How to Read a Fund Return and What Each Measure Hides

A fund return states what one rupee became between two dates. Every published scheme return is a net figure, struck from values that already carry the charge, so nothing is deducted from it afterwards. Every measure of it assumes something: a start date, an end date, a window length and a yardstick. Read each measure with its assumption named.

Here is what sits underneath that. A return is not a property of a scheme. A return is a property of a record: two values, two dates, and a convention for turning the pair into a proportion. Change any one of those three and the number changes while the scheme carries on doing exactly what it was doing. Four measures exist rather than one, and each conceals something the others do not.

One scheme runs through this guide from end to end. Girnar Asset Management Limited, an invented fund house, operates the Girnar Large Cap Equity Fund, an open ended equity scheme. Its net assets are Rs 4,200 crore and it has 120.00 crore units outstanding. Divide the first by the second and one unit is worth Rs 35.00 exactly. Rs 35.00 is the net asset valueThe value of one unit, worked out as what the scheme holds less what it owes, divided by the units in issue. that every measure below starts from. The scheme is held across 3,80,000 folios, so the average holding works out at about Rs 1,10,526/-, and it carries an expense ratioThe running charge against a scheme's assets, quoted as a percentage of assets a year rather than billed to anyone. of 1.65 per cent a year. Kalyani Bhagat manages the portfolio and Sohail Merchant heads operations.

Four things sit underneath a published return without being part of it. How the net asset value is struck, and which day's value an order gets, are covered separately. How the expense ratio reaches the daily value is covered separately, and the one thing to carry forward from it is that the charge is taken against the scheme's assets every day rather than billed to a holder. A scheme's legal shape, and who the trustee company protects, are covered separately. Compounding is arithmetic already in hand. Reading a return starts one step after all of that, at the moment a figure is published and somebody has to decide what it says.

What is a fund return actually measuring?

A return compares a value at one date with a value at another date and expresses the change as a proportion of the first. The whole measure is that one division. Take the value at the end, subtract the value at the start, divide by the value at the start, multiply by a hundred. If a unit stood at Rs 35.00 and finished at Rs 39.69, the change is Rs 4.69. Rs 4.69 divided by Rs 35.00 is 0.134, or 13.4 per cent.

Notice how much of that arithmetic is a choice rather than a fact. The scheme did not choose the two dates. Somebody did. The answer is decided as much by the two dates as by anything the scheme did, and nobody chooses the dates by accident. A figure struck from a low starting value looks generous. The same record measured from a high starting value looks poor. Neither is a lie and neither is the truth on its own; each is one reading of one record, and the reading that gets shown is the reading somebody decided to show.

Here is the everyday version. A shop measures this Diwali's takings against last Diwali's takings. The two dates are alike, so the comparison is a real and useful one. The same shop measured from a slow Tuesday to a busy Saturday produces a much larger number about the same shop, and nothing about the shop has changed between the two statements. The shopkeeper is not being dishonest by choosing Diwali. But a figure handed over without its dates gives no way to tell which comparison is in view, and that is the position a reader is in every time a return arrives without its period attached.

How Mutual Fund Returns Are Reported: on what basis does the published figure stand?

On a net basis, always, and that single word changes how every other figure here has to be read. A scheme return is computed from net asset values. Net asset values are struck after the day's expenses have been set against the scheme's assets. So the charge is already inside the value at the start of the period and already inside the value at the end of it. Both ends of the calculation carry it, and the proportion that falls out of them is therefore a net returnA return figure with the charge already taken out, so no further deduction for the charge belongs on it..

Nothing is deducted from a published scheme return afterwards, and a reader who subtracts the expense ratio from it has taken the same charge twice. Taking the same charge twice is not a subtlety. The double deduction is the difference between reporting 13.4 per cent and reporting 11.75 per cent for the same year on the same units, and the second of those figures is simply wrong by the full 1.65 points.

The mirror image of a net figure is a gross returnA return figure quoted before costs, so a separate fee still has to come off it before it describes what anybody kept.. A gross figure is quoted before costs, so a deduction is still waiting for it. Both kinds circulate. A scheme's published figure is net. A fee-charging arrangement that bills separately is usually quoted gross, and its costs come off afterwards. The number itself does not announce which it is, and that is exactly why the label has to travel with it in the same sentence.

Point-to-Point Return: what does one date to another date leave out?

