Why a Fund Holds Cash, and What That Cash Really Costs

A scheme holds cash because money arrives and leaves on days that have nothing to do with when the manager wants to deal. Cash does not follow the market, so it drags the return down when the market rises and cushions it when the market falls. The cost is real, it changes sign, and it is rarely separable from the other reasons a scheme falls behind.

Most readers meet this idea backwards. A reader sees a line on a disclosure saying a scheme was holding money rather than shares, and reads it as a decision: the manager thought the market was expensive, so the manager stepped aside. Sometimes a manager does think that. But the ordinary reason a pooled scheme is holding money on any given day is that money moved in or out on a day nobody chose, and the arithmetic of what that money did to the return is identical whatever the reason was.

A single asset manager carries this guide from its first block to its last. 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. Dividing those net assets by the 120.00 crore units in issue makes one unit worth Rs 35.00 exactly. Girnar Asset Management also operates the Girnar Broad Market Index Fund, a tracker following a broad index that is not named here. Kalyani Bhagat manages the equity scheme and Sohail Merchant heads operations.

Four points are settled elsewhere and are carried into this guide as they stand. A holdings disclosure is a picture of one date rather than a description of a period. Money arrives and leaves an open ended scheme continuously. Continuous dealing is what open ended means. A scheme return that has been published already carries the running charge inside it, so nothing further comes off it later on. And where a tracker falls short of the index it follows, that shortfall splits into a charge and a leftover.

What is a cash position in a fund?

A cash position is the part of a scheme's assets held as money rather than as the assets the scheme exists to hold. In an equity scheme that means rupees sitting in a bank account or parked in very short dated instruments, rather than shares. The cash positionThe share of a scheme's assets held as money or near money rather than as the assets the scheme was set up to hold. is normally written as a percentage of the scheme's net assets. Every other weight on a disclosure uses that same base.

Cash is a holding like any other and is valued daily like any other, and exactly one property separates it from the rest: it does not move with the market the scheme is exposed to. Everything that follows rests on that one property. The custodian values the shares each day and values the money each day, and both go into the same net assets figure that is divided by the units in issue. Nothing about cash sits outside the machinery. The difference is that when the market a scheme is exposed to moves, the shares move with it and the money does not.

Think about a shop for a moment. A cloth merchant has stock on the shelves and float in the drawer. If the wholesale price of cloth rises overnight, the shelves are worth more in the morning and the drawer is worth exactly what it was worth yesterday. The merchant did not make a call on cloth prices by keeping float in the drawer. The float was there because customers pay in notes and suppliers get paid on Thursdays. The float still changed what the morning was worth, and it would have changed it the other way if cloth had fallen.

Why is that money sitting there at all?

There are four answers, not one, and keeping them separate is what stops a reader turning a cash weight into a forecast. The first is subscriptionA purchase of units in a scheme, which brings new money into the pool. money that has arrived and not yet been put to work. Somebody bought units on Tuesday; the money reached the scheme; the holdings it will become have not been bought yet. Until they are, that money is cash.

The second is money held against redemptionA holder selling units back to the scheme, which takes money out of the pool. requests that have to be paid. Units are being sold back and the scheme has to find the rupees. The third is the gap between a sale and a purchase: a holding was sold, the proceeds landed, and the next holding has not been bought yet, so for a stretch of days the proceeds are money. The fourth is the least obvious and the most interesting. Sometimes the intended deploymentPutting money that has arrived in a scheme into the holdings it is meant to be invested in. is large enough that buying it all at once would push the price up against the scheme, so it is done in pieces across days and the remainder waits as money in the meantime.

Not one of those four is a view about the market, and reading a cash weight as though it were is the commonest misreading of a disclosure there is. The everyday version is the stall outside a college gate. The owner keeps float in the drawer because customers turn up when they turn up and change has to be given, not because the owner has a theory about tomorrow. Ask the owner what the float predicts and the question makes no sense. The float is the cost of being open to whoever walks past.

What share of a market move does a scheme actually take?

All of the move on the part that is exposed, and none of it on the part that is money. The whole mechanism is that split, and it needs no market figure to state. If a scheme holds a weight of w per cent in cash, then 100 less w per cent of its assets are exposed, so the scheme takes 100 less w per cent of whatever the market did. The name for that fraction is captureThe share of a market move that a portfolio actually takes, expressed as a fraction of the move rather than as a return., and a cash weight of four per cent means a capture of ninety six per cent.

A proportion holds whatever the market did and a return does not, so write it as a proportion and never as a return. Writing it as a proportion saves a reader from the commonest sloppy thinking about cash dragThe part of a market move a portfolio does not take because some of its assets are held as money.. Drag is not a fixed number of percentage points a year. The weight fixes the proportion; the market fixes the size. The same four per cent weight costs a scheme a great deal in points in a year the market ran hard, and almost nothing in points in a year the market barely moved, and it was the identical position both times.

