The Valuation Policy: Pricing What Has No Clean Price

A valuation policy is the written statement, approved before the day it is needed, of how each kind of holding a scheme carries is valued and where the valuer goes when the usual source yields nothing usable. Operations applies it. The trustee company and the auditor test whether it was applied. Its purpose is that identical holdings get identical treatment, every time.

Begin with something anyone has priced. Gold has a rate available every morning, printed in the same place for everyone, off trades that happened hours ago in sizes that mean something. An old flat in a lane where two flats have changed hands in the last two years has no such rate. Two honest neighbours asked what it is worth will give two honest and different answers, and neither of them is lying. The difference between those two situations is not that one thing is valuable and the other is not; it is that one has evidence that can be pointed at and the other does not. A scheme holds both sorts of thing, and it has to publish one figure a valuation day whether the evidence is there or not.

Doing exactly that job is what a valuation policyThe written document a scheme approves in advance saying how each kind of thing it holds is to be valued, and what to do when the usual source gives nothing usable. exists to do. The document does not make the difficult things easy. The document settles, in writing and before the difficulty arrives, what will be done about them. The answer then does not depend on who happens to be at the desk.

A single invented scheme does the work below. The Girnar Large Cap Equity Fund is an open ended equity scheme operated by Girnar Asset Management Limited, an invented manager. Its net assetsEverything a scheme holds once everything it owes has been taken off. stand at Rs 4,200 crore, there are 120.00 crore units in issue, and 3,80,000 folios hold those units between them. The first over the second gives a value per unitThe published price a holder deals at, reached by splitting net assets across the units in issue. of Rs 35.00, exact to the paise, and the division is worked rather than recited, here and at every point below. Portfolio decisions belong to Kalyani Bhagat and operations report to Sohail Merchant, both invented.

Three matters are settled elsewhere. The striking of the daily figure itself, and which day an order catches, are covered under fund prices and units. The bookkeeping that receives those figures is covered separately. A holdings disclosure and what it has to show sit under scheme documents. The written method sitting underneath the figure, and what that method is for, is the subject here.

So what is a scheme's valuation policy, exactly?

A valuation policy is a document. The point sounds small and it is the whole foundation, so hold on to it. A scheme could in principle run on shared habit, with the same experienced people doing the same sensible things each valuation day and nobody writing any of it down. The figures might even come out identical. But a habit cannot be tested and a document can, and everything this guide says about checking, about consistency and about answering for a number afterwards depends on there being something on paper to check the number against.

The contents have a shape that is easy to hold. First, it lists the kinds of thing the scheme is able to hold. Second, against each kind it sets out the usual way of arriving at a figure. Third, against each kind it names the destination for a day on which that usual way returns nothing anybody could use. Kinds, usual routes, named destinations. Approved ahead of time, by the people whose job approving it is.

Think of a housing society that writes down, at the start of the year, how a shared repair bill is split: by flat area for the roof, by the number of taps for the water pump, and by a stated committee vote where a bill does not fit either rule. Nobody knows yet which bills will land. Not knowing yet is precisely why the rules can be written calmly. Because a value has to be struck every valuation day and not once a year, the scheme is doing the same thing with a much longer list of cases and a much higher frequency.

Why must the method be on paper before it is needed?

Timing is the intellectual centre of the subject. Suppose the method is chosen on the morning it is used, after the desk has already looked at what the holding might fetch. Everyone involved may be honest and careful. It does not matter. A method selected once the answer is already in view is not a method at all, it is a decision about the answer wearing a method's clothes, and writing the method down in advance is the one thing that takes the person out of the number.

The same principle governs games. Rules agreed before the cards are dealt are rules. Rules agreed after everyone has seen their hand are a negotiation. The person proposing them may mean well and still see which rule gives which result. Once that is seen, the rest follows: the testing, the consistency, the refusal to treat one holding differently from its near twin, all of it is machinery for keeping the choice and the answer apart in time.

There is a second gain from writing it early, and it is quieter. A method fixed in advance can be argued about while nothing rides on it. No live figure is waiting on the outcome, so the people who approve it can ask awkward questions, take a week, change their minds, and none of that costs anybody anything. The same argument held on the day, with a value per unit due and a queue of orders behind it, is a worse argument conducted faster.

What actually makes a holding hard to price?

Not what most readers expect. A reader is tempted to think of hard to priceA description of how much evidence there is about a price, not a description of the thing being priced. as a property of the thing, as though some securities are inherently slippery and others are inherently solid. Difficulty does not work that way. A price is an observable priceA price taken from an actual trade that can be pointed to, rather than one worked out by a method. only while three conditions hold together. A similar thing has to have changed hands lately. The quantity involved has to have been big enough to count, rather than a token parcel. And that dealing has to show up somewhere a person can go and find it. Remove a single one of those three and what remains has stopped being an observation. The remainder is an estimate with an observation lying somewhere behind it.

