ETF vs Fund of Funds: Where You Buy and What You Pay
An exchange traded fund is bought from another holder on an exchange at a price that the market sets. A fund of funds is bought from the scheme itself at a value computed from what it holds. The exchange traded route carries one layer of charge and a price that can sit away from value; the fund of funds route carries two layers and no such gap.
Two structures, six criteria, and one difference underneath all six. An exchange traded fundA scheme whose units are listed and dealt on a stock exchange, so a holder buys and sells them the way a listed security is bought and sold. and a fund of fundsA scheme whose holdings are units of other schemes rather than shares or bonds directly. both let one payment reach a spread of holdings, both are pooled vehicles, and both publish a value per unit. Set beside each other they look like two flavours of the same thing. The resemblance does not survive the first question. The two structures differ on where the transaction happens, on what price the holder receives, on how many layers of charge apply, on what a holder must have in place, on how units come into existence, and on what can go wrong, and every one of those six differences descends from a single fact about who is standing on the other side of the transaction.
Before the criteria, what is already settled. A scheme, a unit, how a value per unit is struck, and which day's value an order gets are all covered separately. So is the expense ratio, including the fact that it runs against a scheme's own assets rather than arriving as a bill, and so is the fund of funds structure itself and the fact that its two charges add. The comparison is the new part: two structures side by side, criterion by criterion, with the arithmetic run wherever the worked figures reach.
Girnar Asset Management Limited, an invented fund house, operates the Girnar Broad Market Index Fund, an index scheme tracking a broad index that is deliberately left unnamed, and the Girnar Large Cap Equity Fund, an actively run equity scheme. Kalyani Bhagat manages the equity scheme and Sohail Merchant heads operations. Every return, charge and asset figure attached to them is a teaching figure rather than an observed one.
Who is actually on the other side of the transaction?
The counterparty question comes first, and the rest of the comparison falls out of the answer. In a fund of funds, the holder deals with the scheme. Money goes to the scheme, the scheme issues units, and the transaction is between a person and a pool. In an exchange traded fund, the holder almost always deals with another holder. Money goes to whoever was selling, units come from whoever was selling, and the scheme is not a party to any of it.
The identity of the counterparty is not a detail of convenience, and treating it as one is where most of the confusion about these two structures begins. It decides three separate things at once: what price the holder receives, where the money actually goes, and whether the transaction changes the number of units in existence at all. Every criterion below is a consequence of it.
Here is the everyday version. Consider two ways to end up with a season pass to a swimming pool. In the first, a buyer walks up to the pool's own counter, hands over the money, and the pool writes a new pass. The pool now has that money and one more pass exists than existed before. In the second, the buyer takes an existing pass from somebody in the queue who no longer wants theirs. The money goes to that person. The pool receives nothing, issues nothing, and the number of passes in circulation is exactly what it was. Both routes leave the buyer able to swim. Only one of them involved the pool.
Where does the transaction actually happen?
On the fund of funds side, through the scheme's own subscription and redemption process. The subscription process is covered separately, and nothing about it changes because the thing being bought happens to hold other schemes: an application reaches the scheme, money reaches the scheme, and which day's value applies is decided by rules the Securities and Exchange Board of India (SEBI) makes about timing and receipt of funds. Rules of that kind move, and the position in force is read at sebi.gov.in.
On the exchange traded side, through an exchange, where the counterpartyThe party on the other side of a transaction, the one whose money or securities meet the holder's. is another buyer or another seller. The scheme is not a party to that transaction at all, and it receives nothing from it. Nothing a holder does on an exchange changes the scheme's assets by a single rupee. The scheme's assets moved when the underlying holdings moved in price, not because two people traded its units between themselves at eleven in the morning.
The distinction sounds like a technicality until it becomes clear what it removes. There is no transaction with the scheme to apply an exit charge to, so a scheme that is not a party cannot apply one. The scheme is not the one paying, so it cannot decide which day's value the holder gets. And it does not receive the money paid, so nothing the holder paid reached the pool. All three follow from the same sentence.
A holder sells units of an exchange traded fund at eleven in the morning. Who pays, and what price applies?
What price does the holder end up with?
Two different processes are at work, and this is the criterion that matters most. The fund of funds holder receives a value computed from what the scheme holds: add up the holdings, take off what is owed, divide by the units in issue. The value per unit is an arithmetic output. Nobody negotiates it and nobody agrees to it. The exchange traded holder receives whatever price the transaction happened at: one buyer and one seller each carrying their own view, one transaction, one number.