Household version. Consider four rupees of every hundred kept in a tin under the bed and ninety six put into a chit that moves with the price of gold. If gold rises, the holder took ninety six per cent of the rise. If gold doubles, ninety six per cent of a doubling is a lot of money not missed and four per cent of a doubling is a lot that was. If gold moves half a per cent, the same four rupees cost almost nothing anyone would notice. The tin did not change. The world did.

What happens to that same money when the market falls?

The identical arithmetic runs the other way. The exposed part is exposed in both directions and the money part is inert in both, so a scheme that takes 100 less w per cent of a rise also takes 100 less w per cent of a fall. A four per cent cash weight that cost four rupees of every hundred the market gained will save four rupees of every hundred the market lost. Nothing about the position changed. The world changed sign.

So cash is a cost only in a rising market and a cushion in a falling one, and the same holding is both of those things depending entirely on something the scheme does not control. Notice how uncomfortable that is for the language people use. A scheme is praised for prudent cash in a bad year and criticised for cash drag in a good one, and in a great many cases it was the same operational weight sitting there for the same operational reasons in both years.

Now the harder point, and it is the one worth carrying away. Which of the two it turned out to be is knowable only after the period has run, so describing a cash position as a cost before the period has ended is a prediction rather than a measurement. Last year's drag can be measured once last year is over. The current year's drag cannot be measured until the current year is over, and any figure written for it now is a forecast with a measurement's clothes on. The distinction between a measurement and a forecast is small in words and very large in practice. Forecasts get written into reports as though they were sums.

Can the cost of a cash position be pulled out of a shortfall?

Usually not, and this limit matters more than anything else in the subject. When a scheme ends a period behind some yardstick, several causes have been acting at the same time and only their total is observed. The gap is visible. The gap broken into parts is not. Nobody was measuring the parts separately as the period ran. Cash is one of those causes. Assigning it a share requires information that the arithmetic does not contain.

One figure in this record carries a cash-related component, and it belongs to the tracker rather than to the equity scheme, so it gets labelled carefully. Across the single year this record covers, the Girnar Broad Market Index Fund came in at 12.12 per cent net. The broad index it follows came in at 12.40 per cent across that same year, and it carries no cost whatever. Nobody can hold an index, so nobody pays a paisa for holding one. A tracker charging 0.20 per cent a year that had tracked exactly would have returned 12.20 per cent. So the tracking differenceHow far a tracker's own result sits from the result of the index it is following, across one period. is 12.12 per cent net less 12.40 per cent costless, or minus 0.28 points. Both bases are named in that subtraction because the charge is sitting inside the gap.

Split that 0.28 points and 0.20 of it is the charge. The remaining 0.08 points carry cash holdings, the timing of money coming and going, and what it costs to follow an index whenever the index itself is changed, all at once and never apart. Read the last four words slowly. The residualWhat is left of a measured gap once the parts that can be named and priced have been taken out of it. is one number covering three causes. Cash is inside it. Nothing anywhere on this platform separates it from its two companions, and any source that supplies a number for the cash part has invented one.

Here is the check that makes the residual worth believing. If the 0.08 were simply the charge under another name, then adding the charge back would make it vanish. It does not. Take the tracker's 12.12 per cent net and add its 0.20 per cent charge back on: that gives about 12.32 per cent on a gross equivalent basis for the Girnar Broad Market Index Fund. Set that against the costless index at 12.40 per cent and about 0.08 points are still standing. The residual survives the charge being restored. A residual that survives is a separate cause rather than the charge wearing a different label.

One honesty note on that addback. Adding a percentage back on is additive. A charge accrues daily, and daily accrual is multiplicative. Backing the charge out properly puts the gross equivalent a little above 12.32 per cent and leaves a residual a little under 0.08 points. The direction of that rounding matters: the additive route understates the gross equivalent, so it slightly overstates the residual. The record fixes the split at 0.20 and 0.08. The finding that survives both routes is that the residual does not disappear.

Here is the whole build in one place, on the Girnar Broad Market Index Fund, for the one stated year, with the check run backwards at the bottom so it can be seen close.

Two routes that share no arithmetic reach the same 0.08 points, and the backwards check closes on it as well. Agreement across three routes is what makes the figure checked rather than merely plausible. Note the word about on the last row and nowhere else: the first two routes are exact on the record's own figures, and only the addback route carries an approximation, so only that row is written as an approximation.

Why do five figures near twelve per cent mean five different things?

