What return has the Nifty actually given over the long run?

What has the NIFTY 50 compounded at over the last two decades?

From March 2004 to August 2026 — 269 months of closes — the NIFTY 50 compounded at 12.4% a year. Measured only over the last ten years to August 2026, the same index compounded at 10.8% a year. These are published index closes, before fund cost, brokerage and tax, so what an investor actually keeps is lower.

Key Takeaways

• NIFTY 50: 12.4% a year over March 2004 to August 2026, across 269 months of closes.
• Over the last ten years to August 2026 the same index did 10.8% a year.
• The window you pick changes the answer — always ask 'from when to when'.
• Index levels carry no expense ratio, no brokerage and no tax; your return does.
• Of 1,518 listed Indian companies with full-year fundamentals, only 22.0% clear a basic quality bar as of August 2026.
• Plan on a long-run number in the low teens before costs, not the last hot year.

How much has the NIFTY 50 actually returned since 2004?

Two numbers answer the question honestly. From March 2004 to August 2026 — a stretch of 269 months of index closes — the NIFTY 50 compounded at 12.4% a year. If you shorten the window and look only at the last ten years to August 2026, the same index compounded at 10.8% a year. Both were computed in September 2026 from index closes held in our own database.

One line matters before you use either figure. This is published index history: these are the index's own closes, not the returns of any service of ours, and index levels carry no fund cost or tax. Nobody earned exactly 12.4%. It is the arithmetic of the index itself, which is the cleanest starting point available and also the most generous one.

Notice that the two numbers disagree by about a percentage and a half, and they describe the same index. That gap is the whole reason people argue about Nifty returns on Twitter. Nobody is lying. They are simply quoting different start dates and not saying so.

NIFTY 50 index history and the shape of the listed universe, as of August 2026

What is measured	Period	Figure
NIFTY 50 return a year	March 2004 to August 2026	12.4%
NIFTY 50 return a year	Last 10 years to August 2026	10.8%
Months of index closes used	March 2004 to August 2026	269
Listed companies with full-year fundamentals	As of August 2026	1,518
Clear ROE >15%, ROCE >15% and D/E <1 together	As of August 2026	22.0% (334)
Median PE of the universe	As of August 2026	24.0

Why is the ten-year number lower than the full-period number?

Because the longer window starts in March 2004, near the beginning of a powerful run, and the shorter one starts later, after prices had already risen a long way. When you begin measuring from a cheap market, your compounded return looks better. When you begin from an expensive one, it looks worse. Nothing about the underlying companies changed between the two calculations. Only the starting price did.

This is why a single headline number is close to useless on its own. '12.4%' with no dates attached is a slogan. '12.4% a year from March 2004 to August 2026' is a fact you can check. Whenever someone quotes you a long-run market return, the first question is not 'how much' — it is 'from when to when'.

The practical takeaway is not to distrust the index. It is to distrust yourself when you pick the window. Most of us unconsciously choose the start date that makes our own decision look smart, and the difference between 12.4% and 10.8% is exactly how much room that choice gives you.

What does 12.4% a year do to a lump sum in rupees?

Here is an illustrative example, using round figures only to show the arithmetic — this is not a projection and not a promise. Suppose ₹10,00,000 grew at 12.4% a year, the NIFTY 50's compounded rate from March 2004 to August 2026. After ten years it becomes roughly ₹32 lakh. After twenty years, roughly ₹1.03 crore. After twenty-two years and change — the length of the actual measurement window — a little over ₹1.3 crore.

Now run the same illustrative ₹10,00,000 at 10.8% a year instead, the index's rate over the last ten years to August 2026. Ten years gives roughly ₹28 lakh. Twenty years gives roughly ₹77 lakh. Over two decades, a gap of 1.6 percentage points a year has quietly removed about a quarter of the final corpus.

That is the real lesson hiding inside a boring rate of return. Small differences in the annual number do very little over one year and enormous things over twenty. It is also why costs matter so much more than most retail investors assume — a fee is subtracted from the compounding rate itself.

