The full CAFE-III rulebook for India PV OEMs — the 105 to 78.9 g target curve, the WLTP transition, super-credits at 3.0× BEV, the 17 off-cycle technologies, block-period averaging, and the new OEM-to-OEM credit market.
The Draft CAFE 2027 standard (notified 25 September 2025) sets a new compliance regime for M1 passenger cars from 1 April 2027 through 31 March 2032. The standard is expressed in petrol-equivalent litres per 100 km — not directly in g CO2/km — and tightens through year-by-year c constants while the slope (a) and reference mass (b) stay fixed across the cycle. Super-credits (volume derogation factors) for electrified vehicles, an off-cycle CO2-reducing-technology derogation, and a new Carbon Neutrality Factor for biofuel-blended fuels all enter the math. Pooling across up to three manufacturers is explicitly allowed. The test cycle remains MIDC until MoRTH separately notifies the WLTP transition, with a conversion factor for CAFE 2027 targets to be issued by the Ministry of Power.
The headline arithmetic
Standard fleet average = a × (W − b) + c, where a = 0.002 L/100 km per kg, b = 1,170 kg, and c steps from 3.7264 L/100 km in FY28 to 3.0139 L/100 km in FY32. At b = 1,170 kg that translates (via BEE's published 23.7135 g/km-per-L conversion) to ~88.4 g CO2/km in FY28 falling to ~71.5 g CO2/km in FY32 on the MIDC cycle.
The Draft CAFE 2027 standard publishes year-by-year c values (in L/100 km) while keeping a = 0.002 and b = 1,170 kg constant. At b = 1,170 kg, the standard reduces directly to c; for any OEM with a different weighted fleet weight W, the target moves linearly via a × (W − b). The g CO2/km figures below are derived using BEE's own petrol-equivalent conversion (1 L petrol / 100 km ≡ 23.7135 g CO2/km, from the FCPetrol = 0.04217 × CO2 formula in the draft).
| Fiscal year | c (L/100 km) | Implied g CO2/km at W=1,170 | Notes |
|---|---|---|---|
| FY27 | ~4.78 | ~113 | CAFE-II close (S.O. 5020(E), 6 Dec 2021) |
| FY28 | 3.7264 | ~88.4 | CAFE 2027 starts — 22% step down vs FY27 |
| FY29 | 3.5737 | ~84.7 | |
| FY30 | 3.4573 | ~82.0 | |
| FY31 | 3.2224 | ~76.4 | |
| FY32 | 3.0139 | ~71.5 | Final notified target |
Where the widely-cited 78.9 g/km figure comes from
The "78.9 g/km FY32" figure that has been widely reported in trade press is not in the official 25 Sept 2025 draft. It traces back to BEE's June 2024 CAFE-III/IV proposal which cited 91.7 g CO2/km on WLTP for CAFE-III at weighted average mass 1,170 kg. The September 2025 draft replaced that with year-by-year L/100 km c values on the MIDC cycle. Until the MIDC→WLTP conversion factor is separately notified, the gross MIDC values above are what apply. The 78.9 figure likely persists in industry discussion because it was the politically-agreed FY32 endpoint at an earlier stage of consultation.
Trying to model what 30% tightening means for your portfolio?
Open TerraNova →MIDC, the cycle CAFE-II runs on, is India's 1990s adaptation of the European NEDC. WLTP — the Worldwide Harmonized Light Vehicle Test Procedure — is the new globally-harmonised cycle, longer, faster, more dynamic, and closer to real-world driving. India's WLTP standard is published as AIS:175 (Automotive Industry Standard 175, March 2026), based on UN Global Technical Regulation 15, Amendment 6 (the same standard the EU adopted via UN R154). AIS:175 covers M1 passenger cars plus M2 / N1 vehicles with GVW ≤ 3,500 kg, including Type I (WLTP cycle), Type IV (evaporative emissions), Type V (durability), OBD, Real Driving Emissions (RDE) and In-Service Conformity (ISC) requirements.
