The Record Date: How a Scheme Decides Who Is Entitled
A record date names one moment, and a scheme reads its own holder list at exactly that moment to settle who counts. Purchases and exits rewrite that list on every working day. Until a moment is named, who holds units has no answer. Whatever stands there counts. Whatever lands later does not.
Start with the record the question is being asked of. A scheme keeps a unit registerThe scheme's own list of holders, one line each, showing how many units that line holds., one line for each account, and each line carries a unit figure that gets written up or written down whenever an application or a redemption is processed. Those write-ups and write-downs happen on every working day the scheme is open for business. So who holds units in this scheme has no single answer. The register has as many answers as there are moments to ask it at, and a scheme that wants to pay something out has to pick one of them. A record date is that pick, made in advance and announced rather than discovered, and once the moment is settled every remaining part of who is due what stops being a matter of judgement.
One scheme runs through everything that follows. Girnar Asset Management Limited, an invented asset manager, runs the Girnar Large Cap Equity Fund, an open ended equity scheme with net assetsWhat the scheme holds less what it owes. A payment comes out of that figure. of Rs 4,200 crore and 120.00 crore units in issue. Divide the first figure by the second and a single unit is worth Rs 35.00 exactly. The scheme is held across 3,80,000 foliosAn account with the scheme, under which one holder's units are recorded.. Kalyani Bhagat manages the portfolio and Sohail Merchant heads operations. The auditor, the distributor, the custodian, the trustee company and the registrar and transfer agent each sit in a separate role, and that separation is what lets one party's work be checked against another's.
Four things are settled elsewhere: what a scheme is and who runs it, what a unit is and how the value per unitNet assets divided by the units in issue, struck each working day and published. is struck each working day, the day's value an application receives under the cut-off rules, and what a folio is and what a statement of one reports to a holder. All four are covered separately.
One figure stays a letter throughout. Whether a scheme pays a distributionA payment a scheme makes on each unit out of what it holds, reducing the pool it is paid from. at all, and how much, is settled by that scheme's own terms, and no amount is set for the Girnar Large Cap Equity Fund. So the amount paid on each unit is written as D: a letter, never a number. The conclusion the arithmetic reaches is a cancellation. A cancellation is true at every value D could possibly take, and a letter carries it as well as any number would.
What does a record date actually decide?
One thing, and it decides that one thing completely: which lines on the register are entitled, and to how much each. Nothing else about the payment is settled by it. Not what is being paid. Not where inside the scheme the money comes from. Not when it reaches anybody's bank account. Only the list.
A cooperative housing society has the identical shape, and the money there is small enough to hold in mind. The committee decides to hand back a share of the money sitting unused in the maintenance fund, so much for each flat. Flats change hands. Somebody moved out recently and somebody else moved in after that. The committee cannot simply pay the residents. The residents is not a list at all: it is a description that fits different people depending on when it is read. So the committee names a date, reads the flat register on it, and pays whoever is standing on that register. Nobody in the building thinks this is strange. The alternative is not a kinder rule; it is no rule.
A scheme is doing exactly that, with far more lines and a computer doing the reading. A record date is the answer to the question when, and once when has an answer, who and how much stop being arguments and become multiplication. The entitlementThe amount a particular line on the register is due, worked out from the units on that line. of one line is that line's unit figure multiplied by the amount paid on each unit, and there is nothing in that sentence for a person to decide.
Why does an entitlement have to be an instant rather than a stretch of time?
Because two lists produce two answers, and nothing anywhere settles which of the two to use.
Suppose a scheme tried to be generous about it and said that everybody who held units at any point during a week is entitled. The register read on one day of that week yields one set of lines carrying one set of unit figures. Read on another day, it yields a different set. A holder who bought and then sold inside that week appears on one reading and is absent from the other. Both readings are honest. Both are complete. The two readings disagree with each other, and the arithmetic offers no way to break the tie. The arithmetic was never handed a well formed question in the first place.
Arithmetic forces the choice, not administrative habit. Some version of this cut turns up in every pooled arrangement ever built, and goodwill on its own has never been able to do without one. A club splitting a surplus among members, a partnership settling a departing share, a company paying on its shares: each of them fixes a moment first, because each of them has to compute over a single list or not compute at all.
The same shape appears in a school photograph, with no money in it at all. The class in the photograph is the class that was standing there when the shutter opened. A student who left before that morning and a student who joined the week after are both missing from the picture, and the picture is not wrong about either of them. A photograph records one instant and has never been able to record anything else. A register read at a record date is the same kind of object, and it is no more unfair than the photograph is.
A scheme wants to be generous and proposes working out who is due what from the register as it stood over a whole week, rather than at a single moment. Why can that not be done?
Whose units count at that instant, and whose do not?