Is the index return the same as the return in my hand?

No, and the gap runs in both directions. On the downside, index levels carry no fund cost or tax. An index fund or ETF charges an expense ratio. Buying and selling costs you brokerage, STT, stamp duty, exchange charges and GST. And in India, listed equity held over twelve months is taxed as long-term capital gains above the annual exemption, while anything sold sooner is taxed at the short-term rate. Every one of those is a subtraction from the printed index number.

On the upside, the plain NIFTY 50 level tracks prices only. Dividends paid by the fifty companies are not added back into it. A total-return version of the index, or a fund that reinvests dividends, would sit somewhat higher than the price-only figure. So the number you hold in your hand is the index return, minus costs and tax, plus whatever dividends you actually received and reinvested.

The honest summary for an Indian investor: treat a low-teens long-run index figure as the gross reference, and assume your kept return is meaningfully below it unless you are disciplined about cost and holding period.

What does this mean for someone running a monthly SIP?

A SIP does not earn the index's compounded rate, because your money is not all invested from day one. Each instalment gets a different holding period. The rupees you put in during year one compound for the whole window; the rupees you put in last month have barely started. Your realised SIP return therefore depends heavily on the path the market took, not just its start and end points.

This cuts both ways and mostly in your favour emotionally. A falling market is when a SIP buys the most units, which is precisely when most people stop it. The instalments that felt worst at the time — the ones paid during a sharp drawdown — usually turn out to be the most profitable ones in the final tally.

So do not compare your SIP statement to 12.4% and conclude you are doing something wrong. Compare it to what the same instalments, on the same dates, would have bought in a plain NIFTY 50 index fund. That is the only apples-to-apples test, and your fund house or broker can usually show it to you directly.

Why does my own stock portfolio usually lag the index?

Three reasons, in roughly this order of damage. First, behaviour: selling after falls and buying after rises turns a decent index return into a poor personal one. The index never panics, never takes a call from a friend, and never doubles down to recover a loss. You do. Second, concentration in the wrong places: a portfolio of ten names you accumulated because they were in the news is not a diversified bet on Indian growth.

Third, churn. Every round trip costs brokerage and taxes, and short holding periods attract the higher short-term capital gains rate. A portfolio turned over several times a year can lose a large slice of its gross return to friction alone — and friction is deducted whether the year was good or bad.

If you are sitting on a portfolio that has lagged badly, this is not a character flaw and it is extremely common. The fix is structural, not emotional: fewer decisions, longer holds, a written reason for owning each name, and an honest benchmark to measure against.

How many Indian companies actually clear a basic quality bar?

This is where the index number stops being abstract. As of August 2026, of the 1,518 listed Indian companies in our universe with full-year fundamentals, 458 (30.2%) had return on equity above 15%, and 464 (30.6%) had return on capital employed above 15%. Debt looks healthier: 1,282 companies, or 84.5%, carried debt-to-equity below 1.

But demanding all three at once — ROE above 15%, ROCE above 15% and debt-to-equity below 1 together — leaves just 334 companies, or 22.0%. Add a simple valuation check, a PE between 0 and 40, and you are down to 241 companies, or 15.9%. The median PE across the universe is 24.0.

One caveat you must apply yourself: a flat debt-to-equity rule unfairly penalises banks and NBFCs, because borrowing is literally their business model. A lender with high leverage is not automatically a weak company; it needs different tests altogether. Screen financials separately or you will throw out a whole sector by accident.

Put those two facts side by side and the picture is clear. The index compounded in the low teens across two decades while roughly four out of five listed companies failed a basic quality-and-value bar in August 2026. The index return is an average of a market where most of the constituents are, frankly, not very good.

How should I use this number when I set my own expectations?

