When a Corporate Action Pays Cash Into a Scheme

Four kinds of issuer event send money rather than securities to a scheme: a dividend, an interest payment, a security repaid at maturity, and a buyback. The published figure per unit shifts at the moment the scheme becomes owed the money, not on the day it lands. Between those moments the amount sits inside the scheme's assets as something owed. Treat the two moments as one and the same rupee is counted twice.

Start with the surprise. Money reaching a scheme is not, by itself, an event. The arrival looks like one. A payment lands, the bank record grows, and somebody could reasonably expect the number a holder watches to jump that morning. The number does not move, and on a well kept set of books it should not. The arrival is the last of three separate moments, and the interesting one happened earlier. If the scheme has already put the amount into its books as something it is owed, the day the cash lands changes the form the money takes and not the amount the scheme has. Nobody is better off on arrival day. The holders were already better off, and the record already said so.

Everything below is worked on a single invented scheme, so the figures stay comparable from the first section to the last. The scheme is called the Girnar Large Cap Equity Fund. Girnar runs on an open ended basis, holds equity, and is operated by Girnar Asset Management Limited. Its net assetsEverything a scheme has, minus everything it owes. The difference is the total that gets shared out over the units. come to Rs 4,200 crore, against 120.00 crore units spread across 3,80,000 folios. Put the first of those figures over the second and a unit comes to Rs 35.00, arrived at rather than quoted. Operations are headed by Sohail Merchant, and responsibility for the portfolio is Kalyani Bhagat's.

Three matters are settled elsewhere and none of them gets built again: what a scheme actually is and who stands where inside one, how the figure per unit gets struck and which day a holder is given, and the fact that a daily charge already runs against what the scheme has. Two things are new: the route a rupee takes from an issuer event into the published figure, and the one place along that route where it can be counted twice.

Which issuer events send cash to a scheme rather than securities?

Four of them, and the list is short enough to hold in mind. A dividend is cash paid out by a company on the securities the scheme holds. An interest payment is cash paid out by a borrower on a debt security the scheme holds. A security repaid at maturity returns the amount borrowed as cash on a date fixed when it was issued. And a buybackAn offer by a company to buy some of its own securities back from holders for cash. is an offer to buy securities back for cash, so a scheme that takes it up hands over securities and receives money.

The contrast between that class and the other one is the quickest route to knowing what to expect. Issuer events that move the count of securities a scheme is holding are covered separately: extra securities arriving, one security becoming several, several becoming one. Opposite things move in the two classes, so working out which class an event belongs to settles which figure ought to react before anyone has gone anywhere near it. An event that changes a count leaves the value of the holding where it was and rearranges how that value is divided up. An event that pays cash leaves the count of securities where it was and adds to what the scheme has.

Think about a household for a moment. A ten rupee note swapped for two five rupee notes is the first class: the household has exactly what it had, in more notes. A relative repaying a loan is the second class: the count of notes in the tin has nothing to do with it, and the household genuinely has more than it did. Nobody confuses those two around a kitchen table. The two classes get confused inside a scheme because both arrive as a notice on a screen, both are called a corporate action, and both are handled by the same operations desk on the same morning.

Why does the figure per unit move before the money arrives?

Because an asset exists for the scheme from the instant it becomes owed, and the figure per unit is built out of whatever the scheme has. An entitlementA right to receive something that has already come into existence, whether or not the thing itself has turned up yet. is not a hope or a forecast. An entitlement is a right that has already come into existence at an instant the issuer fixed. From that instant the scheme is not waiting to have something. The scheme has it already, in a form that is neither cash nor a security.

Now run the alternative and watch it break. Suppose a scheme held the figure per unit still until the money landed. For the whole stretch in between, the published number would be short by the entire amount the scheme was owed. Every holder who redeemed in that window would be paid out of a total that left out an asset the scheme genuinely had, and every holder who subscribed in that window would buy in cheaply for the same reason. Waiting for the cash does not make the figure more careful. Waiting makes the figure wrong in a known direction for a known stretch, and a wrong figure is worse than one that moves early on a right that exists.

The everyday version is payday. Somebody who has worked a full month is not poor on the last day of it and suddenly better off on the first of the next. The work is done, the amount is earned, and payday only decides where the money sits. A household asked to state what it was worth on the last day of the month, leaving out a month of earned pay because the credit had not appeared, would produce a number nobody in the house believed. A scheme is in the same position, except that its number is published and thousands of people transact against it.

What are the three moments an amount owed passes through?

Three, and they are moments rather than periods. The instinct is to reach straight for how long each step takes, and the duration of each step is not settled by arithmetic. The order holds regardless, and the order never changes.

The first moment is announcement. The issuer says a payment will be made and states the terms on which it will be made. Nothing has yet happened inside the scheme, and nothing should: an announcement fixes the shape of a future right rather than creating one. The second moment is the one that matters. The scheme becomes owed the money, and once that is true the amount is written into the books as an asset. RecognitionThe act of putting an amount into a set of books, which is what makes it part of the totals rather than merely known about. is the act of putting it in. Putting it in is what makes the amount part of the totals rather than something the desk merely knows. The third moment is arrival. Cash reaches the account of the scheme.

