How do I analyse my SIP returns properly, and why is XIRR different from the fund's CAGR?

Why your SIP return is not the fund's return

A fund's 5-year CAGR assumes one lump sum went in five years ago and sat there. A SIP puts money in sixty times. The instalment from last month has had one month to compound; the one from 2021 has had five years. Averaging those into a single rate is exactly what XIRR does, and it is the only honest way to measure a SIP.

The gap is not small. On HDFC Flexi Cap Fund, Direct Growth, to 1 October 2026: the 5-year CAGR is 15.54%, but a ₹10,000 monthly SIP over the same five years earned an XIRR of 13.21%. Both numbers are correct. Quoting the first one to a SIP investor overstates what they actually got by more than two points a year.

Four real SIPs, same money, same window

₹10,000 a month for sixty months — ₹6,00,000 invested — run on each fund's own daily NAV to 1 October 2026:

Same discipline, same amount, same five years — and ₹1.06 lakh between the top and the bottom. The SIP did not cause that spread; the fund did. "Start a SIP" is advice about behaviour, not about which fund, and the two decisions are often confused.

SIP or lump sum?

Arithmetically, a lump sum usually wins, because markets rise more often than they fall and the lump sum gets more time invested. That is the honest finding and it is not the whole story: a SIP wins on the things arithmetic does not measure. It removes the need to pick a date, it matches how salaries arrive, and it makes a falling market feel like cheaper units instead of a reason to stop.

The fair comparison is not "which produced more" but "which one would I have actually stuck with". A lump sum invested in January 2020 was down 40% by March. Most people do not hold through that; almost everyone keeps a SIP running through it.

Is my SIP underperforming — and should I stop it?

Three checks before concluding anything, in this order:

1. Compare like with like. Your SIP's XIRR against the fund's SIP return over the same window — never against its CAGR, which will always look better.
2. Compare against the benchmark, not the leaderboard. A fund can lag last year's winner while comfortably beating its own TRI benchmark. One of those is underperformance; the other is a different style having its turn.
3. Check the window is long enough to mean anything. Look at the fund's rolling returns: if one year in six has historically ended negative, a negative year is not evidence of anything.

Two years of lagging its own benchmark, in a fund whose mandate has not changed, is a real signal. One bad year is usually noise, and acting on noise is how SIP investors turn a market dip into a permanent loss. What we will not do is tell you to stop or continue — that is advice, and MF Analyser is not a SEBI-registered Investment Adviser.

How much SIP to reach ₹1 crore

Straight compound arithmetic at a constant 12% a year — not a forecast, and no fund returns a constant anything:

Time does far more work than the amount. Going from 15 years to 25 cuts the monthly requirement by about three-quarters. And the assumed rate matters just as much: over 15 years, ₹1 crore needs ₹28,899 a month at 8% but ₹16,507 at 14%. Any plan built on a single assumed return should be checked at a lower one.

Reviewing a SIP portfolio once a year

Once a year is enough, and more often is usually harmful. What is worth looking at: whether each fund still beats its own benchmark over the windows it has lived through, whether your funds have drifted into holding the same companies, and whether the allocation you intended is still the allocation you have after a few years of unequal growth. The first is a fund question, the second is overlap, and the third is rebalancing.