Everything in between. A point-to-point returnA return measured from one stated date to one other stated date, using only the two values at those dates. reads the value at the start, reads the value at the end, and ignores every value that occurred between them. A point-to-point figure is the most commonly published measure and the easiest to compute, and the price of that simplicity is that the shape of the period is discarded entirely.

A scheme that rose steadily and a scheme that fell hard and then recovered can produce the identical point-to-point figure, and the figure holds nothing that could tell them apart. The identical figure is not a defect anyone can fix by computing more carefully. Two numbers went in, so only what two numbers can say comes out.

The consequence for a person is larger than it sounds. Nobody experiences a point-to-point return. A holder experiences the values in between: the statement in month four that showed less than was put in, the decision about whether to stay, the conversation at home about whether this was a mistake. The steady path and the falling path deliver the same arithmetic and completely different years, and the reason the next measure exists at all is that a great many readers, having been shown one figure, assumed the steady path.

Rolling Return: what does repeating the measurement fix, and what does it cost?

A rolling returnThe same measurement repeated from many different start dates across a longer record, producing a spread of readings rather than one. takes the same measurement and repeats it from every available start date inside a longer record. Instead of one twelve month reading struck from one chosen January, there is a twelve month reading starting in January, another starting in February, another starting in March, and so on to the end of the record. A rolling measurement produces a spread of readings rather than a single number.

The spread removes the choice of start date from whoever is presenting the figure, and that is the single most useful thing a rolling return does. Nobody can present a flattering January when the reading from every other month is sitting beside it. The spread also shows something a single figure never can: how much the answer moved depending only on the month the measurement began in.

The exchange is overlap, and a hundred rolling readings are nowhere near a hundred independent observations. Two neighbouring twelve month windows share eleven of their twelve months. Neighbouring windows are not independent observations; they are one observation looked at twice with a one month shift. A record of 111 months yields 100 twelve month windows if one starts in every month, and that same record contains only nine twelve month stretches that share nothing at all with each other, with three months left over. Counting the hundred as a hundred overstates what is there by roughly a factor of eleven. The detailed comparison of rolling against point-to-point is covered separately.

What is an absolute return, and when does it stop being useful?

An absolute returnThe whole change across a holding period, stated as one figure, with no yearly rate attached to it. is the whole change across a holding period, stated as one figure, with no yearly rate attached. If a holding grew by 13.4 per cent across the period it was held, the absolute return is 13.4 per cent, and the measure says nothing about how long the period was.

Over exactly one year an absolute return is the same number as the annual figure, and the two only separate when the period is not one year. The relationship is that simple, and it is why the stated year for the Girnar Large Cap Equity Fund produces 13.4 per cent on both measures. There is no cleverness hiding in the agreement; it is a consequence of the period, not of the scheme.

Over longer periods the absolute figure accumulates, so it becomes a larger number describing the same experience. Absolute return is therefore the measure most often reached for when somebody wants a figure to sound substantial: a decade of ordinary compounding produces an absolute figure that is arithmetically correct and, quoted without its period, wildly misleading about the pace of anything. The annualised version restates the same change as a yearly rate. A yearly rate is what makes periods of different lengths comparable at all. Absolute return over long periods, and why the gap between it and the annualised figure widens with time rather than with anything the scheme did, is worked out separately.

How Risk Measures Provide Context for Fund Returns: what does each one assume?

A return is one number about a period, and it says nothing whatever about the ride that produced it. The steady path and the falling path in the figure above end at the same place, and a reader handed only the endpoint has no way to know which one they lived through. A risk measure is the second number that attaches to the first: how much the value moved about across the period, how far it fell from its highest point at its worst, how much of its movement looked like the movement of the wider market it sits in.

Every risk measure assumes something, usually a period, sometimes a shape for how the values are distributed, sometimes a yardstick to be measured against, and a risk measure quoted without its assumption is decoration. A measure of how much a value moved about is a measure over a stated window; over a different window it is a different number. A measure of movement relative to the market is relative to a chosen market, and choosing a different one changes the answer. None of that makes the measures weak. The assumptions make them measurements. A measurement only means something once how it was taken is known.

The statistics themselves, how each one is constructed and what it does and does not capture, are built in the portfolio material and covered there. The narrower point is the one readers skip: a return and a risk measure sit beside each other rather than one inside the other, and reading the pair is the only way a single figure becomes a description of a year rather than a summary of two dates.

Why is a scheme return not comparable with a benchmark return as published?