The units standing on a line at the instant count. Everything else about that line is simply not part of the test.
The test carries two consequences that get read wrong more often than any other pair on this subject, and here they are one at a time. A holder who redeemed just before the instant is not entitled, however long they had held the units up to that point. And a holder whose units are written to the register after the instant is not entitled, however early they had sent the money in. Length of holding does not appear in the test. Intention does not appear in the test. The date on which money left somebody's bank account does not appear in the test. There is the register, read once, and that is the whole of it.
The same test applies whether the units sit on a folio maintained by the registrar and transfer agent or are held in dematerialised form with a depository. The record is still read at an instant; only the place it is read from changes. In India those depositories are Central Depository Services (India) Limited (CDSL) at cdslindia.com and National Securities Depository Limited (NSDL) at nsdl.co.in.
A holder redeemed the whole of a folio just before the instant, sees the payment announced afterwards, and complains that years of holding counted for nothing. How should that complaint be answered?
Is either of those two holders treated unfairly?
Neither is, and the reason is worth going over slowly. The exactness of the test is what makes the arrangement honest rather than merely tidy.
Take the holder who left just before the instant. At the moment that redemption was priced, the scheme still held every rupee it was about to pay out. The assets about to go out were inside net assets, so they were inside the value per unit, so they were inside the amount the redemption settled at. The departing holder was paid the value of the entitlement, in the redemption price rather than as a separate credit, and a rupee paid inside a price is the same rupee as one paid into a bank account.
Now take the holder who arrived after. At the moment they bought, the scheme had already parted with what it paid out, so the value per unit they paid was already lower by exactly that amount. The arriving holder bought a cheaper unit. Handing that holder the payment on top would give them the reduction and the money both. The same rupee is counted twice, and this time the double count runs in the holder's favour.
The value of an entitlement sits inside the value per unit right up to the moment it leaves, and reading the register at an instant simply follows it out. The test looks harsh from either side and is exact from both.
A scheme puts money out on each of the units standing on its register. What does that do to the value per unit?
What does paying an amount on every unit do to the value per unit?
Work it as an identity and the answer arrives in three lines.
Write the amount put out on each unit as D. D goes out on every unit standing on the register, so net assets drop by D times however many units are standing there. Paying money out neither cancels anybody's units nor creates anybody else's, so the number of units does not change. The value per unit is net assets divided by units. Reduce the top by D times the units, leave the bottom exactly where it was, and the value per unit is lower by D.
Written out on the scheme carried through this guide: the value per unit was Rs 35.00, and once the payment is made it is Rs 35.00 less D. Not approximately. Not usually. Exactly, and for arithmetic reasons that no operating decision can soften.
The result is true for any D at all. The size of the payment never enters the argument. Ten paise on each unit, one rupee on each unit, five rupees on each unit: at every one of them the drop in the value per unit matches what went out on each unit.
Why is a holder no better off once the payment has been made?
The sum can be done on any holding, so nobody's word need be taken for it.
Before the payment a holder has a number of units, each worth the old value per unit, and nothing else. After the payment the holder has the same number of units, each worth D less, and D in cash for every one of them. Taken together: the fall in the units is the unit count multiplied by D, and the cash is the unit count multiplied by D. Two identical products of two identical factors, so they cancel, and the total is the figure it was before anything happened.
Here is the everyday version. A shopkeeper takes Rs 500/- out of the till at closing and puts it in his pocket. He now has Rs 500/- in his pocket that he did not have in the morning, and the till has Rs 500/- less in it than it had a minute ago. The pocket and the till are both his, so nobody would look at the pocket on its own and conclude the shop had a good day. Money put out on each unit carries value out of the unit and into a bank account without creating a rupee of it, so the value per unit comes down by precisely what left the pool and a holder is not made better off by the payment itself.
The name for the mistake is double counting, and it is the commonest misreading on this whole subject. A holder who reads a payment as something the scheme produced has counted one rupee once as value inside the unit and then a second time as cash outside it. Nothing was produced. Something was moved.
Cash received on a folio plus what its units are worth after the payment, set against what those units were worth before it. Which side is larger?
The cancellation can be watched happening on one illustrative line. The line below holds 3,158 units, the average line of this register rounded to whole units. Rounding it once keeps anything inside the arithmetic from being rounded twice. At Rs 35.00 a unit that line is worth Rs 1,10,530/- before anything is paid, and at the default setting nothing has been paid at all: the cash bar is exactly zero, drawn as an empty dashed outline, and the two together stands at 100.0 per cent of where it started.
What happens to the three bars as the amount paid on each unit rises?
Move the control and watch the first bar shrink and the second bar grow. The third bar is the two added together, and it does not move at any setting. An unmoving third bar is the whole of the claim.