Because a number is not a measurement until what it measures and on what basis are stated. Five figures here all read close to twelve per cent, and letting any two of them shake hands manufactures a finding out of arithmetic. Each stands on its own. The Girnar Large Cap Equity Fund's stated benchmark returned 12.1 per cent for the stated year, and it is costless. The Girnar Broad Market Index Fund itself came in at 12.12 per cent net. A perfectly tracking version of that fund would have shown 12.20 per cent. Its approximate gross equivalent is about 12.32 per cent. The broad index it follows returned 12.40 per cent, and that is costless too.

The stated benchmark of the equity scheme and the broad index behind the tracker are two different measuring sticks, and nothing in this record ever suggested otherwise, so 13.4 per cent net never gets held up against 12.40 per cent and 12.12 per cent net never gets held up against 12.1 per cent. A benchmark for a large capitalisation scheme and a broad market index measure different segments of the world. How near or how far their two figures land in any particular year decides nothing at all, and the rule reads identically whether the pair looks alike or looks nothing like each other.

A near miss sits one step along from that rule and it catches readers in exactly the way the rule exists to stop. Look at where the Girnar Broad Market Index Fund's 12.12 per cent net lands: two hundredths of a point away from that 12.1 per cent stated benchmark. One of those two is what a scheme actually delivered once its own charge had come out of it; the other is a free yardstick built for a different slice of the market. Two mismatches are piled on each other rather than one, and their reading alike is a coincidence carrying nothing with it.

There is a second near miss and it deserves saying out loud. The tracker's approximate gross equivalent of about 12.32 per cent is arrived at by adding a charge back on. A discretionary mandate held in this platform's portfolio material also reads 12.32 per cent, that one is net and was reached by taking fees off, and the match between the two 12.32 figures is a coincidence and nothing else. Set them side by side and the equality carries no information at all: unrelated vehicles, records that never touch, and bases pointing opposite ways. One figure was reached by putting a charge back on and the other by taking fees off. The sums produce that match on their own and cannot be nudged aside to avoid it, so it stands as a coincidence and nothing more.

What does a disclosed cash weight actually show?

A disclosed cash weight shows the state on one date. The state on one date is genuine information and should not be dismissed: somebody counted, and on that date the scheme held that much money. Almost everything else an analyst would want is missing. An average needs many dates and the disclosure supplies one, so the weight does not give the average across the period. A weight carries no reason attached to it, so the disclosure does not say why the money was sitting there. And a cost needs the market's move as well as the weight, so the cost of the position is missing too.

A cash weight is a picture of one date and it inherits every limit a picture of one date carries. Holdings disclosure settles that point already, and the cash line is one more line of the same disclosure. Two schemes showing the identical weight on the identical date may have run completely different weights on the days either side, and a snapshotA record of what something looked like on one date, which says nothing about the dates before or after it. cannot distinguish between them. The reader who most needs to hear this is the one who has just found the weight and feels well informed.

Why is no cash weight given for the scheme?

Because there isn't one to give. The record for the Girnar Large Cap Equity Fund holds net assets, units in issue, the number of folios, the expense ratio and one year's return. There are no portfolio holdings in it, no weights by sector or by stock, and no cash position, for any date at all. The same is true of the Girnar Broad Market Index Fund. There is no second scheme in the range and no second year for either of them.

The capture arithmetic works as a proportion and needs no weight at all, and the one cash-related figure the record does hold can be pointed at without pretending that figure belongs to the equity scheme. The one cash-related figure is the 0.08 points of residual on the Girnar Broad Market Index Fund for the one stated year, and it belongs to the tracker and to the index the tracker follows. The residual is not the equity scheme's, it is not the equity scheme's stated benchmark's, and it does not travel.

How small the sample is belongs right beside the figure itself, not tucked away in a closing caution, so here it sits. The record holds a single year covering a single scheme. Behind that year lie no further years, no series broken down by month, no companion scheme on the shelf and nothing real at any point. One year covering one scheme will not stretch into an annual rate for anything, will not reach forward or back into years nobody measured, has no business standing beside a real vehicle of any kind, and settles nothing whatever about what picking holdings or following an index produces. Two things hold together here: the sums close to the last decimal, and what they reach beyond themselves is very close to nothing.

How should a shortfall with several causes be reported?

As a total, with the causes named beside it, and with no figure set against any one of them. A total with its causes named is a smaller output than most readers want, and it is the only one the arithmetic supports. AttributionSplitting a measured result into the separate causes that produced it, so each cause carries its own share. is a real technique and it works where the parts were measured separately as the period ran. Where only the total was ever observed, there is nothing to attribute with, and dressing an assumption in the language of measurement does not create the missing information.

Girnar Asset Management Limited, the Girnar Large Cap Equity Fund, the Girnar Broad Market Index Fund, the index that tracker follows, the equity scheme's stated benchmark, Kalyani Bhagat and Sohail Merchant are invented.
Educational material. Not advice on any investment, tax, budget or market position.