Each of the three states between those moments is a perfectly legitimate place for a scheme to be, and nothing goes wrong because a scheme was sitting in one of them; it goes wrong on the way from one to the next. A scheme that is owed money and has not yet recorded it is not doing anything wrong if the conditions for recording it are not yet met. A scheme carrying a recorded amount that has not arrived is in the ordinary state of any business that has sold on credit. The faults are a move made twice, a move made once and left half done, or a move never made at all.

What is an amount owed doing inside a scheme's net assets?

Sitting there as an asset, exactly like the securities and exactly like the cash, and being invisible in the published figure. The accounting name for it is a receivableAn amount somebody else is due to pay, carried in the books as an asset until the payment turns up., and the word carries no drama. A receivable is what a shopkeeper has after supplying a canteen on credit on Tuesday and expecting the money on Friday. The supply happened. The asset is real. The asset is simply not cash yet.

Now the consequence that the rest of the argument turns on. Net assets are one figure. Net assets are not published as a list broken into cash, securities and amounts owed, and the figure per unit is that one total divided by the units in issue. A holder who subscribes on a day when a large amount is owed to the scheme has bought a share of money that has not turned up, at a price per unit that already counts it, and that is the right answer rather than a fault. The scheme is entitled to the money. Leaving it out would shortchange the holders who were there when the entitlement arose, in favour of the one buying in today.

The split is not available anywhere. Net assets for this scheme are a single figure of Rs 4,200 crore, and how much of that is cash, how much is securities and how much is owed to the scheme cannot be read off it. A holder looking at the published total is looking at a number that does not separate the three.

How does one rupee end up inside the published figure twice?

By making the second entry without undoing the first. Follow it slowly. Neither entry is wrong, and that is the whole point.

On the day the money becomes owed, somebody writes an amount owed into the books. The total of net assets goes up by it. Correct entry. Later the money turns up, and somebody writes the cash into the bank account of the scheme. Up goes the total again. Taken by itself, that entry is correct too. The first entry has to come out alongside the second. The amount owed stopped existing the moment it turned into the cash. Clearing an itemTaking an amount out of the books because the thing it stood for has been settled, so it is not left standing alongside its own replacement. is that removal, and when it is skipped the scheme is carrying both the money and the promise of the money.

What makes a double countThe same amount sitting in a total twice, usually because a replacement was added without the thing it replaced being taken out. so easy to commit is the spacing: one entry lands days after the other, the two come off different desks, and taken singly on the morning it was written each is perfectly sensible. Staring at either entry alone will therefore never turn up the fault. The entry that recorded the amount owed was right when it was made. The entry that recorded the cash was right when it was made. The fault is not in an entry at all; it is in a missing one, and a missing entry has no line to inspect. A missing entry determines the shape of every control that catches this class of error: the check has to be a comparison of two records rather than a review of one.

Back to the shopkeeper. He writes the Tuesday supply into his book. On Friday the canteen pays him and he writes the cash in as well, and he never ticks off the Tuesday line. His book now says he did two supplies that week. He did one. Ask him to total his week and he will overstate it by exactly the amount of one supply. Neither line is odd, so nothing on either line looks odd. Only laying the book against the day's takings finds it.

What does the arithmetic do on the day of the entitlement and on the day of the receipt?

Two very different things, and setting them side by side in rupees is the cleanest way to see it. No holding of this scheme is on record, and no issuer event against one. Without a holding there is no dividend, no security, no rate and no per security amount to name. The amount owed is carried as a letter, R, and a size is put on R only at the end, so that the conversion can be shown.

The scheme starts with net assets of Rs 4,200 crore, or Rs 42,00,00,00,000/-, and 120.00 crore units, or 1,20,00,00,000 units. Division gives Rs 35.00 a unit exactly, rather than a quoted figure.

Recognition day first. Net assets become Rs 4,200 crore plus R, nothing at all happens to the units in issue, and the figure per unit therefore gains R over 1,20,00,00,000. Receipt day next, with the item cleared as it should be. In one movement the cash side gains R and the owed side gives up R. The total is left flat. Flat total, unchanged units, so the figure per unit is flat as well: handle a receipt properly and the day the money lands is a day on which the published number does absolutely nothing.

Put a size on R to see the conversion. Take R at Rs 1.20 crore, or Rs 1,20,00,000/-. Across 1,20,00,00,000 units that is exactly one paise a unit. Net assets go from Rs 4,200.00 crore to Rs 4,201.20 crore and the published figure goes from Rs 35.00 to Rs 35.01, both exactly. Now leave the item uncleared when the cash arrives. Net assets go to Rs 4,202.40 crore and the figure reads Rs 35.02 when the correct reading is Rs 35.01. The error is exactly one paise: Rs 1.20 crore too much, or 0.02857 per cent of the net assets of the scheme.

