Absolute return and why it misleads

Six chapters on a question that sounds trivial and is not: what rate of return did I actually earn? Almost every number people quote about their investments answers a different question from the one they think.

What an absolute return is

The simplest measure there is. You put in ₹1,00,000, it is now worth ₹1,18,000, so your absolute return is 18%.

$\text{absolute return} = \frac{\text{value now} - \text{amount invested}}{\text{amount invested}}$

It is correct, and it contains no time. That is its one virtue and its one defect: 18% says nothing about whether it took two years or two decades.

Where it is the right figure

Over short and unequal periods. For something held three months, the absolute return is the honest statement. Annualising it requires pretending the next nine months will behave like the first three, and you have no evidence for that.

When the period is already understood. "Up 4% this month" is a complete statement in context.

For a single-period comparison. Two investments both held for exactly the same window can be compared on absolute return directly, and converting both to annual rates adds nothing.

Where it misleads

Comparison across different periods. This is the main failure, and it is pervasive. 18% and 11% cannot be ranked until you know over how long each was earned.

Long periods, where it flatters enormously. A 240% absolute return sounds extraordinary. Over twenty years it is about 6.3% a year — below many deposits. Anchoring and framing in the Behavioural finance subject explains why the big number lands harder, and chapter 2 here does the conversion.

Short periods, where annualising flatters even more. A 6% gain in three months annualises to roughly 26%. Nothing was earned at 26%; a three-month result was multiplied by four and relabelled.

What SEBI requires, and why it is the same point

The regulator has taken a position on exactly this, and it is a useful external check on the argument.

For mutual fund performance advertising, clause 14.2.1 of the Master Circular requires that performance "shall be advertised in terms of CAGR at least for the past 1 year, 3 years, 5 years and since inception", and that "Point-to-point returns on a standard investment of Rs. 10,000/- shall also be provided."

And below a year the rule inverts:

For a scheme which is in existence for more than 1 year, the returns given will be Compounded Annualised Returns and for scheme which is in existence for less than 1 year, the returns would be absolute returns since inception.

Read that twice, because it is the whole chapter in a regulation. Above a year, absolute returns are not sufficient and must be annualised. Below a year, annualising is not permitted and absolute is required.

The regulator is drawing exactly the line this chapter draws: annualising a period shorter than a year projects a result you do not have evidence for, and refusing to annualise a period longer than a year hides the rate. Each measure has a domain, and using it outside that domain is how performance gets oversold.

Two further requirements in the same clause are worth noticing, because they close other gaps: information must be computed from the last day of the month-end preceding the advertisement (so the start and end dates cannot be chosen for effect), and if the scheme was not managed by the same fund manager for the whole period, that must be disclosed in a footnote.

Working the problem

18% over two years against 11% over eleven months.

You cannot rank them as given. One figure covers 24 months and the other 11, so they are not the same kind of quantity. Ranking requires converting both to a common unit, and that means a rate per year.

The two-year investment. 18% over two years is an annual rate of $1.18^{1/2} - 1 = 8.6\%$ a year. This conversion is safe: you have two full years of evidence and you are dividing it, not extending it.

The eleven-month investment. 11% over eleven months annualises to $1.11^{12/11} - 1 = 12.0\%$ a year. But this conversion assumes the twelfth month resembles the first eleven, which is an assumption about the future rather than a restatement of the past.

So on the arithmetic, the second ranks higher — 12.0% against 8.6%. And the assumption required to say so is precisely the one SEBI forbids a fund to make, because the period is under a year.

What you had to assume, stated plainly: that the shorter investment's rate continues for the remaining month, that both returns are comparable despite covering different market conditions, and that neither figure needs adjusting for risk taken or for money added along the way. Chapters 4, 5 and 6 take those last three apart.

The honest answer to the ranking question is that the eleven-month figure is the stronger result and the weaker evidence, and eleven months is far too short to tell you anything about either investment. Which is the uncomfortable thing about return measurement: the number improves in reliability exactly as it becomes less exciting.

The point

An absolute return divides gain by amount invested and contains no time, making it the right figure over short or equal periods and useless for comparison across different ones. Long periods flatter it — 240% over twenty years is about 6.3% a year — and annualising short periods flatters more, since a 6% quarter becomes 26% by projection rather than by measurement. SEBI draws the line at a year in both directions: above a year returns must be annualised, below a year annualising is not permitted and absolute returns must be shown.