CAFE 2027 vs AIS:175 — an important separation
AIS:175 is the test standard. Its scope (M1 + M2 + N1 ≤ 3,500 kg) is broader than CAFE 2027, which applies only to M1 passenger cars. The CAFE 2027 draft itself remains on the MIDC cycle until MoRTH separately notifies the WLTP switch-over for CAFE purposes, and a MIDC→WLTP conversion factor for CAFE 2027 targets will be separately notified by the Ministry of Power in consultation with BEE (Section 4(8) of the draft). Manufacturers must already report both MIDC and WLTP performance values (pi(MIDC) and pi(WLTP)) for models sold from 1 April 2026 onwards.
| Parameter | MIDC | WLTP (per AIS:175) | Change |
|---|---|---|---|
| Cycle duration | ~1,180 s | ~1,800 s | +53% |
| Cycle distance | ~11.0 km | ~23.3 km | +112% |
| Maximum speed | 90 km/h | 131.3 km/h | +46% |
| Standards basis | AIS-137 / CMVR 115 | UN GTR 15 Amendment 6 / UN R154 | — |
| ICE CO2 multiplier | 1.00 | TBD — conversion factor to be separately notified by MoP | — |
Why the 1.18× multiplier appears so widely
A ~18% uplift on ICE CO2 readings under WLTP vs MIDC is the working assumption circulating in trade press, derived from the EU's NEDC→WLTP shift (where the JRC measured ~21% on average). It is not a number formally adopted in the CAFE 2027 draft or AIS:175. Until BEE / MoP issue the formal MIDC→WLTP conversion for CAFE purposes, treat any per-OEM WLTP CAFE projection as an estimate.
CAFE does not set the same target for every OEM. The mass curve adjusts each maker's individual target up or down from the headline c value based on its weighted-average unladen mass W. Heavier fleets get a more lenient target on the basis that heavier vehicles are physically harder to make efficient. Per the 25 September 2025 draft, the formula constants a and b are held constant across all five years — only c moves.
| Parameter | CAFE-II (S.O. 5020(E)) | CAFE 2027 (Sept 2025 draft) |
|---|---|---|
| Constant multiplier a (L/100km per kg) | 0.0028 (notified) | 0.002 (flatter, helps small-car OEMs) |
| Fixed constant b (kg) | 1,037 | 1,170 |
| Units of output | L/100 km (petrol equivalent) | L/100 km (petrol equivalent) |
| Maruti effective target FY28 (W=1,000 kg) | ~4.84 L (industry-target equivalent) | 0.002 × (1,000 − 1,170) + 3.7264 = 3.386 L/100km ≈ 80.3 g CO2/km |
| Mahindra effective target FY28 (W=1,700 kg) | ~4.78 L | 0.002 × (1,700 − 1,170) + 3.7264 = 4.786 L/100km ≈ 113.5 g CO2/km |
| Spread across 1,000-1,700 kg | ~2 L/100km gap | ~1.4 L/100km gap (~33 g/km) — heavier fleets get a much more lenient target |
The b = 1,170 kg revision clause
Section 4(7) of the draft lets the Central Government revise b downward if the average unladen mass of all M1 vehicles sold during calendar 2026 turns out to be below 1,170 kg. In that case, the actual industry-average mass for 2026 becomes the new b. This is a fairness backstop in favour of small-car OEMs if the small-car share of the market unexpectedly grows.
⚠ Layer 2 — February 2026 revised draft (per The Hindu, 25 Feb 2026)
BEE circulated a revised draft to industry in February 2026 (not yet made public). Per reporting by Jagriti Chandra in The Hindu: the slope a is no longer 0.002 flat across all years — it now starts at 0.00153 in Year 1 (FY28) declining annually to 0.00128 in Year 5 (FY32). The effect: permissible emissions for heavier vehicles tighten, while small cars get some relief. This re-balances the lobbying outcome and inverts the impact shown in the table above. The PMO held a review meeting on 25 February 2026 with MoP, BEE, and MHI but reached no final decision. Until the revised draft is notified, both the September 2025 published text and the February 2026 circulated revision sit on the table.
