly different question from chosen by the manager. at the end of that process means one thing only, and it is worth saying in the plainest words available: the scheme is allowed to hold it. Eligible does not mean attractive, does not mean intended and does not mean present in the portfolio. Every holding a scheme has is eligible; almost nothing that is eligible is a holding.

Now the consequence that makes the arrangement worth having, and it reads as a cost before it reads as a benefit. Suppose Kalyani Bhagat arrives at real conviction about an instrument the four steps have ruled out. She cannot act on it. Not with a small amount, not with a note explaining herself, not with the approval of somebody senior. The fence was agreed with the people who bought units before she formed the view, so her certainty is irrelevant. Their agreement was to a scheme of a particular kind rather than to her judgement in general. A manager who is stopped by the fence is not experiencing a failure of the arrangement, she is experiencing the arrangement working exactly as the people who bought units were promised it would.

Say it from the other side and it lands harder. Somebody put Rs 1,00,000/- into an equity scheme. If a strong enough conviction could move the fence, then what they bought was not an equity scheme at all; it was a manager, wrapped in a document that turned out to be advisory. The fence being unmoved by conviction is the entire difference between those two products.

What does an investment policy actually contain?

Six headings, and the headings are worth learning because the values under them are not. Any scheme's written policy works through the same six questions in some order or other. Which kinds of asset may be held at all. Within a permitted kind, what actually qualifies. How much of the pool any single position, or any set of connected positions, is allowed to become. How much of what the scheme holds has to be capable of being turned back into money quickly. Whether derivatives may be used, and if so for what purposes. And whether the scheme may borrow, in what circumstances and for how long.

Each heading answers a different worry. Take them one at a time. The first heading is about the shape of the product: an equity scheme that could quietly become something else would not be the thing anybody bought. The second is about quality inside that shape. Permitting a kind of asset is not the same as permitting every instrument of that kind. The third is a concentration limitA cap on how large one position, or one connected group of positions, may become as a share of the pool. Set by rule, not by preference., and it exists because a pool that leans hard on one thing has stopped being a pool in any meaningful sense. Every household living on one salary already understands this heading without needing the word.

The fourth heading is about the promise to get money back. An open ended scheme has to be able to pay people who want out, so some proportion of what it holds has to be readily saleableAble to be turned back into money quickly, without a forced discount, which every pooled vehicle needs some of.. Saleable in that sense means sold without a fire sale and without waiting for a buyer to appear. The fifth and sixth headings, derivatives and borrowing, both deal with the same underlying question in different clothes: whether the scheme may take on an exposure larger than the money sitting in it, and under what conditions.

Now the part that is easy to skip. Under every one of those six headings there is a specific value: a percentage, a cap, a rating condition, a period. Every one of those values is set by the Securities and Exchange Board of India (SEBI), published at sebi.gov.in, and revised from time to time. A limit printed from memory would not merely become dated, it would become wrong, and the reader who trusted it would be the one carrying the error.

So the working habit is a split. The six headings are stable and they settle what questions to ask of any scheme document, so they are worth memorising. The values are specific to the scheme, to the kind of scheme and to the date, so they are looked up every time, in the scheme's own document and at sebi.gov.in. Anybody who has the headings in their head and the values in a browser tab is reading better than somebody who has half remembered numbers and no structure.

Where does what a scheme may hold part company with what it holds?

The written policy and the portfolio disclosure are two different documents answering two different questions, and readers run them together constantly. The written policy answers what is possible. The portfolio disclosure answers what is actual. One is an outer edge and the other is a photograph, and the distance between the edge and the photograph is unknown unless both are read.

Here is the everyday version. A driving licence permits somebody to drive a car. The licence says nothing about whether they drove today, how far they went, or whether the car has left the compound in a month. Somebody who inspected the licence and then described the week's journeys would be inventing, and would feel entirely reasonable while doing it. The licence is a real document and it does genuinely say something about driving.

Two specific misreadings come out of this, in opposite directions, and both of them mislead the same way. The first assumes the permissions are being used, so a reader who sees that derivatives are permitted starts picturing a scheme that trades them, and builds a mental portfolio nobody has described to them. The second assumes the permissions are decoration, so when the scheme finally does something it always held the written right to do, the reader takes a permitted action as a betrayal and turns angry at a document they had already been handed.

Notice what those two have in common. The shared element is the actual lesson. Both of them are treating a permission document as a description of a portfolio. One reads the permission as a plan and the other reads it as noise. Neither is reading it as what a permission actually is, a boundary. A permission that has never been used is not a promise that it never will be, and a limit that has never been reached is not evidence that the limit is tight.

The practical rule that comes out of this is short enough to carry around. The written policy gives the boundary of the possible. The portfolio disclosure gives the description of the actual. Neither stands in for the other, and when somebody states what a scheme holds, the question worth asking is which of the two documents they were reading when they formed the view.

How does a scheme end up outside a limit with nobody trading?

Because a limit written as a share of the pool is a ratio, and a ratio has two terms, and prices move both of them. A weightOne holding sized against the whole pool, which makes it a fraction with two moving parts rather than one. is not a quantity the manager sets and then leaves alone. A weight is a division done again every single day, using whatever the prices happen to be that evening, and the result can wander across a line that nobody went near on purpose. A crossing of that kind is a passive breachA line crossed by price movement alone, with no purchase and no sale anywhere behind it.. Nothing else about a written limit surprises a first reader as much.