Use it as a ceiling on optimism, not a floor under it. If a plan only works when equities deliver 18% a year, the plan is broken — nothing in 269 months of NIFTY 50 history supports that as a base case. Build your goals around a long-run figure in the low teens before costs, and treat anything above that as a pleasant surprise rather than a budgeted item.

Then apply the honesty test in the other direction. Subtract the expense ratio of whatever you actually own. Subtract realistic taxes for your holding pattern. What remains is your planning number, and it is the only one that should appear in a goal spreadsheet for a child's education or a retirement corpus.

Finally, write the number down with its dates attached — 12.4% a year over March 2004 to August 2026, and 10.8% a year over the last ten years to that same date. Keeping both in front of you makes it much harder to be talked into a fantasy return by anyone, including yourself.

Click Here – See BossInvestor's Data-Driven Stock Screens

---

Conclusion

The long-run Nifty answer is a range, not a slogan: 12.4% a year from March 2004 to August 2026, and 10.8% a year over the last ten years to that date, both from published index closes with no cost or tax removed. Anchor your plan to a figure in that neighbourhood, subtract your real costs and taxes, and judge your own portfolio against the same window rather than a flattering one. If you want to see which specific companies clear the quality and valuation bars described here, that work sits behind KYC in the app.

BossInvestor is a SEBI Registered Research Analyst (INH000024143). This article is for educational and informational purposes only. It explains a method and is not investment advice, nor a recommendation to buy, sell or hold any security. Market-wide figures are computed from our universe as of August 2026; past metrics do not guarantee future results. Always do your own research or consult a registered adviser before investing.

---

Frequently Asked Questions

What return has the NIFTY 50 given over the long run?

From March 2004 to August 2026, covering 269 months of index closes, the NIFTY 50 compounded at 12.4% a year. Over the last ten years to August 2026 it compounded at 10.8% a year. Both figures come from published index history — the index's own closes, not the returns of any service. Index levels carry no fund expense ratio, no brokerage and no tax, so an investor's kept return is lower than either figure.

Why do different sources quote different Nifty long-term returns?

Almost always because they use different start dates and do not say so. Our own figures show it plainly: 12.4% a year measured from March 2004 to August 2026, but 10.8% a year measured over just the last ten years to the same date. Starting from a cheap market flatters the number; starting from an expensive one depresses it. Also check whether dividends are included, since a price-only index excludes them. Always ask for the exact window.

Does the Nifty return include dividends and taxes?

The plain NIFTY 50 level tracks prices only, so dividends paid by the fifty companies are not added back. It also includes no costs of any kind. In practice you pay an expense ratio on a fund, brokerage and statutory charges on trades, and Indian capital gains tax — short-term on holdings under twelve months, long-term above the annual exemption. So your real outcome is the index figure, less costs and tax, plus dividends actually received.

Is 12% a year a realistic expectation for Indian equity?

It is a reasonable reference point, not a guarantee. The NIFTY 50 compounded at 12.4% a year across 269 months from March 2004 to August 2026, which is the longest clean window we measure. But it did 10.8% a year over the last ten years, and any individual decade can land well above or below. Use a low-teens gross figure for planning, subtract your costs and taxes, and never build a goal that only works at higher rates.

If the index did well, why did most stocks not?

Because an index return is an average that hides enormous variation underneath. As of August 2026, of 1,518 listed Indian companies with full-year fundamentals, only 334 — 22.0% — cleared ROE above 15%, ROCE above 15% and debt-to-equity below 1 at the same time. Adding a PE-under-40 check leaves 241, or 15.9%. Median PE is 24.0. Note that a flat debt rule unfairly penalises banks and NBFCs, whose business model is borrowing.

Should I just buy an index fund instead of picking stocks?

That is a personal decision, and this page does not make it for you. What the data says is narrower: an index fund gets you close to the index return minus a small expense ratio, while stock picking requires you to consistently land inside the minority of companies that clear a quality bar — 22.0% of the universe with fundamentals in August 2026. Whichever route you take, measure it against the same index over the same window, honestly.