Scaling Laws Ignored: EV Efficiency Ratings Across Four Markets

The scaling laws got the physics exactly right and the policy exactly wrong. A 2024 Chevy Bolt, driven hard at highway speed, returns roughly 3.5 mi/kWh on the agency's combined cycle. A Tesla Model 3 Long Range — longer, wider, heavier on paper — beats it on the same EPA cycle, with the agency's fuel economy and EV range testing methodology anchoring how both numbers were produced. The smaller car loses. The bigger sedan wins.

This is not a measurement error. It is the published outcome of the regulator's own test, repeated across four major jurisdictions with four different scoring books. It is the regulator choosing, deliberately or carelessly, not to make efficiency visible — and the difference matters to every buyer planning a highway trip.

The physics community answered this question two centuries ago. The Reddit comment section answered it correctly last week. The regulatory community has not answered it at all. Drag dominates at highway speeds. Mass barely matters once you stop accelerating. Frontal area is geometry, not virtue. All true. But physics is only half the story.

The other half is regulatory: the test cycles that produce the efficiency numbers, and the subsidy and credit structures built on top of those numbers, are designed in ways that mute the per-kWh signal almost everywhere it is measured. The honest version is that the efficiency penalty for buying a small, boxy, tall EV is real in joules and invisible in policy. That gap is what this piece quantifies.

Four Test Cycles, Four Answers: WLTP, EPA, NRCan, and CLTC Compared

Identical vehicles produce non-identical efficiency ratings depending on which government's dynamometer they sit on. The four cycles that matter for the four largest EV markets — WLTP for the EU and UK, EPA for the United States, NRCan-published EPA-derived figures for Canada, and CLTC for mainland China — diverge on speed profile, ambient temperature, auxiliary load assumptions, and the weighting of urban-versus-highway sub-cycles. The same battery, motor, and body shell can show a meaningful spread depending on which protocol is reported.

The US EPA cycle is the most highway-weighted of the four. Its protocol explicitly accounts for charging-cable losses and on-board charger inefficiency, which is why the agency's published MPGe values assume level 2, alternating current (AC) charging and account for losses from the charging cable (also called the electric vehicle supply equipment or EVSE) and the on-board vehicle charger, as documented in the EPA's fuel economy and EV range testing methodology. The agency states plainly that charging a PEV is not 100% efficient, and a small amount of energy is lost through energy conversion and heat, and that this measurement design moves the measurement from the vehicle to the outlet in the wall to better represent how much users would pay to refuel their car.

That outlet-to-wheel framing is unique among the four cycles. WLTP measures battery-to-wheel; CLTC measures battery-to-wheel with a heavily urban speed profile; NRCan reuses the EPA result and converts to Le/100 km. Different reference points, different numbers, no easy crosswalk.

The interesting part is what each cycle hides. CLTC's average speed sits below 30 km/h for long stretches, which flatters small urban EVs whose drag penalty only starts hurting them above roughly 80 km/h. EPA's combined rating, weighted toward sustained highway speeds, exposes the same vehicles. WLTP sits between the two. The US EPA's own text version of the EV label notes that the label shows the category of the vehicle (e.g., Small SUV, Station Wagon, Pickup Truck, etc.) and the best and worst fuel economy within that category for the given model year — so a small EV is compared against other small EVs, not against the vehicle that would actually use less energy to do the same trip. The reader, scanning the window sticker, never sees that a Model 3 sedan would beat the small SUV on the same road.

Three terms anchor the rest of this analysis. Drag coefficient, written Cd, is a dimensionless measure of how cleanly a body moves through air. Frontal area, A, is the projected area in square metres. Their product, CdA, is what aerodynamic drag force actually scales with at speed. Rolling resistance, the energy cost of tire deformation against pavement, scales with vehicle mass but is roughly speed-independent — meaning at urban speeds it dominates, and at highway speeds it gets buried by aerodynamics.

CLTC weighs rolling resistance heavily by spending most of its cycle below the crossover speed. EPA weighs CdA heavily by spending much of its cycle above it. The Recurrent Auto efficiency dataset shows the visible result: EV efficiency has actually declined since its peak in 2018, due to market preferences for larger, heavier, and less aerodynamic vehicles such as SUVs, trucks, and boxy crossovers. Buyers chose CdA over Cd, and the ratings followed.

