I Ran Fourteen Diwalis. The Gold Coin Won.
Gift a SIP, not a gold coin — that was the article I sat down to write. Then I priced every Diwali since 2012 using NSE's own muhurat closes and the RBI's own gold series. The coin won every single window. Here is what that actually proves, and what it doesn't.
The article I sat down to write was called This Diwali, Gift a SIP Instead of a Gold Coin. It practically writes itself. Making charges, GST, a locker, a jeweller who decides what your gift is worth on the day you sell it — versus a compounding machine with a published NAV.
Then I built the model, and the model disagreed.
Every Diwali since 2012, the NSE has run a one-hour muhurat trading session and published a closing level for the Nifty 50. The RBI publishes the average price of standard gold in Mumbai for every financial year. Put those two series side by side, buy ₹11,000 of one or the other on each of the last fourteen Diwalis, and price the result on 14 August 2026:
The coin won by 59%. Not by a rounding error. Not because of a cherry-picked start date — I checked all fourteen possible start dates and the coin won every one of them.
I am an AMFI registered mutual fund distributor. This is the least convenient number I have ever put on this site. So let us go through it properly, because the interesting part is not that gold won. It is when it won.
The machine
Below is the whole model, running live. It buys gold on one side and index units on the other, on every Diwali from your chosen start year, and prices both at whatever finish line you set.
Press play. Watch the two lines. Pay attention to the last two seconds.
Now move the finish line
Drag the finish slider back to Diwali 2024 and read the two counters again.
Gold: ₹2,65,709. Index fund: ₹3,33,146. The fund is ahead by 25%.
Drag forward one notch to Diwali 2025. Gold ₹4,24,655, fund ₹3,64,900 — the coin has taken the lead. Drag to today and the gap opens to 59%.
Here is the full sequence, same start year, same money, only the finish line moving:
| Finish line | Gold coins | Index fund | Ahead |
|---|---|---|---|
| Diwali 2016 | ₹50,817 | ₹66,434 | Fund, +31% |
| Diwali 2018 | ₹73,611 | ₹1,03,764 | Fund, +41% |
| Diwali 2020 | ₹1,37,616 | ₹1,47,685 | Fund, +7% |
| Diwali 2022 | ₹1,69,445 | ₹2,25,963 | Fund, +33% |
| Diwali 2024 | ₹2,65,709 | ₹3,33,146 | Fund, +25% |
| Diwali 2025 | ₹4,24,655 | ₹3,64,900 | Gold, +16% |
| 14 Aug 2026 | ₹5,46,370 | ₹3,43,148 | Gold, +59% |
At every single Diwali from 2016 to 2024, the index fund was winning. Then twenty-two months happened.
Nothing about either asset changed in those twenty-two months. Gold ran from an FY2025-26 average of ₹1,18,421 to ₹1,52,363. The Nifty 50 closed its Diwali 2025 muhurat session at 25,868.60 and sat at 24,366.00 on 14 August 2026 — down 5.8%. One asset had a spectacular year at exactly the moment the other had a bad one, and every one of my fourteen "independent" tests shares that single endpoint.
They were never fourteen tests. They were one test, run fourteen times, all finishing on the same afternoon.
That is the whole lesson, and it is worth more than the number it debunks. A gift is not a bet on where the finish line lands, because you do not get to choose where it lands. If I had published this article at Diwali 2024, I would have told you gold was a 25% mistake. I would have been just as certain and just as wrong.
What the coin costs to be a coin
Gold's return is gold's return. The coin's return is something else, and the difference is the part nobody prints on the box.
Run the fourteen Diwalis again with the frictions switched off — no making charge, no GST, no buyback spread — and the same ₹1,54,000 buys 40.36 grams instead of 36.97, worth ₹6,14,977 today.
₹68,606 — 11.2% of the metal you thought you were buying — never belonged to you. It bought a stamp, a velvet box, a certificate and a jeweller's margin at both ends. It is not a fee you can negotiate away by shopping around; 3% of it is statutory GST and does not come back at any price.
And note what the friction did not do here. It did not change the answer. Gold won by so much over this window that eleven percent of drag was irrelevant. Frictions decide close races. This was not one.
Then why am I still telling you to gift the SIP?
Four reasons, none of them "gold is bad."
One — you already own the gold. Almost every Indian household does, in jewellery, and most people who say they own none are wearing some. Adding a coin to a family that already holds gold is not diversification; it is concentration with a ribbon on it. If you want the exposure deliberately, the last article on this site works through what form to hold it in and how much. The research consensus lands around 7.5–15% of a portfolio — a target, which means there is such a thing as too much.
Two — the coin's return is theoretical for most families. The ₹5,46,370 above assumes somebody actually sells 36.97 grams. In practice the Diwali coin goes into a locker and stays there for a generation. An unrealised gain on an asset nobody will ever sell is a story, not money. The SIP, by contrast, gets redeemed — for a fee, a deposit, a degree.
Three — the tax treatment is not symmetric. Physical gold becomes long-term after 24 months and is then taxed at 12.5% with no annual exemption. Equity funds become long-term after 12 months, are taxed at the same 12.5%, and the first ₹1.25 lakh of gains each financial year is exempt. On a gift that will be sold in pieces over several years, that exemption is worth real money — and it is worth exactly nothing on the coin.
Four — and this is the actual argument — a coin is a full stop and a SIP is a comma. The coin's fourteen-Diwali value came from fourteen separate decisions to walk into a shop. A SIP set up on Diwali 2012 would have made 166 contributions by now without anyone deciding anything. If you are gifting to a child, a nephew, a newly married couple, the compounding you are handing over is only half the gift. The other half is a mechanism that keeps running after your attention moves on. AMFI counted 61.44 lakh new SIPs registered in July 2026 alone, carrying total monthly flows to ₹31,961 crore. Those are not fourteen decisions each. They are one.
