Needed: A 'give it up' on gold

The case for educating citizens about India's obsession with gold

On May 10, at a rally in Hyderabad, the Prime Minister asked Indians to stop buying gold for a year. Three days later the government raised the import duty on gold and silver from 6 per cent to 15 per cent, the steepest single increase on record. Both target the same problem. Neither, on its own, will probably fix it.

 

Again, on September 1, the PM renewed his appeal. The problem is not just arithmetic; it is also cultural, and it is a good deal older than the present worry.

 

Start with the gold, and the hole it sits in

 

Over the 20 years to FY26 India ran a cumulative current account deficit of $689 billion, and spent roughly $749 billion importing gold.

 

In 11 of those 20 years the gold bill alone exceeded the entire deficit. This, of course, is a comparison of magnitudes, not a counterfactual. And these are gross imports — netting jewellery re-exports only eases gold to about 73 per cent of the deficit. Either way, no other discretionary import is anywhere close.

 

The K U B Rao working group set up by the Reserve Bank of India (RBI) found in 2013 that gold accounted for roughly two-thirds of the deficit over the preceding three years, and pressed for converting physical demand into gold-backed financial products. Thirteen years on, very little has come of it.

 

What we sold to pay for it

 

In the 27-odd years to FY2026, foreign institutional investors (FIIs) and foreign direct investors (FDIs), together, put in a net of about $940 billion into Indian equities. Weigh that against an outflow of about $796 billion in gold imports. Anecdotally, we sort of “sold” our equities to foreigners and bought gold in return.

 

From FY1999-2000 to FY08 gold imports ran at about 1.5 times net FII equity inflows; from FY2009 to FY2017, 3.5 times. Over the most recent nine years it is 39 times — $386 billion of gold against $10 billion of FII equity. The last five years, still stark: $257 billion worth gold in, and actually a further $33 billion out on account of FIIs.

 

The retort is that FII money is only half the story and that direct investment, the patient kind attached to factories, has been arriving all along. It has. But look at what ends up leaving alongside it. Through the 2000s, repatriation and disinvestment by direct investors ran under 1 per cent of gross inflow. By FY24 it was an amazing 63 per cent. Gross FDI has never been higher, a record $94.5 billion in FY26, yet net FDI was just $0.96 billion in FY25 and $7.65 billion in FY2026 — $53.6 billion repatriated by foreign investors, and another $33.3 billion invested abroad by Indian companies. Dollars in, dollars out, and almost nothing retained.

 

Not all of what went out, of course, is hot money, but a rise in outflows from 1 per cent to 63 per cent is far too steep and tracks the private-equity exit cycle: In 2025 India’s private equity (PE) and venture capital (VC) industry put in about $36 billion and took out $34 billion. Much of what we counted as foreign equity investment seems to have been a loan against our own future, repayable in dollars, but on somebody else’s schedule.

 

And now the earnings that paid for all of it are at risk

 

For most of this period India could absorb the cost. Services exports compounded, remittances hit records, and the FY26 deficit came in at $25.2 billion, a comfortable 0.6 per cent of GDP. That comfort is going, for two reasons – almost at once.

 

The first is Artificial Intelligence (AI). About a third of India's gross external receipts, i.e. $327 billion of $995 billion, sits in lines directly exposed to GenAI: Computer and information services, business services, tech-linked remittances.

 

A fall of just 8 per cent there would alone have doubled the FY2026 deficit. And remember, FIIs still hold $775 billion of Indian equity (at market value) and FDI’s outstanding claims are an additional $543 billion (at book value — probably three times that if we were to consider current market value), all of which will eventually go out, some of it in equity of the very IT names an AI shock would reprice.

 

The second has already arrived. Hormuz closed in late February; Brent stood at $105 when the Prime Minister spoke in May. Crisil now expects the FY27 deficit at 2.2 per cent of GDP and HDFC Bank at 2.1 per cent, against 0.6 per cent last year. That leaves about $15 billion of headroom before the 2.5 per cent line that former governor Dr C Rangarajan and the RBI treat as the outer edge of what is sustainable. $15 bn of headroom. Last year’s gold bill was $72 bn!

 

A 5 per cent hit to AI-exposed inflows takes the current account deficit (CAD) beyond 2.5 per cent; a 10 per cent hit, to 2.9 per cent. If we can halve the gold bill, then even a 10 per cent shock stays inside the line (Chart 4). We cannot conjure up IT demand, set the price of crude, or stop a PE fund exiting. But we can restrain buying gold.