The amount put out on each unit, D, measured against the value per unit it started from: zero
The control is at zero, so nothing has been paid on any unit. The cash bar is exactly zero and is drawn as an empty dashed outline with its label outside it. The third bar stands at 100.0 per cent of where it started, and it will stand there at every setting.
How does this run on the Girnar Large Cap Equity Fund?
Run it across the actual register and the size of the exercise becomes visible without any payment amount appearing at all.
At the instant, the register carries 3,80,000 lines adding to 1,20,00,00,000 units. Divide the units by the lines and the average line carries about 3,157.89 units. At Rs 35.00 a unit that is about Rs 1,10,526/- of holding. The lines and the units are the whole of the inputs. The entitlement of the average line is about 3,157.89 multiplied by D, and the fall in the value of that line's units is about 3,157.89 multiplied by D, and the two are the same product of the same two numbers.
| The step | Worked on this register | Result |
|---|---|---|
| Lines read at the instant | One line for each folio on the register | 3,80,000 lines |
| Units standing on them | The scheme has 120.00 crore units in issue | 1,20,00,00,000 units |
| The average line | 1,20,00,00,000 units divided by 3,80,000 lines | about 3,157.89 units |
| What that line holds | About 3,157.89 units at Rs 35.00 a unit | about Rs 1,10,526/- |
| Cash it receives | About 3,157.89 units multiplied by D | about 3,157.89 D |
| Value its units lose | About 3,157.89 units multiplied by D | about 3,157.89 D |
| The two together | Cash gained less value lost, at every D | zero |
The row saying the cash equals the fall and the row saying the total is unchanged are one equation written two ways, so the fact that they agree proves nothing at all. A check that cannot fail is not a check. So here is one that can fail, on the same register.
Multiply the average line back by the number of lines and see whether the unit total comes home. About 3,157.89 units across 3,80,000 lines is 1,19,99,98,200 units, against a stated 1,20,00,00,000 units. The signed residue is minus 1,800 units. At Rs 35.00 a unit that is Rs 63,000/- of holding the rounded average has quietly mislaid. The residue does not cancel and was never going to: it is the cost of rounding an average to two decimal places, and it is exactly the sort of gap a real entitlement run has to close before any money leaves. This is why an entitlement is computed line by line from actual unit figures rather than from an average. An average is a way of describing a register and never a way of paying one.
The amount paid on each unit is never stated, and the whole thing is worked with a letter instead. Does the argument still hold?
What has to happen between the instant and the money leaving?
Six things, in that order, and every one of them reads from the register.
One. Freeze the list as at the instant. Whatever is processed afterwards then writes to the live register rather than to the one being paid from. Two. Compute each line's entitlement from that line's own unit figure. Three. Reduce net assets by the total of those entitlements. Four. Work the value per unit out on whatever net assets are left. The drop the identity predicted surfaces there, in a published figure. Five. Send the money out against the bank instructionThe banking details recorded against a line, saying where money due to it is to be sent. held on each line. Six. Record all of it. The entitlement computed, the money that left and the register that produced both can then be shown afterwards to agree.
Steps two and five both read from the same line, and that is the sentence to carry away: the entitlement comes off the unit figure and the money goes out on the banking details, and those two fields sit side by side on one row of one record. Sohail Merchant, who heads operations, is not running six separate exercises here. He is running one register through six readings, and a break in any reading shows up as a difference at the end rather than as a complaint at the time.
A line is plainly entitled, the payment run has completed, and the money has not reached the holder. What is the first place to look?
What breaks when the register line is right and the banking details are not?
The holder is entitled and the money does not arrive, and at no point was the entitlement in any doubt.
Picture a wedding invitation posted to an address the household left two years ago. The guest list was right. The invitation was correctly addressed to the right people. Nobody turns up. The list and the address are two different fields, and only one of them was kept current. A stale bank instruction on a register line does that to a payoutThe money actually sent out against an entitlement, as distinct from the entitlement itself.. The unit figure is right, the entitlement computed from it is right, the money is sent, and it comes back.
A failure of that kind belongs to operations rather than to entitlement, and the difference matters: the two get put right in separate places by separate people. An entitlement question is answered by reading the register at the instant. A failed payment is answered by looking at the banking details on the line, correcting them and paying again. Treating the second as though it were the first sends somebody hunting for a fault in a calculation that was never wrong. The holder waits while they hunt.
How much is a slip of one paise a unit worth on this scheme?
Enough to be worth reconciling, and the size of it is the honest reason operations is careful rather than merely diligent.
Take the smallest movement in a value per unit this arithmetic can express. One paise on every one of 1,20,00,00,000 units is Rs 1,20,00,000/-, or Rs 1.20 crore. On the average line of about 3,157.89 units, the same one paise is about Rs 31.58, a figure nobody would cross a room for.