When the coin is genuinely the right gift
I would be doing exactly what I accused the model of doing if I stopped here.
Gift the coin when the gift is ceremonial — a first Dhanteras, a wedding, a grandchild's naming. Nobody has ever cried at a folio number, and a financial planner who cannot tell the difference between a portfolio and a family is not much use as either.
Gift the coin when the recipient has no bank KYC, no PAN, or no interest in ever looking at a statement. A gift that requires paperwork from someone who will not do it is not a gift.
And buy the gold — as an allocation rather than a gift — when your portfolio genuinely has none. That is an asset-allocation decision, and it should be made in a spreadsheet in June, not in a queue on Dhanteras.
How to actually gift a SIP in India
If you do go this way, the mechanics matter and most people get one of them wrong.
To an adult. You cannot open a folio in someone else's name and fund it from your account — third-party payments are not accepted. The clean route is to gift the money, let them complete their own KYC, and let them start the SIP from their own bank account. Your gift is the transfer; the SIP is theirs.
To a child. The folio is opened in the minor's name, with a parent or legal guardian representing them. No joint holder is permitted. The guardian must be KYC-compliant and the relationship documented, and the minor's date of birth must be evidenced. Redemption proceeds go only to the minor's verified bank account.
The eighteenth birthday is a hard stop. AMCs are required to suspend standing instructions when a minor attains majority. The SIP terminates automatically and no transaction is permitted until the now-adult completes fresh KYC and submits updated bank details. Diarise it — folios sit frozen for years because nobody knew.
Tax while they are a minor. Under Section 64(1A), a minor's income — including capital gains on a folio funded by a parent's gift — is clubbed with the income of the parent who earns more, with an exemption of just ₹1,500 per child per year. The gift is not a tax shelter. It is a habit.
Questions people actually ask
Did gold really beat the Nifty over the last fourteen years?
Over these particular fourteen Diwali-to-Diwali windows, all measured on 14 August 2026, yes — and comfortably, even after coin frictions. Priced point to point, gold compounded at about 12.5% a year from Diwali 2012 and the Nifty 50 price index at about 11.2%. Add back the dividends this model deliberately excludes and equity gets to roughly ₹3.86 lakh against gold's ₹5.46 lakh. Gold still wins. The result is real. It is also entirely a function of an endpoint at which gold is near a record and the Nifty is below its level of ten months ago.
Isn't excluding dividends unfair to the equity side?
Yes, and deliberately so. If the conclusion survives a handicap it does not need, it is a sturdier conclusion. The Nifty 50 dividend yield published by NSE on these dates ran between 1.10% and 1.46%, so the equity figures here are understated by roughly that much a year. It does not change the ranking over this window.
Isn't the gold price you used the wrong one?
It is the RBI's financial-year average, which is the official series that goes back far enough to do this. In a rising year that average sits below the actual Diwali price — for Diwali 2025 it was ₹1,18,421 against a spot around ₹1.27 lakh — which means the model buys gold slightly too cheaply and hands gold slightly too many grams. The bias runs in gold's favour, and gold won anyway.
Should I just buy gold this Diwali, then?
That is precisely the reasoning the finish-line chart exists to break. Every asset looks obvious immediately after it has run. The question worth asking is not which one won the last fourteen years; it is what share of your portfolio you want in an asset that pays no income, costs you money to hold, and has spent whole decades going nowhere — including 2013 to 2019, visible in the chart above as a flat stretch six years long.
What about a gold ETF instead of a coin?
Then you are no longer having this argument. An ETF strips out the making charge and the buyback spread, reaches long-term tax status in 12 months rather than 24, and is a perfectly reasonable way to hold the allocation. It also makes a poor gift, which is the whole subject of this article.
Sources
- NSE — Daily index closing values (
ind_close_all), Nifty 50, for every Diwali muhurat trading session from 13 November 2012 to 21 October 2025, and for 14 August 2026. Dividend yields on the same dates ranged 1.10%–1.46%. - Reserve Bank of India — Handbook of Statistics on the Indian Economy: average price of standard gold in Mumbai, financial-year averages FY2012-13 through FY2025-26.
- IBJA — Gold ₹1,52,363 per 10g (24K), 14 August 2026, used as the terminal price.
- NSE Indices — Nifty 50 Total Return Index calendar-year returns, used to quantify the dividend adjustment discussed above.
- AMFI — Monthly data, July 2026: 61.44 lakh new SIPs registered; total monthly SIP contribution ₹31,961 crore.
- SEBI — Circular on investment in units of mutual funds in the name of a minor through a guardian (May 2023), and AMFI operational guidelines on minor folios, guardian KYC and suspension of standing instructions on attaining majority.
- Income Tax Act, 1961, as amended — 12.5% long-term capital gains; 24-month holding period for physical gold with no annual exemption, 12 months for equity-oriented funds with ₹1.25 lakh exempt; Section 64(1A) clubbing of a minor's income with a ₹1,500 per child annual exemption.
- GST on gold: 3% on metal value and 5% on separately itemised making charges, per current CBIC schedules.
Notes: Past performance over any window, including the one modelled here, does not predict future returns — the article's central point is that the window itself decides the answer. Every figure carries its date; prices, duties and tax rules change. Calculator outputs are illustrative, not forecasts.
Have questions after reading this?
I'm Punit Sharma — financial planner & derivative analyst. Happy to review your portfolio or answer any questions.