The Gulf shut. What did the 2026 oil shock actually cost India?

What actually happened at the end of February?

A shipping lane closed.

The Strait of HormuzStrait of HormuzThe narrow sea passage between Iran and Oman connecting the Persian Gulf to the Arabian Sea. It is the only sea route out of the Gulf, and a large share of the world's oil and LNG passes through it.Its closure in March 2026 is the event this entire article measures. is the only sea route out of the Persian Gulf, and through February 2026 it was carrying about 3.7 million tonnes of cargo a day. The last day it ran normally was 27 February, the trading day before a Middle East conflict began. By April it was carrying about 0.2 million, a fall of roughly 95%. It had not recovered by 24 May, where the World Bank's published series stops. In April and May a year earlier the strait was moving about 3.9 million tonnes a day.

Brent crude averaged about $63 a barrel in December 2025. On 7 April 2026 it printed a daily high of $138. The whole move took about five weeks.

Did this stay an oil story?

No, and that is the first thing worth getting right.

The same strait carries liquefied natural gas out of Qatar and a large share of the world's fertiliser. Indexed to the day before the conflict, Brent roughly doubled at its peak. So did Asian LNG. Urea rose about 85%.

Fertiliser is the channel most readers skip. The World Bank puts the Gulf's pre-conflict share of world urea and DAP exports combined at about 20%, and natural gas is the main feedstock for nitrogen fertiliser everywhere else, so a gas shock raises the price of nitrogen through two doors at once. Hold on to that one. It reappears at the end of this piece in the subsidy line of the Union budget, months after the barrels stopped making news. Fertiliser is also where an energy shock can turn into a food-price shock, which is the same channel an El Nino runs through.

Why does one strait matter so much to this part of the world?

Because more of South Asia's economies lean heavily on Gulf energy than in any other developing region.

The World Bank counts, for each region, the share of its economies that source more than 30% of their oil imports from the Middle East. South Asia tops that count at 50%, and tops the natural gas column too. Latin America is at zero. The same shock reached different parts of the world with completely different force, and India happens to live at the exposed end.

Read that chart carefully, though. It counts economies, not barrels, so it says nothing about how much oil anyone actually buys, and the World Bank does not publish the per-region sample sizes. It is a fact about the neighbourhood, not about India.

How exposed is India, exactly?

India bought about 89% of its crude from abroad in 2024-25: 243 million tonnes imported against 28.71 million produced at home. On that number alone you would expect a closed strait to be an emergency.

The number alone is misleading. India's Gulf dependence is not one figure, it is a gradient, and the two ends of it behaved like different countries.

At the top sits LPG, the cooking gas in the red cylinder: 91.4% of what India imported in 2025 came from the seven Gulf economies. At the other end sits crude oil, at 47.9%. LNG is 60.8%, DAP fertiliser 41.7%, urea only 20.9%. India runs this kind of import dependence in more than one commodity: edible oil is the other big one.

That spread is the whole story. India had somewhere else to buy oil. It had nowhere else to buy cooking gas.

Three cautions on that chart. Gulf share is an upper bound on strait exposure: the World Bank notes that Saudi Arabia can reroute crude through its East-West pipeline to the Red Sea, and the UAE separately has a line to Fujairah on the far side of the strait. Both carry crude, not cooking gas, and neither could replace tanker traffic. Oman is left out of the Gulf group because its ports already sit outside the strait, which the World Bank gives as the reason Oman was less exposed. And 2025 is not a settled structural level: the LPG share has run between 90% and 97% since 2021 and was 97% in 2024, so if anything this understates the dependence.

So why did India's crude keep arriving?

The most likely reason is that India had spent three years quietly rebuilding its supplier list, without ever calling it energy security.

After February 2022 Russian crude began trading at a steep discount, and Europe wound down its purchases over that year. Indian refiners went shopping. By 2025 Russia alone was 32.7% of India's crude import bill and the United States another 5.7%. Iraq at 18.2%, Saudi Arabia at 14.6% and the UAE at 10.8% were still large, but they were no longer the only door.

