10-year Nifty 50 SIPs: every start month since 2007

The question

A ten-year SIP is the textbook advice. What has it actually delivered in the Nifty 50, not once but every time someone started one?

We ran ₹10,000 a month into the Nifty 50 for 120 months, starting on the first trading day of every month from October 2007 to October 2016. That is 109 separate ten-year SIPs. Each is valued on the first trading day after its last instalment, and its return is the XIRR, the annual rate that accounts for each instalment's own date.

The index includes reinvested dividends, rebuilt from NSE's daily dividend yield, so it is close to what an index fund holds before its own costs. The first window starts in October 2007 because that is where our daily Nifty 50 price record begins.

The answer

The typical result barely changes with length; the range does. The median sits near 13.3% whether the SIP ran five, ten or fifteen years. What a longer SIP buys is a narrower spread. At five years the outcomes ran from a loss to almost 22% a year. At fifteen they were all between 11% and 15%.

For ten years specifically:

• The middle half of outcomes fell between 12.02% and 14.45%. In money, the middle SIP turned ₹12 lakh into about ₹24 lakh.
• Nine in ten did better than 10.14%. Only 10 of the 109 came in under 10%.
• The best: October 2014 to October 2024, 16.64%, ₹12 lakh to ₹28.73 lakh.
• The worst: April 2010 to April 2020, 3.89%, ₹12 lakh to ₹14.64 lakh. Still no loss.

The end date did the damage

Here are the ten-year windows that returned less than 10%:

Nine of the ten ended between March and November 2020, inside or just after the Covid crash. The SIPs themselves were ordinary; they were measured on a bad day. The April 2010 SIP, left alone for one more year with no new money, was worth ₹26.72 lakh by April 2021, a return of 13.02% a year.

The same thing is happening now. The ten-year SIP that ended on 1 October 2026 returned 10.14%. The one that ended on 1 January 2026, a day before the Nifty 50's high, returned 14.65%, and the one ending in August 2026 returned 12.34%. Nothing about those SIPs differed except when they were added up. Among five-year SIPs, the one ending on 1 October 2026 returned 4.34%, the fourth-weakest of all 169.

That is the practical lesson. If you need the money on a fixed date, the last year or two before it matters as much as the decade before. Moving money out of equity gradually as the date approaches, rather than on the day, is how people protect against it.

Dividends are part of the return

Run on the price index alone, without dividends, the median ten-year SIP falls from 13.34% to 11.90% and the worst from 3.89% to 2.41%. About a point and a half a year comes from dividends. An index fund gets that, less its expense ratio. Anyone comparing an SIP to a price chart of the Nifty is understating what was earned.

What this does not show

• It is one index over nineteen years. That period held 2008, 2020 and 2026's fall, but every future decade is its own.
• Fund costs are not deducted. A Nifty 50 index fund charging 0.2% a year would earn a little less than these figures; an active fund could earn more or less.
• Taxes are not deducted. Gains on units held over a year are taxed at 12.5% above ₹1.25 lakh a year.

Try your own dates

The Nifty 50 returns calculator runs any start and end date, and the SIP calculator projects forward. Our posts on whether the SIP date matters and on stopping an SIP in a crash follow from the same data.