XIRR or CAGR: Which Number Should Judge Your Portfolio

CAGR answers one question well: if I put in one amount, on one day, and took out another amount on a later day, what constant annual rate connects them? For a single lump sum that is exactly right.

For a SIP, a series of top-ups, or any portfolio you have added to and withdrawn from, CAGR is answering a question you did not ask, and the answer it gives is usually too flattering.

What each one actually computes

CAGR = (ending value ÷ beginning value)^1/n − 1.

Two values and a duration. It has no way to know that half your money arrived last year and has therefore been working for one year rather than ten.

XIRR is the discount rate at which the present value of every cash flow, each with its own date, comes to zero. Every contribution is weighted by how long it was actually invested.

The difference is entirely about time-weighting. CAGR treats all your money as if it arrived on day one. XIRR does not.

The same SIP, both ways

₹10,000 a month for five years. Total invested ₹6,00,000. Ending value ₹8,10,000.

Computed as CAGR: (8,10,000 ÷ 6,00,000)^1/5 − 1 = 6.19%.

That treats the whole ₹6,00,000 as if it were invested on day one for five years. It was not. The last instalment was invested for one month.

Computed as XIRR on the 60 dated instalments and the final redemption: approximately 12.0%.

Assumes 60 monthly contributions on the same date each month and a single redemption at the end of month 60.

Nearly twice the number. Both are arithmetically correct; only one answers the question “how well did my money do”.

Why the gap is so large

On a five-year SIP the average rupee has been invested for about two and a half years, not five. CAGR credits every rupee with five years of work, so to produce the observed ending value it has to conclude that the rate was low.

Think of it as a denominator problem. CAGR divides the growth across too much time, so the annual figure comes out small.

This is the direction people get wrong. The usual assumption is that CAGR flatters. For a SIP it does the opposite: it understates, and badly. CAGR flatters in a different situation, described next.

Where CAGR does flatter

Now run it the other way. You invest ₹1,00,000, it grows to ₹1,50,000 over three years, and then you add ₹9,00,000 and the portfolio returns 2% in year four.

CAGR on beginning and ending values ignores the ₹9,00,000 entirely as a cash flow and blends a period of excellent performance on a small base with poor performance on a large one. The headline looks respectable. The rupees tell a different story, because most of your money experienced the bad year.

XIRR weights that ₹9,00,000 properly and reports something much closer to your actual experience.

The rule underneath both cases: CAGR is wrong whenever cash flows are uneven, and the direction of the error depends on whether the money arrived early or late.

The distinction underneath: money-weighted and time-weighted

The two numbers belong to two different families, and knowing which family you are in explains every disagreement between them.

XIRR is money-weighted. It measures the return earned by your rupees, given when you actually put them in. If you happened to invest heavily just before a good run, your XIRR reflects that good luck, or good judgement. It is the honest answer to “how did I do”.

A fund’s published return is time-weighted. It strips out the effect of when money entered and left, because the fund manager does not control that. It is the honest answer to “how did the manager do”.

Both are legitimate and they answer different questions, which is why a fund can report 14% for a period in which your XIRR on that same fund was 9%. Nobody is misleading anybody. The fund is reporting its own performance; you are measuring the consequences of your own timing.

The practical consequence is that you should not use your XIRR to judge a fund, and you should not use a fund’s published CAGR to judge yourself. Comparing one against the other is the most common mistake in this whole area, and it usually ends with an investor blaming a fund for the timing of their own contributions.

Which to use

The fourth row is worth dwelling on. A fund’s published CAGR measures the fund, not you. It assumes a lump sum on the first day. Your XIRR on the same fund will differ, sometimes a lot, because you bought at your own times and not on day one. Neither number is lying; they measure different things.

How to compute XIRR without special tools

Any spreadsheet has an XIRR function. Two columns: dates, and amounts. Contributions are negative, because money left you. The final value is positive on the valuation date.

Two mistakes account for almost every wrong XIRR: getting the signs the same way round, which makes the function fail or return nonsense, and omitting the closing value, which leaves the function with nothing to solve against.

Or use the XIRR calculator, and the CAGR calculator for the lump-sum case.

One honest caveat about both

Neither number tells you whether the return was good. A 12% XIRR over a period when the market returned 16% is underperformance dressed as a decent number, and 8% during a flat market may be excellent.

A return figure needs a benchmark and a time period next to it, or it is decoration. And before comparing anything to anything, take inflation out of it, because a nominal return says nothing about whether you got richer.

Where this fits

• The side-by-side tool, to compute both on your own figures: CAGR vs XIRR
• Your own figures: XIRR calculator · CAGR calculator
• Converting an annual return: CAGR annual return converter
• What inflation does to the answer: real return after inflation