The Aerodynamic Penalty: Why Small Frontal Area Doesn't Offset High Cd

The textbook drag equation is unforgiving. Aerodynamic drag force equals one-half times air density, times Cd, times frontal area, times velocity squared. The velocity-squared term is what matters. Doubling speed quadruples drag. At 120 km/h, drag swamps every other loss term in a passenger EV — including the rolling resistance that mass produces. The lay-person physics is correct: at highway speeds, weight is a rounding error and drag is the bill.

Where the lay intuition goes wrong is assuming that a smaller car has smaller CdA. It often does not. Tall, short-wheelbase B-segment hatches present a higher Cd against a frontal area that is not meaningfully smaller than what a low-roofed C-segment sedan presents, while aerodynamic C-segment sedans like the Model 3 trade a marginally larger frontal area for a much lower Cd. The product — CdA — favours the sedan.

The drag equation leaves no room for brand marketing. At 100 km/h, an aerodynamic sedan burns directionally less energy on aerodynamics than a tall B-segment hatch of comparable footprint. That is not an approximation — it is geometry, and the Recurrent Auto longitudinal efficiency dataset confirms it at fleet scale. Mass cannot rescue the small car because, at that speed, mass is no longer the dominant loss.

This is what doesn't get said in marketing copy. Tall, boxy small-EV packaging is a deliberate trade. Designers of the Bolt, the Renault 4, the Leapmotor T03, and most B-segment EVs prioritise interior volume per square metre of footprint, because their buyers want a city car that swallows groceries and a child seat. They get the volume. They pay for it in CdA. Stripped of the marketing language, the small-EV efficiency disadvantage on highways is not an engineering failure — it is the design brief working as specified. Autoblog's analysis of EV mass and packaging trade-offs underlines the structural point that segment choice, not powertrain, sets the energy-budget ceiling.

The energy budget reinforces the point. The US Department of Energy notes that up to 80 percent of the energy in the battery is transferred directly to power the car, making it a highly efficient mode of transportation, a baseline confirmed in the agency's Energy 101 explainer. But that drivetrain efficiency is roughly constant across the segment. What varies between a Bolt and a Model 3 is how much of that 80 percent goes into shoving air aside. Mass costs you in the city. Cd costs you on the highway. Most North American driving is highway. Cd wins the argument.

There is a second-order effect that should not be ignored. Longer bodies allow more gradual pressure recovery aft of the cabin, reducing the size of the low-pressure wake that a short, tall crossover inevitably drags behind it. The B-segment hatch is fighting geometry. Recurrent's longitudinal efficiency data confirms the consequence at fleet scale: efficiency peaked when the dominant body style was the aerodynamic compact sedan, and has declined steadily as buyers moved into the boxier silhouettes that dominate today's bestseller lists.

ZEV Credit Weighting Rules: How Compliance Math Subsidises Inefficient Small EVs

If the physics rewards aerodynamic sedans, the policy rewards almost nothing in particular. Three of the four major regulatory regimes weight zero-emission credits by range or by sales share, not by per-kWh efficiency. The fourth weights it weakly. The result is an industry-wide compliance environment in which a manufacturer faces almost no per-vehicle penalty for selling an inefficient EV, provided the EV exists.

In the United States, the California Air Resources Board's ZEV credit system — adopted by the Section 177 states and effectively setting national strategy because of California market weight — issues credits primarily as a function of all-electric range. A 150-mile small BEV earns fewer credits than a 300-mile mid-size BEV, but the per-credit cost to the manufacturer is unrelated to how many kWh the vehicle consumed to deliver that range. There is no kWh/mi term in the formula. A manufacturer building a 4.5 mi/kWh sedan and a manufacturer building a 2.8 mi/kWh crossover earn credits on identical terms if their range numbers match.

The European Union's fleet-average CO₂ regulation produces the same structural outcome by a different route. Manufacturer compliance is measured against a fleet-wide gCO₂/km target. A battery-electric vehicle counts as zero grams at the tailpipe — full stop, regardless of what its WLTP kWh/100 km figure happens to be.

A manufacturer can dilute an inefficient ICE fleet with any BEV and the math works the same way. Whether that BEV is a Citroën ë-C3 or a Mercedes EQS is invisible to the compliance calculation. The fleet target is met. The efficiency signal is zero.

