Issue Twelve  ·  1 October 2026  ·  Fortnightly

Dhansanchay

The Compounding Life

A fortnightly letter from DHANSANCHAY

◆

Inside: the floor that is holding, and the fashion that is not  ·  the worst September since 2018, in numbers  ·  a duty rumour, worked out in arithmetic  ·  the turkey that was fed for a thousand days  ·  the gold that may already be in your equity fund

In This Issue

I  ·  The Founder's Notebook
II  ·  From the Editor's Desk
III  ·  Market Pulse

IV  ·  In Focus
V  ·  Mindset Corner
VI  ·  Regulatory Radar

I

The Founder's Notebook

Bhanu Pratap Jain  ·  CEO & Founder

The Floor Is Holding. Now Comes the Hard Part: Doing Nothing Clever.

S eptember was the Nifty's worst month of September since 2018. It fell about 6% to 22,620, and stands roughly 14% below the record it made in January. Every headline this week is written in the language of fear, so let me say something that may sound out of season: I am looking at that number with cautious optimism, not dread. Two years ago, at the end of September 2024, this index stood near 26,200. Thirty months ago, in the early part of 2024, it stood about where it stands today. In between it has gone up, come back, gone up again and come back again — and each time it has come back to this same broad neighbourhood, around 22,000 to 23,000, it has found buyers. Three visits to one level in thirty months is not a promise. But it is an observation, and I would rather act on observations than on headlines.

Here is what makes me hopeful. The price has gone nowhere for two years, but the businesses underneath it have not stood still. Earnings have kept edging up while the index drifted lower, which means the market is now asking you to pay less for each rupee of profit than it did two years ago. That is the arithmetic of a base being built, not of a structure coming apart. It says nothing about next month, and I will not pretend otherwise.

The second reason is one I have been turning over for weeks: India has quietly become the world's anti-AI trade. Global money has poured into the companies that build artificial intelligence — chips, data centres, the platforms themselves — and India's index has almost none of them. Worse, the one large group India does have in the AI conversation, the IT services companies, sits on the wrong side of it: the market fears that AI will do the work those firms are paid to do. That fear is visible in September's numbers, where the three biggest IT names were down between 11% and 14% for the month as of the 29th. But look at the rest of the index — banks, consumption, energy, autos, healthcare. These are businesses that will use AI to run better, not businesses that need AI to keep spending on it going up forever. If the AI build-out ever cools, or is slowed by regulation or simple fatigue, money that fled to the fashion will look for earnings that do not depend on it. India has those earnings. That is the thesis, and it is why we are optimistic.

A floor is an observation, not a promise. Optimism is a reason to keep investing on schedule — not a reason to bet the household on a thesis.

Now the honest half, because a thesis I cannot argue against is not a thesis, it is a mood. India is not purely anti-AI: Goldman Sachs has counted 42 Indian companies that supply the AI build-out — power equipment, data-centre hardware, chip assembly — and reports them up roughly 60% this year while the Nifty fell. So the cleaner idea is that India's index is light on AI, not that India is absent from it. Foreign investors have also pulled roughly ₹2.4 lakh crore out of Indian equities so far this year, the rupee has touched a record low near 96 to the dollar, and crude oil is above $100. Any of those can push a floor through. Support holds until the day it doesn't.

So what does optimism look like in practice? It looks boring, which is how I like it. The SIPs already running continue, untouched. No one should stake a lump sum on my view of the AI cycle, or anyone's. Where a family has surplus that is meant for equity and does not need it for seven years or more, the conversation we would have is about spreading it over several months, so that being wrong on timing is survivable. I hold this view firmly enough to keep investing and loosely enough to change it.

The other half of my caution is gold and silver, and here I want to be careful, because I am not bearish on gold. Gold has done what a diversifier is supposed to do, and what you already hold stays. What we are not comfortable doing is adding new money to gold and silver funds right now, and the reason is not the price. It is a rumour. In May the government raised the import duty on both metals from 6% to 15%. Since late August there have been reports that it is weighing a return to 6%. Nothing has been announced, and I am not predicting what the government will do. But if it happens, the domestic price of gold — and therefore the value of every gold fund — would reset lower in a single step, whatever the metal did abroad. I do not want to be the one who told you to buy just before a policy notification. Section IV works the arithmetic out so you can see it for yourself. If the duty stays and gold rises, we will have missed some of it, and I am comfortable owning that cost. We never held gold to catch that move.