The practical effect: small-car-heavy OEMs (Maruti and the small-end of Hyundai-Kia) face a much tighter individual target than the industry headline c. SUV-heavy OEMs (Mahindra, Tata's heavier mix, the premium Europeans) get substantial relief because their W is well above 1,170 kg. This is the lobbying fault-line, and we return to it in Part 4. Earlier consultation drafts (June 2024 BEE proposal) had a flatter slope and different constants — many trade-press summaries reference those earlier values rather than the September 2025 notified text.
Super-credits (BEE's official terminology: "volume derogation factors vi for super-credits") reward electrified and alternative-fuel vehicles by inflating their effective unit count in the denominator of the fleet-average calculation. The official Table 3 below is the complete list — earlier consultation drafts had different values and categories that have not survived into the September 2025 text.
| Vehicle type | Volume derogation factor vi |
|---|---|
| Battery Electric Vehicle (BEV) / Range-Extender Hybrid Electric Vehicle | 3.0 |
| Plug-in Hybrid Electric Vehicle (PHEV) / Strong Hybrid Electric Vehicle (Flex Fuel Ethanol) | 2.5 |
| Strong Hybrid Electric Vehicle (SHEV) | 2.0 |
| Flex Fuel Ethanol Vehicles | 1.5 |
| ICE (default) | 1.0 (no super-credit, treated at face value) |
What the September 2025 draft did NOT include
Mild Hybrid Electric Vehicles (MHEVs) are not in Table 3. They are not recognised separately and are treated at 1.0× alongside ICE. The 1.6× SHEV value widely reported in trade press belonged to an earlier consultation draft and has been replaced by 2.0× in the notified text. Range-Extender Hybrids are grouped with BEV at 3.0×, which is the most aggressive super-credit category in the entire schedule.
At a 15% BEV mix, the math: 85 ICE × 1 + 15 BEV × 3 = 130 effective units in the denominator. Combined with BEVs contributing 0 g/km to the numerator, the effect on fleet-average CO2 is much larger than the headline 15% share suggests. A 2.0× SHEV (vs the earlier 1.6×) materially improves the strategic position of Toyota Kirloskar, Maruti's SHEV-anchored hedge, and Honda's e:HEV-only India strategy.
Why the multipliers are the policy hinge
These multipliers are the single biggest determinant of whether the CAFE 2027 credit market clears with structural oversupply or undersupply. If BEE haircuts the 3.0× BEV multiplier mid-cycle, the entire OEM strategy book re-prices. The EU's Reg 2019/631 trajectory (peak 2.0× in 2020, phased to 1.0× by 2023) is the natural precedent. We model both scenarios in Part 6.
The September 2025 draft has three separate downward adjustments to a manufacturer's declared performance pi. Each is governed by a different sub-section and is more nuanced than the 17-tech, 6 g/km cap summary widely reported in trade press.
Each technology must be demonstrated to the Type Approval Agency, must not be entirely driver-dependent, and must achieve a minimum reduction of 1 g CO2/km on MIDC (or WLTP when notified). Savings are certified as a factor (0-1) or in absolute terms (g CO2/km) at the Test Agency's discretion. The draft does not publish a closed list of 17 specific technologies — that list circulated in trade press (Business Standard, Feb 2026) as a likely shortlist but is not in the notified text.
Per the published 25 September 2025 draft, petrol vehicles with unladen mass ≤ 909 kg, engine capacity ≤ 1,200 cc, and length ≤ 4,000 mm could claim an additional 3.0 g CO2/km reduction in pi, over and above any certified ci. This was a direct policy response to small-car lobbying. Per The Hindu (25 February 2026), this 3.0 g/km waiver has been REMOVED in the revised draft circulated to industry in February 2026. The revised draft compensates small-car OEMs instead through the variable slope (see mass-curve section above) — substituting per-model relief with curve-shape relief.