Both readings are true at the same time and neither is the whole picture: a slip too small for one holder to notice is Rs 1.20 crore across the register, and the distance between those two views is the entire reason an entitlement is checked against the register before money leaves rather than after. The conversion gives the scale of an operational slip, and the scale of a slip is a separate matter from the size of any payment.
Who reaches for this on a working day, and what for?
Three people, and none of them is doing it out of curiosity. The head of operations is running the six steps above and watching one figure above all others: whether the total of the line by line entitlements equals the amount by which net assets were reduced. If those two differ, something in the register moved between the freeze and the computation, and it has to be found before any money leaves the scheme.
An analyst reading a year of a scheme's record does the opposite of what most readers do. Instead of adding the payments to the growth in the value per unit, the analyst first checks whether the record has already accounted for them. A return measured from one value per unit to another has the payments taken out of it by construction, and a figure assembled any other way may not.
A household holding units in a scheme that pays needs one sum and no more: put the cash that arrived alongside what the units are worth today, and hold that against what those same units were worth before the money left. If the two come out equal, then nothing was produced and nothing was lost, and the credit in the bank account is exactly what it looks like from inside the scheme: a transfer. The sum settles what a payment did, and settles nothing about whether the scheme is worth holding.
Who settles the rules around the date, and where are they written down?
From the market regulator.
The setting of a record date, the warning that has to go out ahead of it, what holders must be told and the working out of what each line is due are all matters the Securities and Exchange Board of India (SEBI) decides, and requirements of that kind get revised. A printed convention would not merely become dated when it moved; it would become wrong, and wrong is a far worse failure than dated. The current position is published at sebi.gov.in and is read there on the day it is needed. Industry level operating material for schemes is published by the Association of Mutual Funds in India (AMFI) at amfiindia.com.
Tax law then decides what a holder keeps of what is received. The Indian tax authority holds those rules, reachable at incometaxindia.gov.in, and moves them to a timetable of its own. Rates, thresholds, classifications and holding conditions are covered separately.
Three things survive every one of those revisions: an instant has to exist before an entitlement can be computed at all, the value per unit comes down by precisely what went out on each unit, and six readings of one register stand between the moment and the money. Rearranging the rules changes none of the three.
Which body decides, and where is the current position read?
SEBI settles the way a record date is set, the warning that goes with it, the working out of what each line is due and what must be disclosed to holders, and it revises all of that from time to time. Conventions, notice requirements, intervals, thresholds and deadlines are read from the current position at sebi.gov.in.
Whatever a holder ends up receiving is then taxed under rules held by the Indian tax authority, reachable at incometaxindia.gov.in. Rates, thresholds, classifications and holding conditions are set there. Industry level operating material sits with AMFI at amfiindia.com, and where units are held in dematerialised form the two depositories are CDSL at cdslindia.com and NSDL at nsdl.co.in.
Which body decides the way a record date is set and the warning that goes with it, and where is the current position read?
The error that gets made, and what it costs
A payment lands in a holder's bank account. The payment arrives as a credit, with a date against it and an amount, in a statement the holder actually reads. On the same day the value per unit is lower by precisely what went out on each unit, and that half arrives as nothing whatever: no message, no entry, no line anywhere, only a slightly smaller figure inside a number that shifts for a dozen different reasons in any ordinary week. One half of the transaction announces itself and the other half is silent.
So the holder does the natural thing and reads the credit as something the scheme produced that week, sitting on top of units they still hold in full. The misreading is not a careless one, and it is the only reading the visible evidence supports. Nothing in the ordinary flow of documents corrects it.
The cost is a reader who can no longer tell a payment apart from a return, and it goes wrong in both directions at once. The scheme gets credited with the payments. The same scheme then gets blamed for a value per unit that never grew the way the holder expected, when the missing growth is those same payments, already received and quite possibly already spent. The repair is one sum anybody can do: put the cash that arrived beside what the units are worth today, hold that against what the same units were worth before the money left, and watch the two come out level.
A reader describes a scheme's year by adding up the payments it made on each unit and the growth in its value per unit. What has gone wrong?
References
| Where | Body named | What is read there |
|---|---|---|
| sebi.gov.in | Securities and Exchange Board of India | The setting of a record date, the warning that goes with it, the working out of what each line is due and what must be disclosed to holders |
| incometaxindia.gov.in | The tax authority | The treatment of anything a holder receives after a payment on units. Rates, thresholds, classifications and holding conditions are set by this authority |
| amfiindia.com | Association of Mutual Funds in India | Industry level operating material for schemes. This body publishes that material and makes no requirement |
| nsdl.co.in and cdslindia.com | The two depositories | Registers of units held in dematerialised form rather than on a folio |
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.
