How to Compare Two Investments Fairly

Why accurate figures produce misleading comparisons

Kabir has stopped assuming that a misleading comparison contains a false number. Almost none of them do. The figures are computed correctly and the comparison is still worthless, because something differed between the two sides that was not the thing being compared.

There are five such things. Hold all five constant and a comparison means something. Let any one vary and it means nothing, however precise the arithmetic.

One: the same period

Two returns over different stretches of time are not comparable, full stop.

This is the most common failure and the easiest to miss, because the periods are often stated in small print and are often nearly the same — five years against "since inception", or two funds whose track records begin in different years. Nearly the same is not the same. On Indian data, moving a ten-year window across the record moves the annualised outcome enormously, which is what rolling against point-to-point returns measures.

The corollary is that whoever chose the period had a choice, and every choice produces a different figure. A period you were given is a period somebody selected.

Two: the same measure

A CAGR and an XIRR are answers to different questions and must not be placed side by side. Nor may a total gain over total invested, which is not an annual rate at all. The distinction is in CAGR and XIRR.

Two variants to watch. Annualised against cumulative — a cumulative figure over several years looks enormous next to an annual one, and they are sometimes printed in adjacent columns. And point-to-point against average of rolling periods, which are different statistics of the same data.

Three: the same benchmark, in the same version

If the comparison is against an index, it must be an index that matches what is actually held, and it must be the total return version — the one that includes dividends, since the fund receives them too.

Comparing a fund against a price index gives it a free head start that grows with the length of the period: price return against total return. Comparing it against an index of a different kind of company measures the mismatch rather than the manager, which is the argument in what alpha means.

Four: the same treatment of costs and tax

Costs have to be on the same side of the line for both. A return quoted after a fund's charges against one quoted before them is not a comparison; it is a subtraction applied to only one of the two. And the magnitude is not trivial — the share of an outcome an annual charge removes over a long holding period is set out in how investment fees reduce wealth.

Tax is the one most often left out entirely, and it can decide the answer. Two investments with identical returns are not equally good if one is taxed as it earns and the other only when sold, because the money not paid in tax keeps compounding meanwhile. That effect is large enough that our own evidence articles charge tax explicitly on both sides — see the method section of does buying the dip help.

Five: the same risk taken

A higher return earned by taking more risk is not a better result; it is a different trade. Judging the two on return alone rewards whoever took the most risk in a period that happened to reward it.

Adjusting for this is genuinely useful and less straightforward than it sounds, because the adjustment depends on which measure of risk goes in the denominator — risk-adjusted return. At minimum, ask what the worst fall was on each side over the same window.

Two questions that dispose of most of it

If you remember nothing else, these two catch the majority of unfair comparisons.

"Over exactly what dates?" — which catches failures one and, usually, two.

"After the same costs and taxes, against what?" — which catches three and four.

The fifth needs a moment's thought rather than a question, and it is the one to apply when a return looks unusually good: what was risked to get it, and would you have sat through that?

What a fair comparison still cannot do

It can establish what happened. It cannot establish what will happen — the comparison describes two periods that have ended, and the reasons that limits its usefulness for choosing are in why past performance is not enough to choose a fund.

It also cannot tell you which is right for you, since that depends on when you need the money and what else you own.

What to take away

Five things constant: period, measure, benchmark, costs and tax, risk. Any one varying and the comparison is about that variable rather than about the investments.

You will rarely be handed a comparison with all five held. That is not usually dishonesty — it is that the person assembling it had choices at every step and no reason to make the unflattering one. The questions are yours to ask.