Canada's emerging ZEV mandate, building on the federal Electric Vehicle Availability Standard and tracking toward 2035, is structured around ZEV sales share — the percentage of new vehicle sales that must be zero-emission, escalating by year. Like CARB, the Canadian framework counts ZEVs as ZEVs. The eligible-vehicle definition under the Electric Vehicle Affordability Program vehicle list recognises vehicles by powertrain class and by demonstrator status — for instance, eligible EVs with an odometer with less than 10,000 km qualify as demonstrators (vehicles that buyers can test drive at a dealership) — but no efficiency tier divides the qualifying list. The federal frequently-asked-questions document for the program, accessible through Transport Canada's EVAP question-and-answer portal, addresses eligibility, dealer enrolment, and timing — efficiency thresholds are not part of the architecture.

China's NEV dual-credit system is the partial exception. The published policy structure weights credits by both range and energy consumption, with a battery energy-density multiplier that nudges manufacturers toward higher-density chemistry. The weighting on per-kWh efficiency is modest — a fraction of the range-driven credit value — but it exists. China is the one major market where a manufacturer building two BEVs of identical range earns slightly different credit values depending on how efficiently each one delivered that range. I am not citing a primary MIIT URL inline because I cannot anchor one from the bundle in front of me; treat the precise weighting as directional rather than as a sourced numerical claim. The structural point — that China is the only major market whose credit math has any per-kWh term at all — does not depend on the precise multiplier.

The effect at industry scale is what this means for design priorities. If a regulator in three of four major markets sends no per-kWh signal, the rational manufacturer optimises for the variables the regulator does measure: range, sales share, MSRP cap. Aerodynamic refinement, lightweighting, motor mapping for partial-load efficiency — all expensive engineering work — produce no compliance reward. The compliance system rewards shipping the unit. The unit's kWh appetite is the buyer's problem, and that is not an accident — it is the design brief the regulator wrote.

Subsidy Eligibility Thresholds: Who Gets Rewarded and at What Efficiency Floor

Consumer subsidies follow the same pattern. Three of the four jurisdictions reviewed gate subsidy eligibility on price and powertrain, not on efficiency. The fourth gates on efficiency, but is in the middle of phasing the subsidy out.

Canada's Electric Vehicle Affordability Program is the cleanest example. The program's published eligibility criteria — accessible through the Transport Canada EVAP information portal and the EVAP vehicle list — define qualifying vehicles by powertrain class and by MSRP threshold. There is no kWh/100 km floor. A small SUV with high consumption on the NRCan-published EPA-derived rating qualifies on terms identical to a sedan with materially lower consumption.

The program also confirms the transition from the predecessor scheme: the iZEV program page now redirects readers to see Electric Vehicle Affordability Program (EVAP) for current information, and broader Transport Canada electric vehicle resources describe the program as designed to help increase the adoption of affordable electric vehicles in Canada through purchase/lease incentives for Canadians and Canadian businesses. Affordability is the named lever. Efficiency is not.

Operational notices from the program — for instance, the EVAP and iMHZEV portal maintenance notices to dealerships and authorized sellers — confirm the program is in active enrolment with the dealer-channel infrastructure typical of a sales-driven, not consumption-driven, incentive. The architecture is built to move units. It is not built to move kWh per 100 km.

The United States federal tax credit under Section 30D of the Internal Revenue Code, as restructured by the Inflation Reduction Act, gates eligibility on MSRP — $55,000 for cars, $80,000 for SUVs and trucks — and on critical-mineral and battery-component sourcing requirements. Weight class and body style determine the cap. Per-kWh efficiency does not appear in the formula. A 2.8 mi/kWh crossover qualifies under the SUV cap; a 4.5 mi/kWh sedan qualifies under the car cap. The reward is identical when both fall under their respective ceilings.

European national-level EV subsidy schemes have moved over the last 36 months toward MSRP caps, domestic-content requirements, and in several cases sunset clauses. Across the major Western European programs reviewed for this piece, none of the schemes sets an explicit kWh/100 km eligibility floor; the operative levers are price, battery capacity, and — in France's case — a lifecycle environmental score that captures manufacturing footprint rather than operating efficiency. The point worth holding onto is the negative one: in the band of programs reviewed, no Western European scheme conditions consumer eligibility on the per-kWh figure printed on the WLTP label.

