Macro · 7 min read
An Indian portfolio with zero exposure to US equities is running a concentrated bet — that the Rupee holds firm and that domestic earnings alone can carry returns. Thirty years of data across the Nifty, Nasdaq, S&P 500, gold and USD/INR say both legs of that bet are currently working against the domestic-only investor. A modest, unhedged allocation to US equities and gold — sized as structural diversification rather than a tactical trade — is the data-backed response, not a call to chase the last three years of Nasdaq returns.
Three things are lining up at the same time. First, domestic and US equities have decoupled: the Nifty is ~9% below its November 2025 all-time high of 26,203, while the Nasdaq sits within a few percent of its own May 2026 record — this is a divergence, not synchronised weakness. Second, the Rupee's slide has sharply accelerated: USD/INR moved from 85.46 to 94.91 between March 2025 and September 2026, an annualised pace of 7.2%/yr, more than double the 30-year trend rate of 3.17%/yr. Third, crude oil — India's largest structural import bill — has climbed back to a ~$85/bbl average in 2026 from ~$64 in 2025, re-widening the current account gap just as the India–US rate differential has compressed to its thinnest cushion on record.
| Period | Nifty 50 | Nasdaq (₹-adj.) | S&P 500 (₹-adj.) | Gold (₹-adj.) | USD/INR |
|---|---|---|---|---|---|
| Full sample (1996–2026) | 11.5% | 14.5% | 11.9% | 11.5% | 3.2% |
| Since 2015 | 8.9% | 20.3% | 16.4% | 15.2% | 3.7% |
| Since 2020 (post-COVID) | 10.9% | 22.1% | 18.7% | 21.4% | 4.3% |
| Since 2023 (AI capex boom) | 8.5% | 30.0% | 23.6% | 30.0% | 4.2% |
"₹-adj." converts the US-dollar index or gold price into Rupees at each month's USD/INR rate — the actual, unhedged return an Indian investor would have earned. The gap between Nifty and Nasdaq (₹-adj.) has widened in every successive window since 2015, not narrowed.
The variable, in plain English. USD/INR is simply what one dollar costs in Rupees. When the Rupee depreciates, anything priced in dollars — a Nasdaq holding, a bar of gold, a barrel of oil — becomes worth more in Rupee terms, purely from the currency move, regardless of what that asset does in its own currency. This is why Gold (₹-adj.) and Nasdaq (₹-adj.) CAGRs above are consistently higher than their US-dollar CAGRs: the Rupee leg has been adding, not subtracting, for three decades running.
Historical precedent. The Rupee has not had a single multi-year appreciation cycle in this 30-year sample — every decade closed weaker than it started. But the pace hasn't been constant: the current acceleration echoes the 2011–13 twin-deficit/taper-tantrum episode and the 2018 oil-spike/EM-selloff episode. In both prior cases, the currency didn't revert — it stepped down to a new, weaker plateau and stabilised there for two to three years before the next leg down. That pattern, if it holds, argues against betting on a snap-back to 85 or 90.
Transmission mechanism. Two forces do the work. First, the twin deficit: India imports roughly 85% of its crude, so every dollar of oil-price increase widens the trade gap and adds direct demand for dollars — the 2025→2026 move from ~$64/bbl to ~$85/bbl average is squarely in this camp. Second, and less discussed, the India–US interest-rate differential — the "carry" that has historically compensated foreign investors for holding Rupee assets — has compressed from an average of ~6.2% in 1996 to ~2.7% in 2025, the narrowest cushion in the dataset. A thinner carry means FPI debt and equity inflows are less automatically stabilising, so the Reserve Bank has increasingly let the exchange rate itself absorb the imbalance rather than defending a level.
Cross-asset read. Nifty's correlation with Nasdaq and S&P 500 (in dollar terms) sits at a moderate 0.47 — meaningful co-movement, but far from the 0.85+ correlation the two US indices have with each other. Gold's correlation with both Nifty (0.13) and US equities (0.04) is close to zero — it remains the cleanest diversifier in the set. Most useful for portfolio construction: USD/INR moves inversely with the Nifty (correlation −0.49), meaning the Rupee tends to weaken precisely when Indian equities sell off — which mechanically cushions unhedged foreign-equity and gold holdings at the moment domestic equities hurt most. Debt and Bitcoin sit outside this dataset, but the same currency mechanism applies directionally to any dollar-priced asset; Rupee-denominated debt, by contrast, carries none of this FX cushion.
| Crisis window | Nifty | Nasdaq (₹-adj.) | S&P 500 (₹-adj.) | Gold (₹-adj.) |
|---|---|---|---|---|
| GFC, Oct 2007–Mar 2009 | −48.8% | −31.0% | −33.6% | +50.1% |
| COVID crash, Jan–Apr 2020 | −17.6% | +1.9% | −5.3% | +12.0% |
In both crises, the Rupee's depreciation (+29% during the GFC alone) meaningfully cushioned unhedged foreign holdings relative to what the headline US indices suggest, and gold was outright positive both times.
For a portfolio that is currently 100% domestic (Nifty/momentum strategy plus INR debt and gold), the data supports carving out a structural 15–25% sleeve in unhedged US equity index exposure (Nasdaq/S&P 500-tracking instruments), sized as a standing allocation rather than a tactical overweight — the "since 2023" numbers are a tailwind, not the thesis itself. Critically, this sleeve should be left unhedged: currency hedging removes the exact -0.49 correlation cushion that showed up in both the GFC and COVID tables above. Gold should be held as a separate sleeve from US equities, not a substitute for it — the two behave very differently in crises (gold outright positive, US equities roughly flat to modestly negative), and combining both gives more complete protection than either alone.
Three things would weaken this case. First, if the India–US rate differential re-widens meaningfully — the Fed cutting faster than the RBI — Rupee depreciation should decelerate back toward its 3%/yr trend, closing much of the currency tailwind for foreign assets. Second, if US equity leadership breaks — an AI-capex or earnings disappointment on the scale of the dot-com bust in this dataset (Nasdaq's own worst drawdown was −75% in local currency) — the standalone-return case for US equities weakens even if the currency-hedge argument still holds. Third, if the Nifty–Nasdaq correlation rises well above its historical 0.47 (markets becoming more globally synchronised than this data has shown), the diversification benefit itself shrinks, independent of currency.
This isn't a case for chasing Nasdaq's last three years — it's a case for treating a modest, unhedged US equity and gold allocation as compensation for the Rupee's well-documented structural drift, built on three decades of data rather than the recent headlines alone.