A jeweller keeps a picture of this in his head without ever using the word. His shop holds gold, silver and stones. He buys nothing for a month and sells nothing for a month. The gold price moves and the other two do not, so by the end of the month gold is a visibly larger share of the shop's value. Nothing came in. Nothing went out. The composition of the stock changed anyway, and it changed because of arithmetic rather than because of a decision.

Now the same thing in proportions of the Girnar Large Cap Equity Fund. The position below illustrates the mechanism and is not something this scheme holds. Take a position sitting at 5.00 per cent of the pool. On net assets of Rs 4,200 crore that position is Rs 210 crore. The position rises by half, to Rs 315 crore, and nothing else in the scheme moves at all. The rise itself is Rs 105 crore, and that same Rs 105 crore is also an addition to the pool. The pool becomes Rs 4,305 crore.

A rounding shown without saying which way it went has not really been shown, so take two notes on the arithmetic before the meaning. Rs 315 crore over Rs 4,305 crore reduces exactly to 3 over 41. The decimal runs 7.3170731 per cent and carries on. Quoted to two places it becomes 7.32 per cent, so the rounding went up, by about 0.0029 of a point. The movement of 2.32 points is that same figure less the 5.00 per cent it started at, rounded the same way and in the same direction.

Two things follow from all of that, and no third. The first is that a line drawn as a share of the pool is at the mercy of a price, so a portfolio can finish a day outside something it was comfortably inside that morning while the day's transaction record stays completely blank. The second is that whether 7.32 per cent has crossed anything at all is a separate question. The limit that would answer it is set by SEBI, published at sebi.gov.in, and looked up rather than recalled. A breach is not automatically evidence that somebody did something. A defined process exists for exactly the case where prices did it and nobody acted.

The process exists, in prescribed terms. There is a requirement to notice a crossing of this kind, a requirement to report it and a requirement to deal with it, and each of those carries specifics set by SEBI: the period, the threshold that triggers it and the remedy that follows. Specifics of that kind change, and a half remembered version gets them wrong, so the current position is read at sebi.gov.in.

Why is a limit value looked up rather than printed?

Because each of those limits is a figure that SEBI fixes and later changes. A control that slid a limit up and down would need a starting value and a range, and inventing either one would put a made up number in exactly the place where the looked up one belongs. Sliders always look authoritative, so a reader would carry an invented value away as though it had been settled. The value belongs at sebi.gov.in and is read there.

Nothing is lost by that. The crossing worked above takes one multiplication and a pair of divisions, all of them on figures already printed, and the mechanism stands up without a single limit involved. The same arithmetic runs on any starting proportion at all: take the share, apply it to Rs 4,200 crore, move that position, add the movement to the pool, and divide again. The interesting part was never any particular threshold; it is that a ratio moves when its denominator does, and that survives having no numbers to slide.

What is the one term of this policy that works out exactly?

The charge, and it is worth ending on it because the charge is the one term in the same written document that works out exactly, to the last paisa. The Girnar Large Cap Equity Fund carries an expense ratio of 1.65 per cent. On net assets of Rs 4,200 crore that is Rs 69.30 crore across a year, and spread over the 120.00 crore units in issue it is Rs 0.5775 for each unit. Both figures come out of the pool already given, so no limit is involved and nothing has to be looked up.

Be careful about what that check row proves, because it is easy to oversell. The two routes agree, but they are not independent of one another. The unit value of Rs 35.00 is itself Rs 4,200 crore divided by 120.00 crore units, so taking 1.65 per cent of Rs 35.00 is the same arithmetic with the division done first instead of last. Agreement here confirms that no slip was made in the shuffling; it does not confirm anything twice. Saying that out loud is more useful than letting a reader think two separate methods have converged on a truth.

One more thing about that charge, and then what the written policy will not do. The charge runs against the scheme's assets rather than being billed to anybody, and how that works from day to day is covered separately. The charge is a term of the same document. The written document a holder is handed carries both the fence and the price of standing inside it. Read the whole thing rather than the part with the chart in it.

What does an investment policy not decide?

Almost everything a reader actually wants to know, and that makes this the hardest boundary in the subject. The written policy settles whether a holding is permitted. The policy does not settle how much of it to hold. Nor does it settle when that proportion should change, or in which direction, or how quickly. Nor does it settle how one permitted holding should be balanced against another permitted holding. Each of those questions is real, answerable, and covered separately.

The distinction is easier to feel with the boundary wall again. The wall settles where the field ends. The wall settles nothing about crop rotation, spacing, or which corner to plant first, and a farmer who asked the wall those questions would get no answer and would deserve none. Wall builders and farmers are both doing serious work, and it is different work.

There is a second reason to hold this line hard, beyond tidiness. A written policy is stable and public. Decisions about proportions are continuous and private. Mixing them makes the stable thing look like a set of intentions and makes the continuous thing look like a rule. A reader who has learned to keep the two apart can pick up any scheme document, find the fence quickly, and know exactly which of their questions that document is going to leave unanswered.

And one final limit, the one a reader most wants the document to break on their behalf. A written policy says what may happen to the money, never what will. The document is not a promise about any outcome, it is not a floor under a loss, and it does not narrow the range of results to anything comfortable. The fence settles where the scheme may walk and settles nothing at all about the ground underneath it, so a scheme can follow its written policy to the letter on every day of a year and still hand a holder back less than they put in.

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