Put the two halves together and you have the whole approach: hopeful about the base, humble about the timing, and unwilling to make new decisions on the strength of a rumour or a chart line. We say it often because it keeps proving true. Be smart, do nothing — and when you must do something, do the small, staggered, reversible thing. Sapna has the September numbers in Section II; the behavioural thread, with a turkey in it, is in Section V.

◆    ◆    ◆

II

From the Editor's Desk

Sapna Jain  ·  Editor

A Bad Month Has a Very Ordinary Shape

Last issue I ended with a question about small and mid cap: whether record inflows would survive a harder market. September has now supplied the harder market, and the answer will only be visible in the industry's September data, which arrives in the second week of October. I will report it in the next issue rather than guess at it here.

What I can do is describe the month honestly, because a bad month looks scarier in a headline than it does in a list. The Nifty fell about 6.1%, its weakest September in eight years, and 40 of its 50 stocks were lower for the month as of the 29th. Foreign investors had withdrawn about ₹21,000 crore by the third week of the month, taking their 2026 outflow to roughly ₹2.4 lakh crore — already more than the ₹1.66 lakh crore they withdrew in all of 2025. The rupee slipped through 96 to the dollar for the first time, and crude oil stayed above $100.

Read that list again and notice what it is made of: oil, the dollar, US interest rates, the rupee. Not one of those is a statement about how an Indian company is doing. I am not saying the fall is unimportant — it changed the value of what you own, and pretending otherwise would be unkind. I am saying that a month with this ingredient list is the kind that has come and gone many times, and the households that came through it best were the ones whose instalments kept going.

One piece of plain arithmetic, offered as arithmetic and not as a forecast: a SIP instalment made in a fallen market buys more units for the same rupees than the same instalment made in a rising one. That does not make the next month safe. It only means the discipline you are practising in a month like this is the discipline doing exactly what it was designed to do.

If a month like this has made you wonder about your own SIP amounts, or whether your emergency fund is where it should be, that is exactly the conversation to have with us before the next statement, not after. Write to me any time.

Sapna Jain

Editor, The Compounding Life  ·  [email protected]

III

Market Pulse

The Worst September in Eight Years. Look at What It Was Made Of.

Seven takeaways, compiled and contextualised by Sapna Jain

Market data as at the close of Wednesday 30 September 2026, except where a date is given.

1.

A bad month, fully on the record. The Nifty 50 closed September at 22,620.45, down 6.06% for the month (August closed at 24,080.40) — its weakest September since 2018 — and about 14% below its 5 January record of 26,373.20. Forty of the 50 index stocks were lower for the month as of 29 September.

2.

Since our last issue. On 11 September, the last close we reported, the Nifty stood at 23,398.10. It has since lost a further 3.3%, the drift lower coinciding with oil above $100, firmer US yields and continued foreign selling.

3.

AI fear landed on India's IT names. Information technology took the sharpest damage: as of 29 September, TCS was down 13.7% for the month, Wipro 12.4% and Infosys 11.5%. A brief relief rally on 15 September, when the Nifty IT index rose 4% as fears of AI disruption eased, did not survive the month. The stocks are being priced for a question nobody can yet answer: how much of their work AI takes over.

4.

Foreign money: still leaving. Foreign portfolio investors had withdrawn about ₹21,000 crore from Indian equities by the third week of September, after a two-month spell of buying (final September data was not out at press time). Their 2026 total is roughly ₹2.4 lakh crore of net selling — more than the ₹1.66 lakh crore withdrawn in the whole of 2025.

5.

The rupee crossed 96. It touched a record low near ₹96 to the US dollar this week, its sharpest weekly fall in four months, with Brent crude around $107. A weaker rupee raises the cost of imported oil, and imported oil feeds straight back into the currency: the two are locked together.

6.