None of that was done for strategic reasons. It was done because the oil was cheap. When the strait shut in March, the effect was the same as if it had been deliberate: India had contracts and tested routes with a second set of suppliers, none of whom sail past Hormuz.

That explanation is consistent with what happened, but it has not been proved. The partner data here is a snapshot of 2025. Nobody has yet published where India's April-to-July 2026 barrels actually came from, so it remains possible that Gulf cargoes kept moving in quantities the shipping data does not capture. What can be said is that the one product India never diversified behaved completely differently.

And why could it not do the same for cooking gas?

Because LPG does not travel the way crude does.

Propane and butane move chilled or under pressure, in purpose-built carriers. There is no pipeline bypass around the strait, and there is no discounted Russian cargo to switch to. Of the LPG India imported in 2025, the UAE supplied 37.0%, Qatar 21.3% and Kuwait 16.5%, with Saudi Arabia close behind and the United States a distant fifth. Look at the crude chart and the LPG chart side by side and the difference is not subtle. One has an escape route drawn on it. The other does not.

This matters because of who uses the stuff. Crude becomes diesel and petrol and jet fuel, which is to say it becomes freight and travel and industry. LPG becomes lunch.

What did India actually pay for a barrel?

There are two Indian oil prices. In calm months they sit within two or three dollars of each other, a median gap of about $1.6. During this shock they came apart by eighteen. Getting them mixed up is the easiest mistake available here.

The Indian basket is a reference quote: a weighted blend of Brent Dated and the Oman-Dubai average, reflecting the grade mix refiners took that month. The import price is the money that actually left the country when the cargo landed. Cargoes are priced weeks before they arrive, so in a violently moving market the two separate.

What India actually paid went from $65.6 a barrel in February to $95.7 in March, then $118.6 in May, which was the peak. That is a rise of about 81% on February. By July it was back to $87.4.

Note the timing. The reference basket peaked in April; the money peaked in May. Anyone quoting the April basket figure is quoting a price India had not paid yet.

One number on that chart should not be used at all. PPAC's March 2026 basket print of $113.49 sits above both Brent ($103.7) and Dubai Fateh that month, and the basket is supposed to be a blend of the two. No weighting of $103.70 and $91.90 produces $113.49. Our own pull and IndiaDataHub's independent restatement agree exactly, so this is PPAC's number rather than a transcription error, but it cannot be right.

What did the same barrels end up costing?

This is the cleanest measure of the damage, because the volume barely moved.

Between April and July 2026 India imported 81.9 million tonnes of crude. In the same four months of 2025 it imported 81.5 million tonnes. The difference is 0.45%, which is nothing.

The bill went from $40.5 billion to $63.4 billion. That is about $22.8 billion more, a rise of 56%, for the same oil. In rupees, roughly ₹3.48 lakh crore became ₹5.94 lakh crore, an extra ₹2.46 lakh crore in four months. The rupee figure moves faster than the dollar one because the currency was also sliding, which is a separate story.

Two caveats sit on that figure. The 2026 numbers are provisional and PPAC revises them. And this is crude alone, so it leaves out LPG, LNG and refined products, and it ignores that India earns some of it back by exporting refined fuel.

What happened to the cooking gas itself?

It stopped coming.

India imported 3.7 million tonnes of LPG between April and July 2026, against 7.08 million in the same months of 2025. That is a fall of 47.7%. The monthly series dates it precisely: 1700 thousand tonnes in February, 807 thousand in March, 678 thousand in April. April is the lowest of the 28 months PPAC's current table covers, which begins in April 2024, so it is not an all-time low.

And it cost more. The price India paid per tonne went from $534 in February to $852 in May, up about 59%. Less gas, at a higher price, which is the worst combination available.