 

So should gold imports simply be banned?

 

Shudder! But, India has just run the experiment. In July 2024 the duty was cut from 15 per cent to 6 per cent; imports promptly rose 24 per cent to a record $71.98 billion. In May 2026 it went back to 15 per cent. Gold responds to price. However, a ban is only the limiting case of a very high duty. A ban would not reduce demand, only move the payment beyond what the state can see or tax. The forex would still leave, before it arrives – we would simply lose sight of it.

 

It also fails a fairness test. Gold is not a luxury, especially for the older generation. For very many women it is “stridhan”, the one asset they hold – their protection in illness, or a marriage that ends badly or widowhood. A ban takes security from those with least of it, while the wealthy buy in Dubai.

 

To be fair, gold’s investment returns have matched equity's, give or take, depending on the year from when you start comparing. But the issue is less about its performance as an asset class, it is more about its economic impact over decades.

 

How then do we convince our people of the economic ills of an unproductive asset in the locker? Can we put the country before the individual ? It requires a combination of education and emotion.

 

The emotion that the ‘give it up’ campaign got right

 

When the government asked well-off households to surrender their cooking gas subsidy in 2015 through an excellently orchestrated marketing and emotional campaign, some 10 million did within two years. It worked for three reasons. The ask was bounded. It cost a subsidy, not an asset. And there was a visible person at the other end: A woman cooking on gas, through an Ujjwala connection, instead of on firewood.

 

The gold appeal has none of these. It asks for an asset, not a subsidy, offers nothing in its place, and has no beneficiary anyone can picture. It is exhortation, which defers demand rather than redirecting it. Deferred demand returns larger, with a vengeance, particularly when households are anxious.

 

Make the ask finite — not “stop buying gold”, but “one tola this year” into a financial instrument like a SIP, in your own name. Or, even better, “Buy One Tola Less”. Make the substitute real: An equity-linked mutual fund, liquid and inflation-beating. And give it an emotional face: A mother opening a demat account in her daughter’s name. Beti ko sona nahi, SIP do! 

 

But the real work is education

 

None of this requires persuading Indians that gold is foolish. It is not. For most of our history, it was the only honest store of value on offer, and households, especially women, that chose it were reading the state accurately. The failure was never household behaviour, but policy's, in offering no alternative, and never explaining why it mattered.

 

Rationality for a household and rationality for the country are not the same thing, and that gap is what must be taught. A gold coin in a locker holds its value for the family that owns it, but builds no factory, employs no one, pays no dividend. An equity share or an investment in an MF does both — at once. Gold can be a sound trade for the household and a poor deal for the nation. That’s not a contradiction, but the whole problem.

 

Practically every tola is bought with foreign exchange somebody else earned first: An engineer in Bengaluru, a nurse in the Gulf, a garment worker in Tiruppur. And those earnings are what AI now threatens. The gold bangle and the software contract are connected, and the connection runs the wrong way.

 

There is real reason for optimism, and it is generational. India has roughly 136 million investors and 210 million demat accounts, three-quarters of last year’s openings by people under 30. Household equity in shares stands at nearly ~89 trillion. Our daughters are far less enchanted by the glitter than their grandmothers — opening demats, not accumulating gold chains. Our problem is that the generation that has not made that shift holds most of the metal, and nobody has taught them.

 

Education, then, is not the soft accompaniment to the policy. It needs to be the policy – in the vernacular, at bank branches, in schools, wherever households take financial advice. Twelve months of asking will not undo a habit built over generations. Twelve months of explaining, and maybe twelve more, might.

 

Tijori mein band, ya desh ko karein buland?

 

And, by the way, the RBI itself annually buys 40-50-odd tonnes of gold directly. That features nowhere in the import figures above. We could start by selling our own gold to the RBI, saving forex and shoring up the national reserve.

 

Sources: DGCI&S and World Bank WITS (gold); NSDL (FII equity); RBI DBIE (CAD, FDI); Bain–IVCA (PE/VC); Crisil, HDFC Bank; SEBI. Literature: NBER 13619, NBER 13602, IMF WP 13/113, RBI Working Group on Gold (2013).

 

The authors are retired bankers and have earlier worked at Kotak Mahindra Bank. Views expressed are personal.

 

Disclosure: Entities controlled by the Kotak family have a significant holding in Business Standard Pvt Ltd