What India notified

On 30 September 2026 the government notified the third phase of India’s Corporate Average Fuel Efficiency rules for passenger cars, CAFE-III, administered by the Ministry of Power and its Bureau of Energy Efficiency under the Energy Conservation Act. They run for five years from 1 April 2027 to 31 March 2032, in two compliance blocks, FY2028–30 and FY2031–32. The main features, as reported by The Hindu, The Economic Times and Autocar Professional:

• A tightening fleet target. For a fleet at the reference weight of 1,229 kg, the benchmark falls from 3.996 litres of petrol-equivalent per 100 km in FY2028 to 3.3273 in FY2032, or from about 94.8 to 78.9 g CO₂/km — about 16.7% over five years. Heavier fleets get looser targets and lighter ones tighter, on a weight-based line. Fuel consumption is measured on the Modified Indian Driving Cycle; reporting on the global WLTP cycle starts in April 2027.
• Super-credits. Each battery or range-extended EV counts as 3 vehicles in the fleet average; each plug-in hybrid or flex-fuel strong hybrid as 2.5; each strong hybrid as 1.6; each flex-fuel ethanol car as 1.1.
• Carbon Neutrality Factors. Declared CO₂ is cut by 8% for petrol cars, including hybrids, that run on E20 or higher blends; by 22.3% for flex-fuel ethanol cars; by 5% or the notified compressed-biogas blending share, whichever is higher, for CNG cars; and by the notified biodiesel blending share for diesel cars.
• Technology credits. 1 g CO₂/km for each eligible technology — start-stop, tyre-pressure monitoring, regenerative braking, efficient alternators, LED lamps, electric water pumps, six-or-more-speed gearboxes and others — capped at 9 g/km. Self-declared in the first block, test-validated in the second.
• Credits, trading and buy-out. Annual credits and debits can be carried forward or traded within a block and lapse at its end. Shortfalls can be bought out from BEE at ₹2,500 per g CO₂/km in FY2028, rising to ₹4,500 in FY2032. Trading and buy-outs are allowed only in October each year.
• No small-car carve-out. A proposed 3 g/km concession for cars under 909 kg, which split the industry, was dropped.

Almost every one of these mechanisms has a precedent abroad. Some of those precedents were later wound back. The comparison shows where India has copied what worked, and where it has taken a tool that other regulators decided was too generous.

The same machinery, country by country

Sources for each row are in the sections below and in the sources note. India’s figures are as reported by the press from the notification; the reports do not mention a cap on super-credits or say whether the buy-out price applies per vehicle, as the EU’s does.

Europe: the closest model

India’s structure is closest to the EU’s Regulation (EU) 2019/631. Europe sets each manufacturer a target from an EU-wide fleet figure — 93.6 g CO₂/km on the WLTP cycle for 2025–29 — adjusted for the average mass of its cars. It allows carmakers to pool their fleets, which works like India’s trading: Tesla has sold its compliance surplus to other manufacturers in a pool. It gives credits for “eco-innovations”, capped at 7 g/km until 2024, 6 g/km to 2029 and 4 g/km after, much like India’s 9 g/km technology menu. And it charges an “excess emissions premium” of €95 for every g/km over target, multiplied by every car sold. For a large carmaker missing by 5 g/km on a million cars, that is €475 million.

Two European choices are instructive for India. First, the EU used super-credits but kept them small and short: from 2020 to 2022, cars under 50 g/km counted as 2, then 1.67, then 1.33 cars, with the total benefit capped at 7.5 g/km per manufacturer, and the multiplier ended in 2023. The International Council on Clean Transportation found that eight of ten manufacturer pools had used up their entire super-credit allowance in 2020 alone. Europe then replaced multipliers with a benchmark: from 2025 a manufacturer whose zero- and low-emission share beats 25% gets its target relaxed by up to 5%. Second, the EU has repeatedly loosened when targets bit: in May 2025 it let carmakers average 2025–27 instead of meeting 2025 alone, and in December 2025 the Commission proposed lowering the 2035 target from 100% to 90% below 2021, with the remainder offset by credits for EU-made low-carbon steel and for e-fuels and biofuels, and new super-credits for small EVs built in Europe.

India’s compliance blocks do what the EU’s 2025–27 averaging does, but from the start rather than as a rescue. Its buy-out price is lower: ₹4,500 per g/km, the FY2032 top rate, is less than half of €95 at any exchange rate above ₹95 to the euro.

The United States: a credit market being dismantled

The US has two overlapping systems: NHTSA’s fuel-economy CAFE standards, set by vehicle footprint, and EPA’s greenhouse-gas standards. Both let carmakers bank credits for five years and trade them, and that trading made EV makers such as Tesla and Rivian large sellers of compliance credits. EPA also used EV multipliers, most recently 1.5 for battery EVs and 1.3 for plug-in hybrids in model years 2023–24, capped at 10 g/mile in total.