Gold and silver eased; the duty question is the one to watch. Gold traded near ₹1.49 lakh per 10 grams in India, while international bullion fell about 4% to a seven-week low on expectations of firmer US interest rates. Silver also slipped over the month. Separately, the possible cut in import duty on both metals — reported, not announced — is the subject of Section IV.

7.

What did not fall. Index-level earnings have kept edging up even as the index fell — the Nifty is down roughly 13% this year while trailing earnings per share have risen — so the market is paying a lower multiple than it was a year ago. That is context, not a signal; markets can stay cheaper for longer than anyone expects.

How Far Below the Record Has the Nifty Slipped?

Nifty 50 versus its 5 January 2026 closing record of 26,373.20 · per cent below · Source: NSE closing data

8 Sep
  −10.4%
11 Sep
  −11.3%
30 Sep
  −14.2%

Three closes in three weeks, each lower than the last. A slope like this is why the honest reading is "holding so far" — not "safe".

Sources: NSE/BSE closing data; NSDL foreign-portfolio-flow data as reported by the financial press; Moneycontrol report on the gold-duty review (26 August); Goldman Sachs research on India AI enablers (September), as reported. Figures as available on 1 October 2026 and subject to revision.

Editor's Read

Oil, the dollar, US rates and the rupee drove the month. None of them is a verdict on an Indian business. The one number that is — earnings — kept quietly rising.

◆    ◆    ◆

IV

In Focus

A Duty Rumour, Worked Out in Arithmetic

Educate

On 13 May 2026 the government raised the import duty on gold and silver from 6% to 15%, to protect foreign-exchange reserves and narrow the current-account deficit after gold imports jumped sharply in April and the rupee weakened. It has not worked as intended: imports rose about 34% in May and were still 5.5% higher in June–July than a year earlier, and industry sources attribute part of that to metal moving through unofficial channels. On 26 August, Moneycontrol reported that the government is weighing a return to 6%, and that silver and platinum are part of the same discussion. The Finance Ministry has not commented. As of 30 September 2026 there is no decision — and until an official notification appears, a rumour is all it is. We are working out what a cut would mean, not forecasting one.

Interpret

Here is the mechanism, in simplified form. The domestic price of gold is roughly the international price, multiplied by the rupee–dollar rate, multiplied by one plus the import duty. Gold funds and gold ETFs are valued on domestic spot prices — SEBI moved them to exchange-published prices from 1 April. So if the duty falls from 15% to 6%, the last term drops by a factor of 1.06 ÷ 1.15, or about 7.8%, and the value of every gold and silver holding in India resets by roughly that much on the day the change takes effect. Two other forces then push against it or with it: the rupee (weaker raises the domestic price) and the international gold price. The table shows how the three combine. It is arithmetic, not a forecast.

Domestic gold price if duty falls from 15% to 6% — change from today

Illustrative arithmetic, ignoring GST and local premiums. Not a prediction.

Intl gold −5%Intl gold flatIntl gold +5%
Rupee unchanged−12.4%−7.8%−3.2%
Rupee 2% weaker−10.7%−6.0%−1.3%
Rupee 4% weaker−8.9%−4.1%+0.7%

Reading the table: with nothing else moving, the duty cut alone takes about 7.8% off the domestic price. Only a much weaker rupee, or a much higher international price, can offset it.

The rupee at a record low is exactly why this is not a one-way street: a currency that keeps sliding props the domestic price up. Firmer US interest-rate expectations, which pushed international bullion to a seven-week low this month, pull the other way. Three moving parts, one of them a decision that only the government can make.

Explain

None of this is an argument for selling gold or silver. A holding bought as a diversifier is there to do a job across years, and a one-time repricing on a policy notification does not change the job. The narrower question is about new money. Money added before a notification carries the risk of that step-down on day one; money added after carries none of it. If the duty stays at 15% and the metal keeps rising, waiting costs the gain that was missed. That is the trade, and it is why our present posture in conversations with families is to hold what they own and not add to gold or silver until the position is clearer one way or the other. It is equally not a reason to sell equity to buy gold, or the reverse, on the strength of a rumour.