The computed value and the traded price are produced by two different processes, nothing forces them to agree, and the difference between them belongs to the holder rather than to the scheme. That last clause is the part readers skip. Where exchange traded units are bought at a price above the value of what those units represent, the extra amount did not go into the scheme, did not become part of its assets, and cannot be recovered from it. The extra amount went to the person on the other side. The reverse is equally true: bought below, and the seller carried it.
The gap between price and valueThe difference between the price a transaction on an exchange actually happened at and the value computed from what the scheme holds. exists because the two numbers are made in different places by different means. One is struck by the scheme at the end of a valuation cycle. The other is struck by whoever was willing to deal, at the moment they dealt. Nothing connects them mechanically. Forces do push them toward each other, and the block dealing mechanism described further down is the main one, but pushing toward is not the same as being equal to.
How large is that gap? Not one of the schemes worked through here is exchange traded, so no traded price can be quoted, no gap against value computed, and no spreadThe distance between the price at which somebody is willing to buy and the price at which somebody is willing to sell at the same moment. observed. An invented number would carry the authority of one that had been computed. The figure below therefore draws the shape and leaves both marks unnumbered.
Exchange traded units are bought at a price above the value of what those units represent. Where does the extra amount go?
How many layers of charge does each structure carry?
A fund of funds carries two. There is the outer scheme's own charge, running against the outer scheme's assets, and there are the charges inside every scheme the outer one holds, running against those schemes' assets. The second set was already taken out before the outer scheme ever recorded the value of what it held, so it never appears in the first figure. A layerOne charge running against one pool of assets. Where a scheme holds another scheme, each pool carries its own charge, and the two are recorded in different places. in this context means one charge against one pool, recorded in one place.
An exchange traded fund held directly carries one. Its own charge runs against its own assets and there is no scheme sitting on top of it taking a second cut. On layers alone, that is a genuine and structural difference and it is not close.
One qualification keeps the layer count honest, and dropping it turns a true statement into a misleading one: an exchange traded fund also carries whatever the holder gave up in the gap between price and value. The gap appears in no expense ratio anywhere, and it is not zero. It is not a layer, because a layer is a charge against a pool and this is not charged against any pool. The money went from one holder to another and never entered the scheme at all. But it left the holder's pocket exactly as a charge would.
| What is being counted | Fund of funds | Exchange traded fund held directly |
|---|---|---|
| Charges against a pool of assets | Two, the outer scheme's and those inside what it holds | One, its own |
| Where the second one is recorded | Inside the value of the units the outer scheme holds | Not applicable, there is no second one |
| Amount that leaves the holder but charges no pool | None from this source | The gap between price and value, on every transaction |
| Appears in the printed expense ratio | The outer charge only | The scheme's charge only |
Read the last row twice. The error set out further down lives entirely in that row. Each printed figure covers exactly one of the things that costs the holder money, and in each case it is a different one that is left out.
An exchange traded fund's expense ratio is far below a fund of funds' printed figure. Is it cheaper to hold?
What does a holder need in place before they can hold each?
Exchange traded units are held in dematerialisedHeld as an electronic record in an account rather than as a paper certificate. form and transacted through an exchange, so a holder needs the arrangements that go with that. Dematerialised holding is described by the two depositories, National Securities Depository Limited (NSDL) at nsdl.co.in and Central Depository Services (India) Limited (CDSL) at cdslindia.com. Both set out the requirements, the charges and the providers, and both revise them. A fund of funds is bought and sold with the scheme in the ordinary way, so it needs nothing beyond what any scheme needs.
The requirement is a real difference in reach rather than a matter of preference, and it is the criterion most often waved away as trivial by people who already have the arrangements. For somebody who already deals on an exchange, the requirement costs nothing and feels like no difference at all. For somebody who does not, it is not a small inconvenience sitting between the holder and the holding; it is the whole distance. A household that has bought schemes for fifteen years through an ordinary application has met none of it.
The everyday version: two shops sell the same rice. One takes cash at the counter. The other only accepts payment from an account the customer must first open, verify and keep. For a customer who has that account the two shops are identical. For a customer who does not, one shop is open and the other is a form.
A holder has no exchange arrangements at all. Which structure is still reachable for them?
How do units come into existence in each?
In a fund of funds, exactly as in any scheme. Ordinary subscription creates units and ordinary redemption cancels them. Every application is dealt with the scheme itself, so the count of units in existence rises and falls with what holders do, one application at a time.
In an exchange traded fund, the two things are separated. Units are created and cancelled in large blocks dealt directly with the scheme, in a mechanism known as block dealingThe arrangement by which units of an exchange traded scheme are brought into existence or extinguished in large lots dealt directly with the scheme, rather than one holder at a time.. Ordinary trading on the exchange does none of that; it moves units that already exist from one holder to another. Trading and creation are two different activities, happening in two different places with two different counterparties. A busy morning of trading can move an enormous number of units without a single new unit coming into existence.