That system has been taken apart in 15 months. The July 2025 budget law, the One Big Beautiful Bill Act, set the CAFE civil penalty to zero. In February 2026 EPA rescinded its 2009 finding that greenhouse gases endanger public health, and with it all vehicle greenhouse-gas standards. On 28 September 2026 NHTSA finalised new CAFE standards for model years 2022–31 that reach a fleet average of 34.9 mpg by 2031, against the 49 mpg previously planned for 2026, and end trading of newly earned credits from model year 2028. With no penalty and no buyers, a CAFE credit is worth little.

The US also offers the clearest precedent for India’s flex-fuel treatment. From model year 1993 the Alternative Motor Fuels Act let carmakers count flex-fuel vehicles as if only 15% of their fuel were petrol, which flattered their fleet averages whether or not owners ever bought E85. NHTSA’s report to Congress on the credit found that very little E85 was actually used, so the credit let fleets burn more petrol, not less. Congress phased it out, ending in 2020.

China: the dual-credit system

China runs two linked ledgers. Corporate Average Fuel Consumption sets a fleet target — 4.0 L/100 km in 2025, about 3.2–3.3 L/100 km by 2030 — and a separate New Energy Vehicle mandate requires credits equal to a rising share of conventional production: 48% in 2026 and 58% in 2027 under rules finalised in November 2025. NEV credits can be sold to other carmakers and used one-for-one to cover a fuel-consumption deficit. Like India, China gave EVs a multiplier in the fuel-consumption average, 2.0 in 2021, and stepped it down to 1.0 by 2025; it also gave very efficient conventional models, under 3.2 L/100 km, a multiplier of 1.4 falling to 1.0. India’s 1.6 for strong hybrids is the closest equivalent anywhere, and it does not step down.

Japan, Korea and Brazil

Japan’s 2030 standard requires a fleet average of 25.4 km per litre, about 3.9 L/100 km, a 32% improvement on FY2016, on a weight-based curve. It brought EVs in for the first time and therefore counts energy well-to-wheel, charging EVs for the upstream energy of the electricity they use. India instead counts electricity at the plug and then multiplies EVs threefold.

South Korea lets carmakers comply on either fuel economy or greenhouse gases, and runs super-credits for electric, hydrogen and hybrid vehicles. A recent amendment tightened its 2030 passenger-car limit from a planned 70 to 54 g/km and extended the super-credits to 2029 to ease the transition.

Brazil’s MOVER programme, in law since June 2024, sets efficiency targets for 2027 in megajoules per km and counts emissions well-to-wheel, with an ethanol utility factor for flex-fuel cars. It is the closest parallel to India’s Carbon Neutrality Factors: both credit ethanol for its lower fossil carbon. Brazil applies the factor to a fuel its flex-fuel fleet actually burns in large volumes; India applies 22.3% to flex-fuel cars whose high-ethanol fuel is barely on sale yet.

India’s targets in real g/km

A gram of CO₂ per kilometre is not the same gram everywhere. Each country measures it on its own laboratory driving cycle, and the cycles differ in speed, acceleration and idling. India uses the Modified Indian Driving Cycle, a version of Europe’s old NEDC, which is gentler than the WLTP cycle Europe, Japan and China now use. The same car emits less on MIDC than on WLTP. The International Council on Clean Transportation converts MIDC results to WLTP by multiplying by 1.226, a factor drawn from Indian vehicle data; its earlier simulation work found petrol cars emit about 13% less CO₂ on the NEDC and 15% less on the US test than on the WLTP cycle. Putting each country’s target on a WLTP footing gives a rough like-for-like comparison:

This blog’s conversions, for petrol cars: MIDC × 1.226 (ICCT); US test ÷ 0.85 (ICCT, 2014); fuel consumption at 23.7 g CO₂ per L/100 km, the factor implied by India’s notified 0.04217; mpg at 8,887 g CO₂ per US gallon. China’s 2025 figure is ICCT’s WLTC value. Each figure is the target for an average fleet before credits, and fleets differ in weight and size, so differences of a few g/km mean little. *Japan counts EVs well-to-wheel, so the figure is not purely tailpipe. †The US figure covers cars and light trucks together.

On paper, then, India’s FY2032 target of about 97 g/km on WLTP is close to the EU’s 2025–29 level and Japan’s 2030 level, tighter than China’s 2025 target and far tighter than the new US rules — but about twice the EU’s 2030 level and well above Korea’s. It is also a target for lighter cars: at the 1,229 kg reference weight, India’s reference fleet is lighter than Europe’s, and a lighter car should emit less on any cycle.