So hard to price reports on the evidence rather than on the thing being valued. One holding can therefore be straightforward on Monday and awkward on Tuesday without a single feature of it having moved. Its terms are the same. Nothing was signed. The trading around it moved, and the trading around it belongs to the world rather than to the holding.

The flat in the lane is the same story. The flat became hard to price the month the two comparable sales stopped being recent, and it will become easy again the month a neighbour sells. Nothing happened to the flat. The practical consequence follows: because difficulty arrives from outside and without warning, nobody can wait to see which holdings will need special handling and write the method then. By the time it is known, the scheme is already in the situation the writing was supposed to protect it from.

Where does the policy send the valuer when the price is missing?

Somewhere stated. Those two words are the whole answer. The policy describes an order of resort. The ordinary source is tried first. Where that yields nothing usable, the policy points at another observable input. Where that is unavailable too, it points at a stated construction built from inputs that can still be observed. And where even that runs out, it points at a judgement set down in advance, reached by a group named for the purpose rather than by whoever happens to be nearest the screen.

Something climbs as that list descends. The amount of judgement rises at every step and the amount of evidence falls. The order runs in that direction for exactly that reason, and no step may be skipped to reach the comfortable end of it. The idea here is usually called fair valuationValuing a holding at what the evidence supports, using a stated route, rather than at whatever figure is convenient.: the aim is the figure the evidence supports, reached by a route agreed in advance.

Any step of a fallbackThe stated alternative a policy sends the valuer to when the usual source gives nothing usable that day. is acceptable not because it is clever but because it was fixed before the day and applies to every instance, so a fallback invoked for one holding and quietly skipped for its near twin has stopped being a policy and become a decision about one number. The last step insists on a group for the same reason. One person can be persuaded, hurried or simply mistaken without anyone noticing. A named group leaves a trail, and the trail is what the testers read.

Who applies this policy, and who checks the work?

Different people, and the separation is the point. The operations team at Girnar Asset Management, headed by Sohail Merchant, applies the policy and computes the value per unit from the result. Operations holds the producing job. Kalyani Bhagat, who manages the portfolio, decides what the scheme buys and sells; she does not strike the figure the scheme is valued at, and that split is deliberate for the same reason the writing has to come first.

Then come the testers. The trustee companyThe party a scheme is held in trust with, on behalf of the people whose units it is, and which answers for how it gets run. has to answer for whether the written policy was in fact obeyed, and it puts that question to the manager instead of settling the matter alone. The auditor examines the applying of the policy against the scheme's own records. The auditor's question is a third one again: not whether the route is a good route, but whether the route on paper is the route that got walked. Securities the scheme has bought sit with the custodian, and the holder records sit with the registrar and transfer agent. Both are roles rather than firms, and what teaches is the split between the jobs.

Neither checking party puts out a figure of its own, and that is the design rather than a gap in it: anybody who supplies the answer has stopped checking it, and a scheme holding two rival numbers has no way left to decide which of them is the number. The trail is what gets examined. Did the route on paper get walked, on every instance, in the manner the document lays down.

Where the requirement itself comes from

Because five separate things here are settled by the Securities and Exchange Board of India (SEBI), and each of the five gets revised from time to time. The contents a policy is obliged to carry. The test that puts a holding into the category with no price anybody can use. Which alternative governs in which circumstance. How long an exceptionA departure from the ordinary method, permitted only where the policy says so, for a stated reason and with stated approval. is allowed to stand before it stops being one. And whose approval it takes. The five points end there.

The reason a rule is needed at each of those five points is stable and can be taught. The position each rule sits at today is not, and it is a one minute check at sebi.gov.in. The distinction travels well beyond valuation policy, separating the part of any regulated subject worth learning from the part worth looking up.

Say it without dressing it up. A printed statement of today's requirement would not merely go out of date on the morning the requirement shifted. The statement would go false. Worse, somebody who picked the requirement up second hand and repeated it a year on would be citing a source that cannot keep the requirement current on their behalf. Naming the place instead costs a minute and costs nothing in understanding. The understanding sits above and the figure is a two click errand.

What does one valuation choice do to the published figure?

Principle without arithmetic will not stick, so now put rupees on it. Take the Girnar Large Cap Equity Fund again: Rs 4,200 crore of net assets, 120.00 crore units in issue. One over the other lands on Rs 35.00 a unit, exact to the paise and with nothing trailing. Rs 35.00 is the settled figure every line below points back to.

