Most of this year's growth jump is higher prices, and war prices are about a quarter of it

Median revenue growth for listed Indian companies went from 10.7% to 20.9% in three quarters. We held commodity prices at their February 2026 level, before the Iran–US war, and recomputed growth for every company, using price links tested on 20 years of results and India's own wholesale prices.

How we worked it out

• Learn each company's price link from its past. Over 2006–2025 we measured how much a company's sales moved when the price of its main product or input moved, after removing what moved every company that quarter. Prices are world prices in rupees, or India's wholesale price index (WPI) where it has the product, such as steel, polymers, yarn and paper.
• Keep only links that work on unseen years. Each link had to predict years it was not trained on, in two separate tests. Links that failed count as zero.
• Replay 2026 without the war. Hold world commodity prices at their February 2026 level, and let WPI prices keep their pre-war trend. Apply each company's link and recompute its growth. The gap between reported and replayed growth is the war's share.

Which way is the error? Probably still a little too low. Edible-oil revenue, LNG, coal, commodity traders, price rises passed on by auto, durables and FMCG makers, and the weaker rupee all count as zero. For Sep-26 revenue, read 2.7 points as the floor and about 3.5 as the top. Profit is understated by more, because most margin gains from wider spreads are not removed.

Growth was rising before the war; war prices added about 3 points on top

Sep-26 median growth is 20.9% as reported and 18.2% without war prices. Of the 10.2-point rise since Dec-25, 2.7 points are war prices.

About 1,900 listed non-financial companies per quarter (Sep-26: 1,364 reported by 10 Nov). Without war = world prices held at Feb-26, WPI on its pre-war trend. Source: Cofacto financial database, World Bank, WPI, FRED.

For manufacturers, most of the jump is higher prices; volumes picked up only in Sep-26

Price-adjusted growth stayed near 10% from Dec-25 to Jun-26 and rose to 14.0% in Sep-26, while reported growth went from 10.3% to 22.7%.

About 1,000 manufacturers per quarter (Sep-26: 740 so far). Each company's growth deflated by its industry's WPI group. The price line includes war prices and every other reason prices rose.

Profits at commodity companies pulled away once the war began

Commodity industries' median PAT growth jumped from 4.8% to 35.8% in two quarters; everyone else went from 8.2% to 16.0%.

Companies with positive year-ago PAT. Commodity industries = oil, gas, metals, steel, chemicals, polymers, fertilisers, cement, power, edible oils, sugar, yarn, jewellery, cables and similar. Raw medians, no model.

Polyester, plastics, oil and metals got the biggest revenue lift

Industries with at least 5 reporting companies; company count in brackets.

Dyes, metals, polyester and chemicals gained profit; jewellers and home-care makers paid for it

Reported median PAT growth minus median PAT growth without war prices, after each company's own trailing tax rate. Oil refiners and chlor-alkali left out (regulated pump prices; small-sample outlier).

Oil accounts for 89 of every 100 extra rupees of revenue

₹2.62 lakh crore in total, about ₹2.2 lakh crore after removing listed subsidiaries counted twice.

Largest: Reliance ₹63,563 cr, ONGC ₹40,206 cr, IOC ₹39,847 cr, BPCL ₹28,971 cr, HPCL ₹23,472 cr. HPCL and MRPL sit inside ONGC's consolidated figures, CPCL inside IOC's.

We kept only the price links that predicted unseen years in two separate tests

Industry × quarter cells, FY2007–FY2026 (WPI links from FY2014; GST changeover and Covid lockdown quarters left out). Kept for revenue when both tests give correlation ≥ 0.3 and slope ≥ 0.3.

Method, in detail

• Data. Quarterly results from the Cofacto financial database, 2005 onward. World Bank commodity prices and FRED USD/INR from 2005. WPI from Apr-2012, the 2011-12 base spliced onto the 2022-23 base.
• Common cycle removed. Each company's growth is measured against the median of companies with no price link in the same quarter.
• Company sensitivity. Fitted on 2006–2025, leaving out the GST changeover and Covid lockdown quarters. Where a company's history is short or noisy, its estimate leans toward its industry. A separate model links EBITDA margin to the price level and to the price move within the quarter.
• Profit without the war = PAT − change in EBITDA × (1 − the company's own tax rate over the last four quarters).

Caveats

• Profit is understated for businesses that earn a spread. A margin link passed both tests for only 70 of about 1,050 linked companies. For the rest, wider margins from refining, phenol, polymers or zinc count as normal profit. Margin gains the model does not explain add up to as much as ₹36,900 cr in the latest quarter, against ₹25,300 cr it attributes to war prices; some have other causes, so that is an upper bound.
• Crude has fallen since, and spreads are narrowing. Brent futures were $92 on 11 Nov, down from $113 on 30 Oct. On 10 Nov, China futures showed the PX margin over crude back near its February level and benzene's 16% above it; polypropylene's was still 44% above. Caustic soda is below its pre-war level. The reported quarters above do not change. October's spike keeps Dec-26 crude above Jul–Sep, so the war effect should shrink from the Mar-27 quarter if prices hold.
• Domestic prices move for more than one reason. WPI steel, yarn and polymer prices include duties, crop size and demand as well as the war, which overstates those industries somewhat.
• Second-round price rises are visible but not linked. Autos, durables, food and machinery revenue does not track their own output prices in the tests, so they count as zero.
• Totals count some revenue twice where listed subsidiaries also report on their own. Medians do not change.
• February is a choice of baseline. Copper, aluminium and gold were already up 25–59% before the war; that counts as normal.
• Oil marketers are regulated, and their profit reflects under-recoveries and policy, which the model does not capture.
• Sep-26 is partial: 1,364 of about 1,960 companies had reported by 10 Nov 2026.

Not a SEBI-registered Research Analyst or Investment Adviser. This is not a recommendation to buy, sell or hold any security, and not investment advice. Consult a SEBI-registered RA or IA before any investment decision. Complaints cannot be raised with SEBI for this content. Built with AI from public filings and public price data; it may be incomplete or incorrect.