CAGR and XIRR — Which One Answers Your Question

Neha has two numbers and they disagree

Neha has been investing monthly for three years. Her app shows one return figure, the fund's own page shows another, and they are not close. Neither is wrong. They are answers to different questions.

The fund's figure describes what happened to a single sum left in from the start. Hers describes what happened to money that arrived in instalments, each of which has been invested for a different length of time. Those are genuinely different quantities and there is no reason for them to match.

CAGR: one sum, two dates

CAGR — compound annual growth rate — answers: if I had put in one amount at the start and taken it out at the end, what single yearly rate would explain the change?

It needs exactly three things: a starting value, an ending value and a length of time. It ignores everything in between, including how bumpy the path was, and that is deliberate — the question it answers has nothing to do with the path.

The critical constraint is that CAGR assumes no money went in or came out during the period. If any did, the calculation is being asked to describe a situation it does not model, and it will return a number confidently anyway.

CAGR is the right measure for a fund's own performance, for an index, or for a single lump you invested and left alone.

XIRR: many amounts, many dates

XIRR answers a harder question: given money going in and coming out on various dates, what single yearly rate, applied to each amount for the time it was actually invested, accounts for where I have ended up?

It takes every cash flow with its date, and finds the rate that makes them all balance against your current value. There is no formula for it — the answer is found by trying rates until one fits, which is why it lives in a spreadsheet function rather than in your head.

Because it weights each amount by how long it was invested, XIRR is the right measure whenever the amounts and dates are irregular: a monthly investment plan, a plan you paused, a portfolio you added to when you had spare money, or one you have partially redeemed.

The rule for choosing

Did money go in or out during the period? If yes, XIRR. If no, either works and they agree.

That is the whole of it. Everything else is consequence.

One consequence worth naming: a fund's published CAGR and your XIRR in that fund are not comparable, and finding them different tells you nothing about the fund. If you began investing monthly three years ago, your money has been invested for an average of well under three years, so your figure will differ from a three-year CAGR even if the fund did exactly what it says. Comparing the two and concluding you have underperformed is a mistake about arithmetic, not a finding about the fund.

Three ways people go wrong

Dividing total gain by total invested. If you have put in a sum over five years and it is now worth more, the gain expressed as a percentage of the amount contributed is not an annual rate and is not comparable to one. It ignores time entirely — the same figure could describe a fine result over three years or a poor one over ten. This is the most common error and it is usually made by people being careful.

Annualising it by dividing by the number of years. A total return of some size over five years is not that size divided by five per year, because returns compound. That mistake goes in the opposite direction from the previous one, and people sometimes make both at once.

Comparing an XIRR against a CAGR. Covered above, and worth repeating because fund marketing and app dashboards routinely place the two side by side without saying which is which.

What neither of them tells you

Both compress a period into one number, and both therefore hide the same three things.

They hide the path. A steady result and a terrifying one that recovered can produce the same rate, and they were not the same experience — the distinction that why good returns do not make a portfolio safe takes apart.

They hide the period's dependence on its own dates. Any single rate is one draw from a distribution of possible periods, and moving the start date moves the answer, sometimes a great deal. That is the subject of rolling against point-to-point returns.

And they hide inflation. Both produce a rate in money terms, and money terms are not what you buy things with — nominal return, real return and purchasing power does that conversion.

What to take away

Ask which situation you are in before choosing the measure. Money going in and out means XIRR; a single sum left alone means CAGR.

When you are shown a return figure by anybody, the useful question is which of the two it is and over which dates — and if the answer is a percentage that turns out to be total gain over total invested, you have not been shown a rate at all.