Block dealing is covered separately. Only the consequence matters for the comparison: on one side a holder's transaction changes the size of the scheme, and on the other side it usually does not.
A thousand people buy exchange traded units on an exchange in one morning. How many new units were created?
What can go wrong, and are the two kinds of risk comparable?
Risk is where the comparison is decided. On the fund of funds side, the second layer is certain in direction and unknown in size to the reader. The second layer only ever subtracts. It never once adds. The size of it sits inside the value of the units the outer scheme holds rather than on the outer scheme's own printed figure, so the reader usually cannot see it.
On the exchange traded side, the price received can sit above or below the value of what the units represent. So the amount the holder gives up is uncertain in both direction and size. The gap might be a cost. It might be a windfall. Averaged over many transactions and many holders it is neither reliably one nor reliably the other, and no ratio anywhere records it.
The two structures carry different kinds of uncertainty rather than different amounts of the same thing, and that distinction is what stops the comparison collapsing into a ranking. A subtraction of unknown size and a two-sided gap of unknown size cannot be netted against each other, added, or ordered. The fund of funds layer runs continuously while the gap is paid only when a transaction happens, so netting them would need both sizes and the number of times the holder transacts. Neither of those is in this record.
Here are the six criteria in one place, in the order they were taken, so the pattern underneath them is visible at a glance rather than reconstructed from memory.
Which of the two structures carries the more uncertain cost?
Can this comparison be run on real numbers?
On one side yes, exactly. On the other side no.
The arithmetic on the fund of funds side is exact, so take that side first. The Girnar Broad Market Index Fund returned 12.12 per cent net for the stated year, charging 0.20 per cent of its own assets. The unnamed broad index it follows returned 12.40 per cent for the same year, and an index carries no charges at all because nobody pays anything to hold one. The gap of 0.28 points between the two is the tracker's own business, and tracking shortfall is taken apart properly under its own heading elsewhere. What matters now is the 12.12 per cent, the figure a holder of that scheme actually received after everything.
Now put an outer scheme on top of it. A fund of funds holding that tracker records the value of the tracker's units. The recorded value is already net of the tracker's charge, and the outer scheme then runs its own charge against its own assets. So the holder reaching the tracker through an outer scheme received 12.12 per cent less the outer charge, for the same year. The second layer subtracts precisely its own size, no more and no less. Nothing else in the comparison can be stated as an exact identity.
How large is the outer charge? No outer charge has been fixed anywhere above, so the figure below runs three arbitrary outer charges to show the shape rather than asserting any single number. Each one is a point chosen to illustrate, not a scheme's actual term.
A build that reconciles twice can be trusted, so check the same build from the other end. The distance from the costless index at 12.40 per cent down to the holder is the tracker's own 0.28 point gap plus the outer charge. At an outer charge of 0.25 that is 0.53 points, and 12.40 less 0.53 is 11.87. The ladder shows exactly that. At 0.35 it is 0.63 points and 11.77. At 0.50 it is 0.78 points and 11.62. Three rows, two routes each, six agreements.
Now the exchange traded side, and here the honest output is a list of absences rather than a number. Every scheme worked through above is an ordinary index or equity scheme, and not one of them is exchange traded. Not a traded price, not a gap against value, not a spread, not a block size. The Girnar Broad Market Index Fund is an ordinary index scheme rather than an exchange traded one, so none of its figures may be quietly reused as though they belonged to an exchange traded scheme.
Set the two sentences side by side and the comparison is genuinely useful even though it produces no winner. The fund of funds holder gives up an amount that is certain in direction and unknown in size to them. The exchange traded holder gives up an amount that may be positive or negative, is unknown in size, and appears in no expense ratio anywhere. One is a subtraction. The other is a distribution around zero. A reader who wants to know which is cheaper is asking a question that needs the missing figures and their own holding period, and neither of those is here, so the ranking cannot be made.
One more caution about the numbers that are here. The 12.12 per cent and the 12.40 per cent are one year on one scheme against one index, both invented for teaching. The two figures are not evidence about tracking, about active management, or about either structure compared here. Their whole job is to make the two-layer subtraction concrete, and they carry no weight beyond that.
What was computed for the exchange traded side of the comparison?
Can a fund of funds hold an exchange traded fund?
Why is this not a straight either-or choice?