The 9 g/km cap that matters
No model i (including its variants) may claim a cumulative reduction exceeding 9.0 g CO2/km in any reporting period. This is the actual draft cap — not the 6 g/km figure widely cited in trade press (which referred to an earlier consultation proposal). The 9 g/km cap is binding on the sum of ci + small-car bonus.
Entirely separate from ci, the CNF discounts tailpipe CO2 in recognition of biofuel blending:
For an OEM with a heavily E20-petrol fleet, the 8% CNF alone meaningfully reduces compliance distance. For Toyota's flex-fuel Hycross or any future SHEV flex-fuel, the 22.3% CNF is a very large discount — this is the policy signal that India is using CAFE to push ethanol blending alongside electrification.
Important correction to widely-circulated summaries: the September 2025 BEE draft does not introduce block-period averaging. Section 4(9) of the draft is explicit: "In any fiscal year commencing from the 1st day of April, 2027 onwards, the Annual Average of Actual Fuel Consumption shall be less than or equal to Annual Average Fuel Consumption Standard ... of the respective fiscal year." Compliance is annual.
What the draft does introduce, in Section 7, is pooling:
The Block I / Block II story — where it came from
The "Block I FY28-30 + Block II FY31-32" structure described in some trade press traces back to a Ministry of Power presentation to the PMO reported by the Times of India on 30 March 2026. That presentation appears to envisage block-period pooling of surplus credits across the 5-year cycle (first 3 years, then 2 years). The published 25 September 2025 draft does not yet codify that — under the current draft, surplus carry-forward across years is not allowed for an individual OEM, only intra-year pooling across up to 3 manufacturers. Watch the final notification for this provision.
Pooling fundamentally changes the credit-market dynamic. Skoda-VW + Audi + a third European can net into a single position. Tata + a smaller surplus-side OEM can stabilise credit revenue across a thin year. The 3-OEM cap is the cap that matters — without it, a single "compliance trust" structure could absorb the entire industry, which BEE clearly intends to prevent.
The TerraNova View
Block-period averaging changes the year-by-year cash flow profile materially. TerraNova projects BAU vs With-plan across both blocks side by side.
See the Compliance Manager →For the first time, OEMs in deficit can buy credits from OEMs in surplus instead of paying the penalty. The mechanism is OTC (over-the-counter) with BEE running a passbook ledger and acting as a market-maker through a direct sell-leg.
| Element | Design |
|---|---|
| Trading venue | OTC, peer-to-peer, recorded in BEE passbook |
| BEE direct sell-leg | ₹2,500-4,500 per g/km-unit — serves as soft floor and ceiling on the OTC market |
| Banking | Surplus credits from Block I carry forward to Block II |
| Pooling across OEM groups | Not allowed in the current draft — subject to lobbying (Skoda-VW, premium Europeans pushing for it) |
| Cross-market fungibility (CCTS, RECs) | Not allowed in the current draft |
| Tenor | Credits valid within the block they are earned and the immediately following block |
How the market clears
When supply exceeds demand by 1.5× or more, the market clears at the BEE floor of ₹2,500. When supply is below 0.85× demand, it clears at the ₹4,500 ceiling. In between, the price moves linearly through that band. Part 6 walks through the full clearing math, the four scenarios, and what each implies for OEM strategy.
Part 3 maps every passenger-vehicle OEM in India against the CAFE-II margin, BEV mix, curb weight, and 5-year strategic role — from Maruti's 1.82M-unit volume exposure to Tata's structural credit surplus, the premium European pooling lobby, and the pure-EV import pack.
Part 1 — CAFE Primer · Part 4 — The Open Issues · Part 6 — The Credit Market Modelled
Primary regulatory sources and verified analysis cited above.
From per-OEM cost exposure to the 5-year credit-market view, Climate Decode helps Indian passenger-vehicle OEMs sequence the CAFE-III response with finance-grade clarity.