How long should you stay invested in equity? What every Nifty 50 holding period since 2012 shows

The question, asked properly

"How long should I stay invested?" usually means "how long before I can be fairly sure I will not lose money?" The honest way to answer it is not to pick one start date. It is to take every start date in the record and look at what happened after a fixed number of years.

That is what rolling returns do. For each trading day from 21 Feb 2012, I took the Nifty 50 close, found the close one, three, five, seven and ten years later, and worked out the yearly rate in between. That gives 3,267 one year windows and 1,071 ten year windows, each a different investor who started on a different day. The module on performance measurement explains why rolling returns are fairer than a single CAGR, and the blog note rolling returns vs CAGR shows the same idea on our own strategies.

One caution before the numbers. These figures use the Nifty 50 price index, without dividends, from NSE's daily closes. With dividends added, every figure would be about a point a year better.

How often each holding period lost money

The drop is steep. Holding for one year, roughly one start date in six ended lower. Holding for three years, about one in sixty. Holding for five, about one in five hundred. In this record, no investor who held for seven years or more ended with less than they put in, before costs and tax.

The worst, the middle and the best

Two things in this table are worth sitting with.

First, the middle hardly moves. The median investor earned about 12% a year whether they held for three years or ten. Time did not raise the typical outcome; it removed the extremes on both sides. The best one year window, 94.67%, started at the 2020 low; the best ten year window was 14.04% a year.

Second, every worst case in the table ends on the same day, 23 Mar 2020, when the Nifty 50 closed at 7,610.25 in the pandemic crash. The worst one year window started on 22 Mar 2019; the worst five year window on 23 Mar 2015; the worst seven year window on 21 Mar 2013. A crash does not just hurt the people who bought just before it. It hurts everyone who needs their money on that day.

Why the worst case matters more than the average

A table of averages tells you what a typical investor saw. Your plan has to survive the investor who needed the money in March 2020. That is the logic behind the usual rule: money needed within about three years stays out of equity. In this record a three year holding could still be 5.86% a year down, about 16.6% in total, at exactly the wrong moment.

There is a second reason the worst case matters: what you do in it. The crash guide shows how the most expensive mistake is selling after the fall. An investor with a ten year horizon who sold on 23 Mar 2020 turned a temporary 38% fall into a permanent one. The rolling table only holds for people who stay.

What changes with a SIP

Most people do not invest one lump sum; they invest every month. A SIP spreads the start dates across the whole period, so its result is an average of many rolling windows rather than one. That smooths the outcome but does not remove the end date problem: if the market is down 15% in the month you need the money, the whole pot is down with it, including the instalments made years ago.

This is why goal based investors move money out of equity gradually as the goal approaches, rather than all at once on the last day. The goal based investing note and the guide on investing for a child's education work through a glide path. The SIP calculator shows what a monthly amount could grow to at a fixed rate, and the sequence of returns note shows why the order of good and bad years matters more near the end.

Matching money to horizon

A simple way to use the table:

• Under 3 years: the money you need belongs in deposits or short duration debt. A one year equity window lost a third of its value in the record.
• 3 to 5 years: some equity is defensible, but plan for a loss in the year you need it. A mix with debt narrows the range.
• 5 to 7 years: the record gets steady, with the worst five year window close to flat.
• 7 years or more: in this sample, every window ended ahead. That is where most of a long term equity allocation sits.

Limits of the record

Fourteen and a half years is a short history, and the windows overlap heavily: the 1,071 ten year windows all start between Feb 2012 and Oct 2016, so they share most of their path. The sample contains one crash, from which recovery was fast. A crash with a slow recovery would make the worst case for every holding period worse. Before 2012, the Nifty took years to regain its January 2008 high. The table is a description of 2012 to 2026, not a law.

How this shows up in RupeeCase

Every RupeeCase strategy page shows rolling returns of the model next to the Nifty 50 TRI line, so you can see the spread of outcomes rather than one CAGR. Largecap Hybrid runs a momentum rule on large companies with a debt and gold sleeve, 21 holdings rebalanced every 4 weeks: model CAGR 22.74%, deepest fall −22.04%, as of 27 Apr 2026. The back test covers about five years, shorter than the windows in this guide; read it as one window, not a law.

Questions investors ask