Because the structures compose. A fund of funds can hold an exchange traded fund, and that is a common enough arrangement to matter rather than a curiosity. The arrangement puts the exchange traded portfolio inside a scheme that a holder can buy without an exchange at all, through the ordinary subscription process, with no dematerialised holding required of them. And it does that by adding the second layer back.
The two structures are not two answers to one question. Both are components, and either can sit inside the other, so a reader looking for a winner between them has misread the question. The comparison states what each component does. Whether one, the other, or one inside the other suits a particular situation is a different question, covered separately.
Who reaches for this comparison on a working day, and what can they actually conclude?
Three people, with three different questions, and none of them gets a ranking out of it. Sohail Merchant, who heads operations, needs the structural facts rather than the arithmetic: where a transaction lands, who the counterparty is, whose records change, and which of the two produces a subscription his registrar and transfer agent must process. For his purposes the exchange traded side is largely somebody else's process, and that is precisely the point he has to design around.
An analyst comparing two cost figures wants to know whether the comparison is available at all. The answer above is that it is not, as printed, and the useful output is knowing why: each figure is complete about a different thing. An analyst who accepts that will go looking for the two missing pieces before ranking anything, and an analyst who does not will publish a ranking that cannot be signed.
Reach is binary and cost is a matter of size, so a household choosing how to reach a spread of holdings uses the reach criterion first and the cost criteria second. If there are no exchange arrangements, one structure is available today and the other is a set of forms away. Reach is a fact about their situation, not a judgement about either structure.
None of the three can say which structure costs less. Doing so requires the traded price, the gap against value, the outer charge, and how often the holder transacts. Not one of those four has been fixed anywhere above.
The error that gets made, and what it costs
A reader compares the expense ratio of an exchange traded fund with the printed figure of a fund of funds, finds the first much lower, and treats the difference as the whole cost difference. Two things have been left out, and crucially they are different things on the two sides.
The fund of funds figure is missing its inner layer. The inner layer sits inside the value of the units the outer scheme holds and never appears on the outer scheme's own ratio. So the true cost on that side is higher than the printed figure, and by an amount the printed figure does not disclose. The exchange traded figure is missing the gap between the price the holder actually transacted at and the value of what the units represent. The gap is a real amount of money and it appears in no ratio anywhere. So the true cost on that side may be higher or lower than the printed figure, depending on which way the gap ran.
Who makes this error: anyone comparing two printed percentages. Two printed percentages invite exactly that. The cost: a structure chosen on a comparison in which both figures were incomplete, in opposite directions, so the reader cannot even say which way their error ran. An ordinary mistake at least has a sign, and a mistake without one is worse.
The fix is one habit. Before comparing two cost figures, ask of each what it leaves out. If the answer is the same for both, the comparison stands. If the answer differs between them, say the comparison is not available rather than making it anyway.
Who decides the conditions attaching to either structure?
SEBI does. The rules on what either structure may hold, what must be disclosed about it, what conditions attach to block dealing, and where each sits in the categorisation of schemes are all set by SEBI, and rules of that kind are revised. The current position is read at sebi.gov.in on the day it is needed. Industry level disclosure and the distributor framework sit with the Association of Mutual Funds in India (AMFI) at amfiindia.com. AMFI publishes rather than decides.
The dematerialised holding of exchange traded units, and everything that goes with it, is described by the depositories, NSDL at nsdl.co.in and CDSL at cdslindia.com. The requirements, the charges and the providers are read there.
Tax is covered separately. How either structure is treated depends on what the vehicle holds. The classifications, holding periods and rates all sit in tax law, and tax law moves. The tax authority at incometaxindia.gov.in is where the current position is read, and a wrong number in this area would be worse than a wrong one anywhere else.
Which of the two structures is put forward as the better one?
References
| Source | Document | Where |
|---|---|---|
| Securities and Exchange Board of India | The rules governing what either structure may hold, what must be disclosed about it, the conditions attaching to dealing in blocks, and where each sits in the categorisation of schemes | sebi.gov.in |
| Association of Mutual Funds in India | Industry level disclosure of scheme charges and of the distributor framework, published by an industry association rather than made into rules | amfiindia.com |
| National Securities Depository Limited | Where the dematerialised holding of units is described, along with the requirements, the charges and the providers that go with it | nsdl.co.in |
| Central Depository Services (India) Limited | Where the dematerialised holding of units is described, along with the account arrangements a holder needs before dealing on an exchange | cdslindia.com |
| The tax authority | Where the tax treatment of a pooled vehicle is set out. The classifications, holding periods, rates and thresholds are read there | incometaxindia.gov.in |
Girnar Asset Management Limited, the Girnar Broad Market Index Fund, 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.