The honest limit on this: these are imports, not consumption. India produces LPG domestically and holds stocks, and both filled part of the gap. What households actually experienced is not visible in this data, and the retail price is administered anyway. What is visible is that the import channel for the one fuel India could not substitute closed by half.

Why did none of this show up at the pump?

Because someone else absorbed it.

In July 2026 retail petrol inflation was 7.5% and diesel 8.4%. Cooking gas was 5.0%. Headline inflation was 4.4%, food 5.2%. Compare like with like: the crude India landed in July cost 25% more than a year earlier, against 7.5% at the pump. At the May peak the crude gap was 81% over February, and the pump never went near it.

Indian pump prices are formally deregulated, but they do not behave like market prices. They are set by the state-owned oil marketing companies that dominate retail fuel, and through this shock they moved in small steps or not at all. So the question was never whether the shock would be absorbed. It was who would absorb it.

A word on the numbers themselves. MOSPI rebased the consumer price index to 2024 in January 2026, so these figures are not continuous with the older 2012-base series that most published charts still use, and year-on-year comparisons only begin from December 2025. These are also inflation rates, not a pass-through estimate. A real pass-through calculation would need pump prices and tax rates, which is a different piece of work.

So who paid for it?

The exchequer, mostly, and not in the way you would guess.

Union excise dutyunion excise dutyA central government tax charged on goods made in India. After GST absorbed most of it in 2017, roughly nine-tenths of what remains is levied on petrol and diesel, with tobacco making up most of the rest.It is where the cost of holding pump prices down actually landed, as revenue the government chose not to collect., which after GST is levied overwhelmingly on petrol and diesel, fell 22.4% between April and June 2026 against the same months of 2025: about ₹55,605 crore became ₹43,149 crore. That is consistent with the fuel tax cut the World Bank lists among India's responses to the shock, though the fall is not decomposed and timing and volume will account for some of it.

The fertiliser bill rose. Urea subsidy spending went from about ₹31,523 crore to ₹53,034 crore over the same three months, up 68.2%. That is the urea price on the second chart in this piece, arriving in the budget about four months later.

The thing that did not happen is the one most people would predict. India's petroleum subsidy line stayed at almost nothing, ₹282 crore across three months. There was no fuel subsidy surge. India held pump prices down by giving up tax revenue, which does not appear as spending anywhere, rather than by writing cheques.

Two honest qualifications. April posts as a near-zero or negative month for excise in these accounts every single year, which is why this compares April-to-June totals rather than single months. And customs collections rose 36.1% over the same window, so excise and customs together were roughly flat. Customs covers all imports and India charges little basic duty on crude, so that rise should not be read as an oil effect without separate evidence.

Then why did the World Bank raise India's forecast?

Of the 146 developing economies with a 2026 forecast in the June 2026 Global Economic Prospects, 93 were cut and 39 were raised. 14 were left where they were. India is one of the 39.

Most of the other upgrades are commodity exporters, which gain when oil is dear, or economies small enough that a single project moves the number. India is neither. It is not alone though: fourteen commodity importers were upgraded, and the biggest of those upgrades, Jamaica's, was 1.3 percentage points against India's 0.1.

What changed behind that number?

Almost everything, which is what makes the unchanged number interesting.

India's forecast for 2026-27 went from 6.5% in January to 6.6% in June, a move of 0.1 of a percentage point. The January forecast was made before the conflict, when American tariffs were the live risk. By June the US Supreme Court had struck down the tariffs imposed on economic-emergency grounds, though the administration promptly reimposed a temporary 10% surcharge under a different law; the effective US rate had fallen from about 14% to about 12%; India had signed trade agreements with the European Union and the United Kingdom; and GST rates had been cut. An energy shock arrived to take the place of a trade shock, and the two roughly cancelled.

Do not read the upgrade as good news. India still slows from 7.7% to 6.6%, a deceleration of 1.1 percentage points, which on an economy this size is a large amount of missing output. The upgrade is against January's forecast. It is not against last year.

