Foreign investors sold ₹3.16 lakh crore of listed Indian equities in the first nine months of 2026. The Nifty fell about 13 per cent in rupees and 18 per cent in dollars. The easy conclusion is that foreign selling is the problem. It is not the whole problem.
South Korea has seen roughly $126.7 billion of foreign outflows, yet the KOSPI is up about 72 per cent in dollar terms. Taiwan lost $36.9 billion and is up 64 per cent. Indonesia lost a modest $4.7 billion and is down 30 per cent. The size of the selling does not separate winners from losers. Earnings and the currency do.
Selling can be absorbed
Domestic institutions bought more than ₹5 lakh crore of Indian equities by early August, with mutual funds taking a large share. That is a powerful counterweight, and the Nifty still fell 13 per cent.
Korea shows the other side. About 83 per cent of foreigners’ June selling there came from two stocks, Samsung Electronics and SK Hynix, after enormous gains. Foreign ownership of Korean equities still hit a record in May, because prices rose faster than investors could sell. Trimming concentrated winners is not the same as losing faith in a market.
India’s domestic buying provided liquidity, but it could not offset weaker earnings momentum, valuation compression and a falling rupee.
Earnings set the ceiling
Korea and Taiwan sit at the centre of the AI and memory chip cycle. Memory prices, advanced packaging and AI server demand have lifted profits at Samsung, SK Hynix and TSMC. Foreigners sold into markets where earnings were being revised up.
India’s cycle is positive but ordinary. Revenue growth has cooled from the post pandemic surge, margins are harder to expand, and several large sectors face demand or pricing pressure. When profits grow at a normal pace, a market cannot absorb persistent selling without losing valuation. A rising market does not need foreign money every month. It needs a credible reason to own future cash flows.
The rupee changes the arithmetic
The rupee ended 2025 at 89.87 to the dollar and was near 95.94 by September, a fall of more than 6 per cent in nine months. A 13 per cent fall in the Nifty is a 13 per cent loss for an Indian investor. For a dollar investor it is closer to 18 per cent.
A foreign investor’s return is roughly the local return minus the rupee’s depreciation. The Nifty trades at about 20.5 times trailing earnings, against about 25 times at the September 2024 peak and a 10 year median of 23.1 times. That is a real de-rating. But cheaper is not the same as cheap, especially when the currency has also lost ground.
Oil is the rupee’s swing factor, not its cure
Oil is where the currency pressure is most visible. India’s crude import bill rose 48.4 per cent to $74.8 billion between April and August, even though volumes fell 0.4 per cent. In rupee terms the same bill rose 62 per cent, to more than ₹7 trillion. The gap between 62 and 48 per cent implies the rupee was about 9 per cent weaker than a year earlier. The currency is amplifying the oil shock, and the oil shock is feeding the currency.
Headline Brent also understates India’s cost. The Indian crude basket reached $123.86 a barrel on September 18, while Brent traded near $102 on September 21. The basket is the number to watch.
What would cheaper oil be worth? Estimates of the sensitivity range from about $1.2 billion to about $2 billion a year per $1 a barrel. From August’s import volume of 19 million tonnes, we estimate about $1.7 billion. Applying that to a Brent starting point near $100:
| Brent | Annual import saving | Share of FY27 run rate |
| $90 | about $17 billion | about 9% |
| $80 | about $34 billion | about 19% |
| $70 | about $51 billion | about 28% |
The run rate is the five-month bill annualised, about $180 billion. Brent at $80 would save about $25 billion over nine months, roughly three quarters of the $34 billion that foreign equity sellers took out this year at about ₹93 to the dollar. It would work through four channels. Refiners would buy fewer dollars each day. The trade gap, which hit $31.98 billion in July, would narrow. Lower inflation would give the RBI room to ease without needing high rates to hold the currency. And foreign investors would demand less carry for holding India.
But oil cannot carry the argument alone, for two reasons. First, the price depends on geopolitics. Brent crossed $100 on September 10 on US-Iran tensions and fears for Hormuz, and the rupee fell to 95.33 that day. India cannot influence that. Second, cheap oil did not hold the rupee in 2025. The rupee fell from about 85.6 to 89.87 that year, a drop of about 5 per cent, in a period when the August crude basket averaged only $69.11. History makes the same point. When crude stayed above $100 from 2011 to 2013, the rupee lost nearly 20 per cent over 18 months. Policy then waited for relief that did not arrive.
