The rupee-oil link only appears once you remove the dollar

Oil rises and the rupee weakens — except the raw data says the opposite. Twenty years of monthly observations show what is actually driving both.

TradeTurbulence 28 July 2026 4 min · 825 words Updated 14 Oct 2026 AI-assisted · reviewed by Hritik TTSI today 81.5 · Severe

A laden crude tanker lies broadside along the left of the frame at dusk, bow to the right, its hull dark and heavily rust-streaked above the waterline. Deck pipework, railings and a raised crane stand in silhouette against a band of amber light low on the horizon. Four black discharge hoses run from the ship's side into brown-grey water and sweep across the foreground toward the lower right. On the far shore a row of large cylindrical storage tanks runs to the right under sodium lighting, with a flare stack burning orange and trailing smoke. The sky is heavy and overcast. — India's crude import dependence reached a record 88.7% in 2025-26. Almost none of the rupee's recent move traces to it.

India’s crude import dependence reached a record 88.7% in 2025-26, so a rising oil price should mean a weaker rupee. Twenty years of data say the opposite: since 2006, months when Brent rose have been months when the rupee tended to strengthen. The textbook relationship is real, but it is buried underneath a larger one, and separating them changes what the rupee is actually telling you.

The raw correlation runs backwards

Take every month since January 2006 for which Brent, the rupee and the broad dollar all have observations — 5,052 aligned trading days, 246 month-ends — and compare the change in oil against the change in rupees per dollar.

The correlation is −0.108 over one-month windows, −0.249 over three months and −0.402 over twelve. Negative, and increasingly so as the window lengthens. Since the series is quoted as rupees per dollar, a negative correlation means that when oil went up, it took fewer rupees to buy a dollar.

That is the opposite of the received account. Extending the daily series back to 1987 gives the same sign: −0.070 across 9,677 observations.

The dollar is doing most of the work

Oil is priced in dollars, and the dollar is not a neutral yardstick. Over the same monthly sample, changes in Brent and changes in the broad dollar index correlate at −0.311 over one month, −0.499 over three and −0.644 over twelve.

That is a strong relationship, and it points the same way at every horizon: when the dollar weakens, dollar-priced oil rises. A weak dollar also lifts emerging market currencies, the rupee among them. Both halves of the original comparison are being moved by the same third thing, and it is moving them harder than anything passing between them directly.

Strip the dollar out and the textbook link appears

Measure the rupee against the broad dollar index rather than against the dollar alone — what the rupee did on its own, net of what every currency did — and the sign flips.

Against that adjusted series, oil correlates at +0.177 over one month, +0.204 over three and +0.143 over twelve. Positive at every horizon: oil up, rupee weaker, exactly as the import-bill account predicts.

So the mechanism is there. It was never visible in the raw comparison because the dollar effect is roughly two to three times larger and runs the other way.

How much this actually explains

Not much, and the honest version of this finding has to say so. A correlation of +0.204 means oil accounts for about 4% of the variation in the dollar-adjusted rupee over three-month windows. The one-month figure is nearer 3%.

That is a real effect and a small one. It is enough to say that oil is one of the things pushing on the rupee; it is not enough to explain a given month’s move, and nothing here supports reading a rupee level off an oil price.

Two limits are worth stating plainly. Correlation across two series that both respond to global risk appetite does not establish which causes which, and subtracting the broad dollar index is a rough control rather than a clean one — the rupee is itself a component of that index, though a small one.

Where the numbers sit today

Brent last printed $86.99 on 20 July 2026, the 70.87th percentile of its trailing five years. The rupee was at 95.92 per dollar on 27 July, its 98.89th percentile — close to the weakest it has been in five years. The dollar index sits at its 51st percentile, almost exactly typical.

A weak rupee against an ordinary dollar and moderately expensive oil is a combination this analysis does not attribute to oil. The TradeTurbulence Stress Index reads 48.9, squarely in its Normal band, which is its own way of saying the same thing: nothing in the broader trade picture is straining.

What to watch

The broad dollar index leaving its normal band. At the 51st percentile it is currently neutral. A move above the 80th would put the largest driver identified here back in play, in the direction that lifts oil and supports the rupee at the same time.
Brent above $100. That is the 87th percentile of the trailing five-year window, which has run between $59.93 and $138.21. The relationship measured here is estimated on ordinary variation; a supply shock of that size is a different regime, and the small dollar-adjusted coefficient should not be extrapolated into one.
The rupee’s percentile against the dollar index’s. The gap between them is the part of the rupee’s move that is not the dollar. It is the only part this analysis finds oil in.

Every correlation above is computed from three public FRED series — Brent, the rupee and the broad dollar index — aligned by observation date. The import dependence figure is the PPAC’s. All four sources are listed with links in the front matter of this piece, and the daily histories are downloadable as CSV from each indicator page.