Did India go into this strong?

Yes, and that is part of the answer too.

Year-on-year growth ran 6.8% in the second quarter of 2025, 8.3% in the third and 8.0% in the fourth. The World Bank's estimate for the first quarter of 2026, the quarter in whose final month the conflict began, is 7.8%. India hit this shock accelerating.

Momentum is not immunity, and one month of a quarter tells you very little. But an economy growing at close to 8% absorbs a terms-of-trade hit differently from one growing at 2%, and the timing here was lucky rather than clever.

How to read these numbers

The single most important thing this article cannot tell you is what the shock did to India's current account.

The Reserve Bank publishes balance of payments data about a quarter in arrears. The most recent published quarter, January to March 2026, contains only one month of the conflict, and it recorded a surplus of about $7.1 billion. The first full quarter of the shock was not published when this was written in late August 2026. A $22.8 billion increase in the crude bill over four months has to go somewhere, and the current account is the obvious place, alongside the gold India keeps buying, but that is an inference from monthly trade data and not a measurement. Anyone telling you India's current account deficit widened by a specific amount this year is guessing.

On the rest:

Crude import volumes and values are PPAC monthly tables. They are provisional for recent months and get revised, and PPAC notes that its June and July 2026 figures are prorated from DGCI&S data rather than measured, so two of the four months in the 2026 comparison are estimates.

The realised import price is India's average crude oil import price, published by PPAC and distributed by IndiaDataHub. It reproduces exactly from PPAC's own volume and value tables divided by 7.33 barrels per tonne, which confirms the extraction but is not independent corroboration: it is the same source arriving twice. One month does not reconcile. March 2026 implies $93.81 a barrel from the trade tables against $95.73 published, a 2% gap, and March is quoted above. The Indian basket price is a separate FOB reference series, is not what India paid, and its March 2026 value is inconsistent with its own stated blend, so it is not used.

Partner shares are UN Comtrade for calendar 2025, filtered to drop the duplicate customs and mode-of-transport rows the API returns, with Comtrade's own reported world total as the denominator. They describe the position India was in when the strait closed. Three things to know about them. They are shares of value, not volume. Recomputing on tonnage moves most of them by well under a point, and LNG by about two. They rest entirely on India's own customs declarations, because no Gulf state publishes partner-level LPG exports for 2025, and Russia has not reported to Comtrade since 2022 at all, so the largest single number here cannot be checked against its counterparty. And excluding Oman, whose ports lie outside the strait, matters most for urea: Oman alone supplies 18.7% of India's urea, so a Gulf-plus-Oman figure would be about 40% rather than 20.9%.

Consumer prices are MOSPI's 2024-base CPI, which begins in December 2024 and is not splice-compatible with the 2012-base series ending December 2025. Fiscal figures are the Comptroller General of Accounts monthly accounts, reaching us through IndiaDataHub rather than from CGA directly. April is an accounting artifact for excise there, so only multi-month totals are comparable. These figures have not been checked against CGA's own published monthly account: its downloadable report path no longer resolves, its replacement dashboard renders in JavaScript, and no second distributor carries the union monthly series. The internal checks that can be run do pass, in that the monthly flows annualise to within a percent or two of the Budget Estimates for excise, customs and revenue receipts, which rules out a units or scale error. Confirming the individual months against CGA's own release is a check that remains open.

Growth forecasts are the World Bank's June 2026 Global Economic Prospects. India reports on an April-to-March fiscal year while most economies in the same tables are on calendar years, so the columns are not strictly like for like. The Hormuz shipping series is a seven-day moving average in millions of metric tons a day. That unit is printed on the y-axis of the World Bank's own figure, though not in the PDF's text layer, because the panel is an image.

Every figure above was recomputed from the underlying data files before publication, and the check is re-runnable. That is not the same as every figure being independently corroborated: several rest on a single source, and those are named as such above. Where two sources disagreed, both are given.