The third row and the fourth row are the same morning handled two ways. R is Rs 1.20 crore throughout. In the handled row the total is unchanged from the row above it, and that is the result the whole argument turns on. In the not cleared row the total has gained R for a second time and the reading is one paise too high. Everything else in both rows is identical, including the units in issue. No event of this class touches them.

Can a check on this arithmetic ever fail?

One of the two checks cannot, and only one of them is a test. The difference matters. The amount owed falls by exactly what the cash rises by, so the change in net assets is zero, and zero divided by any number of units is zero. A check that the figure per unit does not move on the day the cash arrives therefore cannot fail. That is an identity. The identity is true before it is computed, and it stays true whatever R is. Running it proves nothing about the books; it only confirms that subtraction works.

A check that can fail runs the other way, on the same numbers: the one paise is converted down to the average folio and then built back up. The scheme has 1,20,00,00,000 units across 3,80,000 folios, so the average folio holds 3,157.894736 units and a little more. Rounded half up to two places, that is 3,157.89 units. One paise a unit on that folio is 3,157.894736 paise and a little more, or Rs 31.578947 and a little more. Rounded half up to the nearest paise, that is Rs 31.58. Multiplied back across 3,80,000 folios that comes to Rs 1,20,00,400/-, against a true amount of Rs 1,20,00,000/-.

Signed, the residue comes to plus Rs 400/-, and nothing cancels it out. The residue is the rounding on one average folio multiplied back across every folio, all of it in the same direction. Two further things are worth saying about it. No distribution of folio sizes is on record, so the average folio is not any actual folio. The rebuild is arithmetic performed on an average, and the residue is part of what that costs. And the residue is a rounding artefact rather than an error in the books: nothing was lost, the per folio figure was simply stated to the nearest paise and then treated as though it were exact.

Who has to agree the money turned up, and agree with what?

Two records have to agree, and neither of them is the one that made the entries. The scheme keeps a record of what it expects, item by item, built from the entitlements it recognised. The bank account keeps a record of what actually turned up, item by item, built from nothing except money moving. MatchingSetting two records side by side and pairing each item on one with an item on the other, so anything unpaired stands out. is the act of pairing each expected item with an arrived item, and a paired item is one that can be cleared.

Matching tests one record against another rather than recomputing either, and that distinction is the whole reason it works on a fault that lives in a missing entry. A recomputation asks whether the arithmetic inside a record is right. Here the arithmetic is right. The fault is that two records disagree about how many times a thing happened, and only setting them side by side can show it. An item on the expected side with nothing opposite it is either money that has not come yet or money that is not coming. An item on the arrived side with nothing opposite it is money nobody was expecting. Both are questions, and both are found the same way.

The working discipline behind this is covered separately, where units, cash and holdings are all put side by side and the operational quantities are real rather than a letter. The shape is what carries over: the check that catches a double count is a comparison, not an inspection.

What happens when money that was announced never arrives?

One case is left, and it is the one where the arithmetic stops being tidy. Everything so far has assumed the money turns up. Sometimes it does not.

If the money never comes, then the amount owed was an asset the scheme did not have. The books have to be put right, and putting them right takes the amount back out of net assets, and taking it out drags the published figure per unit down by whatever R comes to over 1,20,00,00,000. Here, with R at Rs 1.20 crore, it is that same single paise travelling in reverse: Rs 35.01 falls back to Rs 35.00. Every holder who bought or redeemed between the day the amount went into the books and the day it came back out transacted at a figure that was later found to be wrong. A holder who redeemed in that window was paid out of a total that included an asset the scheme never received. A holder who subscribed in that window paid for one.

Notice what that does to the shape of the risk. The double count and the never arrived amount are opposite faults with the same consequence: a stretch of days on which people transacted against a published number that did not describe the scheme. In the first case the number is too high because a replacement was added without the original being removed. In the second it is too high because the original should never have been standing in the first place. Both are corrected by a movement, and a movement in a published figure has a transaction history sitting underneath it.

The market regulator sets what is done about a correction of that kind, whether anything is done at all, and on what terms. The position of the day is published at sebi.gov.in. Structure is what survives whatever it says: the amount sat in the total, the total went out, and people transacted against it.

Why can the amount owed on a single holding not be worked out?

Because three inputs are needed to work it out, and none of the three is on record for this scheme.

Three inputs decide what this scheme is owed on any one holding: which issuer, how many securities of that issuer the scheme holds, and how much is payable on each. The record here carries none of the three, and holds one figure for net assets, one figure for units in issue and one figure for folios, and nothing else. A made up company and rate would produce a worked example that looks like evidence and is not, and a number carried away from it would have come from nowhere.

The conversion works at the level of the scheme instead, holding the amount as a letter until the last step, and that costs the argument nothing at all. The three moments exist whatever the amount is. The amount owed sits inside net assets whatever the amount is. The double count is a risk whatever the amount is.

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