Because the two figures are not on the same basis, and the difference is larger than the difference they appear to show. The Girnar Large Cap Equity Fund returned 13.4 per cent net for the stated year, computed from values that already carry its 1.65 per cent charge. Its stated benchmarkThe measuring stick a scheme states it will be read against, usually an index, which nobody can hold and nobody pays to hold. returned 12.1 per cent for the same year and carries no cost at all. An index is not something a person can hold, and nobody pays anything to hold one. Subtracting the second from the first leaves 1.3 percentage pointsThe unit for a difference between two percentages. A move from 12.1 per cent to 13.4 per cent is 1.3 percentage points., and that subtraction has set a net figure against a costless one.

Put both on one basis and the picture changes size. Adding the 1.65 per cent charge back on to the scheme's net figure gives about 15.05 per cent. Against 12.1 per cent that is about 2.95 points. The basis mismatch is not a caveat on the answer. Since 2.95 points is more than twice the 1.3 points the headline shows, the mismatch changes the answer by more than the answer itself. Dividing 2.95 by 1.3 gives about 2.27.

Say the approximation out loud rather than hiding it. Adding the ratio back is an arithmetic add-back, and the charge does not work that way: it accrues daily against an asset base that moves every day. A daily accrual against a moving base is multiplicative rather than additive. A stricter restatement divides instead of adding. 1.134 divided by 0.9835 is about 1.15302, so about 15.302 per cent, and against 12.1 per cent that is about 3.202 points. Both of those last two figures round downward at the third decimal from an exact value a shade above them. Even the stricter route still treats a daily accrual as one flat annual deduction, so it is not exact either. The two routes disagree by about 0.252 of a point, roughly a twelfth of the gap they are both trying to describe. Neither route ties exactly, and what survives both is the finding rather than the figure: the honest gap sits near three points, more than twice what the headline shows.

What does one stated year look like run through every measure?

Here is the whole record in one place, with the basis written on every line. Start with the scheme's scale so there is something to hold on to. Net assets of Rs 4,200 crore divided by 120.00 crore units gives Rs 35.00 a unit exactly. The charge of 1.65 per cent on Rs 4,200 crore is Rs 69.30 crore for the year. Across 365 days that is about Rs 18,98,630/- a day, or about Rs 0.0015822 per unit per day, and it comes out of the assets rather than being billed to anybody. Now the measures, all on that same single year.

A build that only works in one direction has not been checked, so the check also runs backwards. Rs 0.0015822 a unit a day across 120.00 crore units is about Rs 18,98,630/- a day, and across 365 days that is Rs 69.30 crore. Rs 69.30 crore is 1.65 per cent of Rs 4,200 crore. And on the returns: 12.1 per cent plus the published gap of 1.3 points returns to 13.4 per cent, exactly where the arithmetic on the published row started. Both directions close.

Two roundings go downward, and both are named where they appear. The stricter gross restatement is 15.302 and a fraction of a thousandth more, written as about 15.302 per cent. The stricter gap is 3.202 and the same fraction more, written as about 3.202 points. An approximate figure and an exact one are two different kinds of claim: the word about marks the approximate ones, and its absence marks the exact. Rs 35.00 is exact. Rs 69.30 crore is exact. 13.4 less 12.1 is exactly 1.3. The gross equivalents are not exact by either route, and they never will be. A daily accrual against a moving base cannot be undone by one annual operation.

What does one stated year on one scheme actually establish?

Almost nothing outside itself, and saying so is not modesty. The record behind these figures holds one year for one scheme. There is no second year, no month by month series, no other scheme in the range and no portfolio holdings. A single year cannot be annualised into anything longer, cannot be extended forward or backward, cannot be set beside any real fund or index, and is not evidence about what selecting holdings delivers or about what tracking an index delivers.

The arithmetic on it is exact and its reach is tiny, and both of those are true at once about the same figure. The pairing of an exact figure with a tiny reach is what to sit with. Rs 4,200 crore over 120.00 crore units really is Rs 35.00. 13.4 less 12.1 really is 1.3. Nothing about the precision of those statements makes them mean anything more than they say. Precision and reach are separate properties, and a well computed number with a one year record behind it is a well computed number with a one year record behind it.

A figure printed with nothing beside it makes a claim that was never written down, so every measure needs its limit attached in the same breath. The unwritten claim is the dangerous one. Nobody has to defend a sentence nobody wrote. Print 13.4 per cent beside 12.1 per cent with no basis and no sample size, and a reader will finish the sentence themselves, in the direction the arithmetic invites. The only reliable guard is to write the limit in the same block as the figure rather than parking it in a caution at the bottom that the reader has already decided to skip.

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