What the flexibilities do to it. The headline target is what the fleet average must show after super-credits, Carbon Neutrality Factors and technology credits. Take the FY2032 target of 78.9 g/km on MIDC for a reference-weight fleet of petrol cars running on E20 and EVs, with the full 9 g/km of technology credits. The table below shows how much the petrol cars could emit on average, and what the whole fleet would then actually emit from its tailpipes, counting EVs at zero, in g CO₂/km:

This blog’s calculation: counted average = (petrol share × petrol CO₂ × 0.92) ÷ (petrol share + 3 × EV share) − 9, set equal to 78.9. It assumes every petrol car qualifies for the 8% E20 factor and the full technology credit, and ignores diesel, CNG and hybrids and the weight adjustment. Real fleets will sit somewhere inside these bounds.

The result is counter-intuitive. With no EVs, the flexibilities loosen the FY2032 target from 78.9 to about 95.5 g/km. With 10% EVs, the petrol cars can average 127 g/km and the fleet as a whole about 115 g/km on MIDC — roughly what CAFE-II required in FY2023 (113 g/km) and below the 121.3 g/km the ICCT measured for India’s new cars in FY2021. Under a threefold multiplier, every EV sold lets the petrol cars beside it emit more, so the fleet’s actual tailpipe emissions can rise even as its compliance figure falls. Europe’s 2030–34 standard has no multiplier, and its benchmark relaxation ends in 2029.

And on the road. Laboratory cycles understate real driving. The European Commission’s first report from fuel-consumption meters fitted to cars registered in 2021–22 found real-world CO₂ about 20% above WLTP for petrol and diesel cars, and 3.5 times the laboratory value for plug-in hybrids, which owners often do not charge. Under the older NEDC, the ICCT found the gap in Europe had grown to around 40% by the mid-2010s. India has no comparable published fleet data for CO₂; its on-road testing so far has focused on pollutants. Congested city driving would be expected to widen the gap, not narrow it.

Where India goes further than anyone

The EV multiplier is the largest in any major market. A threefold count matters most when EV sales are small. Take a carmaker selling 95 petrol cars at 110 g/km and 5 EVs. Counted plainly, its fleet averages 104.5 g/km. At 2×, as Europe did in 2020, it averages 99.5. At India’s 3×, it averages 95.0 — inside the FY2028 target, from a 5% EV share, without improving a single petrol car. Europe capped this effect at 7.5 g/km; the press reports of India’s notification mention no cap.

The ethanol credit applies to almost every petrol car. India has sold E20 petrol nationwide since 2025. An 8% cut in declared CO₂ for cars running on it therefore loosens the target for most of the petrol fleet: a car emitting 103 g/km on test counts as 94.8. The discount credits the ethanol’s carbon as biogenic; it does not mean the car’s tailpipe emits 8% less, which, as this blog has shown, E20 in an engine not built for it barely does. For a flex-fuel car the 22.3% cut and the 1.1 count stack. The US flex-fuel credit was dropped because cars got it whether or not they used the fuel.

The buy-out is cheap and the past is unsettled. Under CAFE-II, BEE identified eight carmakers, including Hyundai, Mahindra and Kia, as having missed their FY2023 targets, with penalties reported at about ₹7,300 crore. The companies argued that penalties notified in December 2022 could not apply to a full year, and this blog found no published resolution. CAFE-III replaces penalties with a priced buy-out and trading, which is cleaner. But a regime that was not enforced last time is now offering a cheaper way out.

None of this makes the targets meaningless. The 17% tightening is real, the technology credits are capped, the second block demands tested evidence, and trading lets carmakers who electrify early earn from those who do not. But how much petrol the rules actually save will depend on how much of the gap between 94.8 and 78.9 g/km is closed by multipliers and fuel discounts rather than by more efficient cars.

What would fix this

• The Bureau of Energy Efficiency should publish each manufacturer’s annual CAFE result: its fleet average with and without super-credits, Carbon Neutrality Factors and technology credits, plus credits traded and bought out and at what price, so the public can see how much of compliance is real efficiency.
• The Ministry of Power should cap the total benefit any manufacturer can take from super-credits, as the EU did at 7.5 g/km, and step the multipliers down between the first and second compliance blocks, as China did.
• The Ministry of Power and the Ministry of Petroleum and Natural Gas should tie the 22.3% flex-fuel factor to the volume of high-ethanol fuel actually sold, through oil-marketing-company sales data, so that the credit does not repeat the US flex-fuel experience.
• BEE should publish the gap between MIDC and WLTP results for each model from the April 2027 WLTP reporting, and start collecting real-world fuel-consumption data from on-board meters as the EU does, so that CAFE-IV targets can be set on the WLTP cycle used by the EU, Japan and China and checked against how cars are actually driven.
• The Ministry of Power should publish the outcome of the CAFE-II FY2023 penalty proceedings against the eight manufacturers, so that the credibility of CAFE-III’s buy-out rests on a known record.

These are proposals from this blog, not recommendations any agency has adopted.

Sources and caveats

About this article: Researched, written and edited by Umashankar Triplicane Dwarakanathan, with AI research assistance; every figure is meant to trace to the primary source cited. See the Editorial Policy for how sourcing, AI use and corrections work.