Framework

Three questions to ask about gold and silver in your own portfolio. What is it there for, and has that job changed? How much do you hold in total — including the gold and silver that may now sit inside your equity and hybrid funds (Section VI)? And could you sit through an eight-per-cent single-day repricing of this holding without acting on it? If the honest answer to the last one is no, the holding is larger than your nerves can carry, whatever the duty does.

◆    ◆    ◆

V

Mindset Corner

The Turkey Was Right Every Day for a Thousand Days

There is a story we keep in the Dhansanchay drawer for exactly this kind of fortnight. A turkey is fed every morning by a kind farmer. On day one it is cautious. By day one hundred it has a hundred observations of kindness and no counter-example. By day nine hundred and ninety-nine its confidence is at its peak, and its evidence is overwhelming — every single data point agrees. Then comes the day before Thanksgiving. The turkey's model was not wrong on any day it was used. It was wrong on the one day that mattered, and it had no way of knowing that from inside its own record.

Read this alongside Section I. A floor that has held three times in thirty months is the market's feeding record. It is genuine information, and it is a fair reason for optimism. It is not a guarantee that day one thousand and one will look like day nine hundred and ninety-nine. The opposite error is just as old: the forecaster who has called nine of the last two crashes and treats the next fall as certain. Both are the turkey, reading a record as a law.

The Compounding Test

The answer to not knowing is not a better forecast. It is a structure in which being wrong does not ruin you.

That is what we mean by the Incerto Check, and it is worth running on any thesis — including ours. Can you survive being wrong? A view held in a SIP and an emergency fund can be wrong for years without hurting you; the same view held in one large lump sum cannot. What can be removed rather than added? Often the best action this month is to leave something out: the gold top-up, the sector fund, the second-guessing. And has it lasted? The habit of investing steadily through falls has a long record. A prediction about this particular floor does not, yet. A good thesis, held with bad sizing, is still a bad decision. That is the whole lesson of the turkey — and it is why we prefer to be boring.

◆    ◆    ◆

VI

Regulatory Radar

There May Be Gold in Your Equity Fund Already

What changed. Under the SEBI (Mutual Funds) Regulations, 2026 and the February categorisation changes, equity and hybrid schemes may now place part of their non-core allocation in gold and silver exchange-traded funds, infrastructure investment trusts and debt. Most fund houses put the change into effect by addendum in late August. The practical result: an "equity" fund you hold may now carry some gold or silver exposure that was not there a year ago. The amount differs from scheme to scheme.

Why it matters this fortnight. If you are watching the gold-duty story in Section IV, your real exposure to the metal is your dedicated gold or silver holding plus whatever your other funds have quietly added. Read the latest scheme addendum or monthly factsheet for the actual figure; it is disclosed. This is the overlap idea we introduced last issue, applied to a new asset.

Valuation of gold and silver funds. Since 1 April 2026, physical gold and silver held by mutual fund schemes is valued on exchange-published polled spot prices rather than international benchmarks. That is why a change in Indian import duty would show up in unit values directly. It is a rule of the framework, not a comment on any scheme.

A reminder on rumours. Only an official government notification changes a duty rate. Reports, however well sourced, are not decisions. We will tell you the moment something is confirmed, and not before.

On the watch list. Industry data for September, expected in the second week of October, will show whether last month's record small and mid cap inflows survived a falling market. Nothing new requires action from you this fortnight.

Your Action  ·  5 Minutes

Want to know how much gold or silver you actually hold across all your funds? Write to [email protected] with the subject line "Overlap Check" and we will map it for every fund you hold with us.

A Note on This Section

This is general market and regulatory commentary for information and education — not personalised investment advice, a recommendation, or a call to buy or sell any security or scheme. Views on the market and on gold and silver are those of the author, are subject to change, and are not a forecast or assurance of any outcome. Please consider your own goals and risk profile, and speak with us directly before acting on anything here. Write to [email protected].

◆    ◆    ◆

This Issue

Bhanu Pratap Jain

CEO & Founder  ·  The Founder's Notebook

[email protected]

Sapna Jain

Editor  ·  Market Pulse, Research & Dispatch

[email protected]