Now put to it the question a written method exists to settle. Say one holding could sit at either of two figures a careful person would defend, and put the gap at Rs 1.20 crore, or Rs 1,20,00,000/-. Take that over 120.00 crore units. The division comes to Rs 0.01 a unit, on the nose, and the published figure walks from Rs 35.00 up to Rs 35.01. Turn the same gap into a share instead and it reads 0.02857 per cent of net assets, rounded half up to five decimal places. The share is precisely one thirty fifth of one per cent.

A gap too small for anyone to raise their voice over inside a single holding becomes a whole paise on the number that gets published, and that paise is the number all 3,80,000 folios buy and sell at. The sensitivity sits in one sentence, and it is why the route behind a figure deserves more attention than the figure does.

Two of those expressions look like independent confirmations and they are not. Rs 4,200 crore over 120.00 crore units is Rs 35.00 by construction, so Rs 1.20 crore over Rs 4,200 crore and Rs 0.01 over Rs 35.00 are one equation rearranged, and the second can never disagree with the first, so it proves nothing. Both come to one thirty fifth of one per cent because they are the same ratio written twice. A check has to be able to fail before agreeing means anything.

So here is one that can fail, built on a number outside that identity. The scheme record carries 3,80,000 folios. Average units a folio is 120.00 crore over 3,80,000. The quotient is 3,157.894736 and runs on. Rounded half up to two decimal places that is 3,157.89 units. One paise on that is Rs 31.5789 for the average folio. Multiplied back out across 3,80,000 folios it comes to Rs 1,19,99,982/-, against the Rs 1,20,00,000/- started from. The gap is minus Rs 18/-, and it does not cancel.

The Rs 18/- is not a slip to be quietly smoothed over. The Rs 18/- is the price of rounding the average folio to two decimal places, once that rounding gets multiplied back across 3,80,000 folios, and carrying it as a signed row is the honest treatment. The exact per unit move is untouched by any of it: Rs 1,20,00,000/- split across 120.00 crore units lands on one paise and leaves nothing behind.

Now say the gap out loud. The actual holdings of the Girnar Large Cap Equity Fund are nowhere set out: no weights by sector, no weights by security, no run of figures across days, and no worked case of a holding on it losing its price. None of that was needed, and that is the point worth carrying away: the effect of a valuation choice on a published figure can be sized from the unit count alone, with the portfolio behind it left a complete blank.

What if somebody transacted on a figure that was wrong?

Something has already shifted, and that is the part readers miss. Units went out, or came back in, priced off a number that ought to have read otherwise. No suspense account caught the difference and held on to it until somebody spotted the problem. The difference went straight into the exchange between the holder transacting and the pool belonging to everybody who did nothing at all that day.

Put the size on it with the figure above. If units were allotted at Rs 35.00 when Rs 35.01 was the figure that should have applied, then for every 1,00,000 units allotted, Rs 1,000/- of value crossed from the holders who stayed to the holder who transacted. Reverse the direction for a redemption and the sentence reads the same way with the parties swapped. Nobody was asked and nobody was told. The transfer is a by product of the arithmetic rather than an instruction anyone gave.

A mispriced valuation is more than a filing problem: it has redistributed the holding before anybody spots it, and setting the record straight later means tracing what each party took from which other party and pushing it back, a slower business than pricing correctly while the day was still open. The register of who holds what, covered separately, is the record that has to be unpicked, and unpicking it is slow precisely because it was correct as a record of what happened. The register faithfully recorded a transaction at the wrong price.

What does a valuation policy not make true?

Two things, and they need saying because the natural reading of everything above is too generous. A valuation policy does not make a value correct, and it does not protect a holder from a loss. A policy makes a value consistent and accountable instead, a different and smaller claim. Consistent, in that the same kind of holding gets the same treatment on the same evidence tomorrow as today. Accountable, in that somebody can afterwards read what was done and check it against what was written.

Correctness is not on offer, and the reason sits in the third section above. Where the evidence for a price is thin, there is no hidden true figure sitting behind the estimate waiting to be recovered by a good enough method. There is a range that the evidence supports. A policy picks a stated route through that range and applies it to everyone equally. Equal treatment is worth a great deal, and it is not the same as being right.

The same answer holds for anybody who wants to know whether a scheme with a good valuation policy is a safer holding. A written method does not settle that. A written method gives the ability to ask a question and get a checkable answer. A written method gives nobody an assurance about anything the scheme carries or what it will be worth.

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.