So India should focus on what it controls. India sources oil from about 41 countries and is seeking more, while petrol demand rose 7.88 per cent and diesel 6.46 per cent in August. Diversified supply, reserve purchases on price dips and slower demand growth matter more than a forecast.
Bonds face the same test
A year ago, India’s 10 year yield was about 2.4 percentage points above the US. With India near 7.2 per cent and the US above 5.3 per cent, the gap is now 1.8 to 1.9 points.
That gap also tells us how much rupee weakness the market will tolerate. A hedged investor pays roughly the yield gap as hedging cost. If the rupee falls about 2 per cent a year, the advantage survives. If it falls faster, an unhedged investor can lose money in dollars while collecting every coupon.
Two false dawns
FPIs put ₹55,524 crore into the primary market and other categories over nine months, including ₹35,409 crore from July to September. The NSDL category is wider than IPOs, and the scale is small. For every ₹1 invested in new paper, nearly ₹6 left listed shares. Foreigners are buying selected issues, not the market.
Foreigners also covered about ₹906 crore of Bank Nifty shorts before the recent policy announcement. Against roughly ₹20 lakh crore of foreign holdings in Indian financials, that is a tactical trade, not a structural return.
How much earnings growth is enough?
This is the question the debate usually skips, so we have put a number on it.
A foreign investor’s dollar return is approximately earnings growth, plus dividend yield, plus or minus the change in the valuation multiple, minus rupee depreciation. We assume a hurdle of about 12 per cent a year in dollars, which is the US 10 year yield of 5.3 per cent plus a risk premium of 6 to 7 points for emerging market equity. We assume a Nifty dividend yield of about 1.3 per cent and no change in the multiple.
| Rupee depreciation a year | Required Nifty earnings growth |
| 0% | about 11% |
| 2% (the yield gap) | about 13% |
| 3% | about 14% |
| 6% | about 17% |
| 8 to 9% (2026 pace) | about 19 to 20% |
Three things follow. First, a durable return needs roughly 13 to 14 per cent earnings growth, which is above recent trend and above nominal GDP growth, so it requires margin recovery and not just volume. Second, the bar is not fixed. Every point of rupee weakness raises it by a point, and oil is what moves the rupee. At this year’s pace the bar is out of reach. Third, India does not need Korea’s upgrades. It needs broad upgrades across banks, consumption, industrials and capex beneficiaries, not a handful of sectors.
A re-rating from 20.5 times towards the 23.1 times median would lower the requirement, but that re-rating depends on foreign buying. It cannot be the starting assumption.
What can steady the rupee
The goal is not a fixed rate. It is depreciation near the 2 per cent the yield gap already prices, with lower volatility. Four groups of measures help.
Reduce dollar demand. Oil comes first. Diversified sourcing, strategic reserve purchases at lower prices and faster substitution lower the bill without needing a price fall. A trade agreement that eases tariffs on Indian exports helps the current account directly, and calibrated curbs apply when gold imports surge.
Add stable dollar supply. The FCNR(B) swap window in 2013 raised about $34 billion, which we cite from memory. Longer tenor foreign currency borrowing and easier approvals for foreign direct investment bring flows that are less sensitive to market mood than portfolio money.
Build credibility. RBI intervention works best when it smooths volatility instead of defending a level, and when its forward book is disclosed clearly. Predictability lowers the premium foreign investors demand.
Cut the cost of holding Indian assets. Wider onshore hedging access, deeper forward markets and tax parity for FPIs on bond and equity gains reduce what investors lose before the currency even moves.
Each carries a cost. Intervention uses reserves, deposit schemes create repayment bulges, tax changes cost revenue, and trade deals involve concessions. None replaces earnings growth. A currency held up only by borrowed dollars is stable only while the borrowing lasts.
The bottom line
Foreign investors are selling, and they will keep selling until India offers two things together: earnings growth of about 13 to 14 per cent and a rupee that depreciates no faster than its carry. Either alone is not enough. Strong earnings on a sliding currency still lose money in dollars, and a steady currency without earnings has no reason to attract capital.
Oil sits between the two. Brent at $80 would remove about $34 billion a year of dollar demand, but India cannot count on it. The rupee is steadied by what policy controls: sourcing, reserves, credibility and the cost of holding Indian assets.
Domestic savings can support the market in the meantime, but they cannot replace what foreign capital is waiting for.