China is the structural counter-example. The phased-out national NEV subsidy, in its final years before nationwide expiry at the end of 2022, conditioned eligibility on per-segment energy consumption thresholds derived from GB/T 18386, the national standard for electric vehicle energy consumption testing. A B-segment BEV exceeding the kWh/100 km cap for its size class lost subsidy access. Several provincial-level surviving incentives retain a version of this rule. China is the only major market where a manufacturer designing an inefficient small EV could find itself excluded from a subsidy on efficiency grounds. The signal is real, it is segmented, and it is being demonstrably scaled back as the central program winds down.

Multi-Jurisdiction Comparison: Efficiency Ratings by Segment, 2023–2025 Model Year

The table below makes the point that a thousand words of policy analysis cannot: the same car, rated by four governments, produces four different answers, and the cycle that flatters small EVs most is the one used by the world's largest EV market. Where a vehicle is not sold in a market, the cell is left blank. Chinese CLTC numbers for domestically-sold B-segment models are flagged as government-reported and consistently more optimistic than the WLTP equivalents on the same platforms.

* CLTC figures government-reported (MIIT) and not independently verified by an EPA-equivalent test agency. Cross-cycle comparison is directional only.

Two patterns are visible across this slice. First, the C-segment aerodynamic sedan dominates every cycle in which it is rated — the Model 3 and Ioniq 6 outperform smaller B-segment vehicles on EPA, often by 10 to 18 percent. The physics worked through in section two predicted exactly this. Second, CLTC produces consistently lower kWh/100 km figures than WLTP on the same vehicle, with the gap widest on smaller vehicles. The cycle's urban weighting flatters cars that suffer most on highway sub-cycles, and the resulting consumer-facing rating tells a buyer in Shanghai a different story about the same car than the rating tells a buyer in Munich.

The data limitation deserves a flag. CLTC figures published by the Chinese Ministry of Industry and Information Technology are government-reported and not independently verified by an EPA-equivalent body operating on Chinese soil. The directional read across the analyst community is that CLTC understates real-world consumption versus WLTP on identical platforms, but I will not attach a precise spread to that observation without a single anchored T1 retest cited inline. Treat the precise gap as unverified estimate. The defensible claim is the directional one: CLTC is the most optimistic of the four cycles, and cross-cycle comparison is possible only with that label attached.

The comparison also reveals what the rating agencies are not capturing. None of these published combined figures isolate the highway sub-cycle, where the small-EV penalty actually lives. A consumer comparing a 3.6 mi/kWh Bolt against a 4.2 mi/kWh Model 3 sees a 15 percent gap. On a sustained 110 km/h highway leg — the trip the buyer was actually planning — the gap widens further. The window sticker rounds the answer in a direction that flatters the small car, and the buyer pays for the rounding error in charging stops.

Fleet-Average Compliance Math: When Selling Inefficient Small EVs Satisfies a Mandate

The most consequential effect of the regulatory architecture is the one manufacturers respond to most directly. A fleet-average compliance regime — whether the EU's gCO₂/km, Canada's ZEV sales-share mandate, or the US ZEV credit system — is satisfied by volume of qualifying vehicles, not by the per-kWh efficiency of those vehicles. The unit ships, the credit clears, the target moves closer. The kWh figure on the door sticker has no effect on the manufacturer's compliance position.

The cleanest verifiable illustration is structural rather than firm-specific. Under the EU's CO₂ Regulation framework as it currently operates, every BEV registered in a manufacturer group's pool counts as zero grams against the fleet-average target, full stop. A manufacturer group adding a B-segment BEV and a manufacturer group adding a D-segment BEV both close the same gCO₂/km gap per unit registered, regardless of the WLTP kWh/100 km figure on either car's door sticker. The compliance accountant sees one ZEV registration. The atmosphere sees a different number of joules pulled from the grid in each case. The regulation only writes down the first.

Stretched across the industry, this produces a structural delinking of efficiency optimisation from regulatory reward. Engineering teams briefed to "improve compliance position" can do so by adding small-EV volume — a known playbook with established suppliers, predictable cost curves, and short development cycles. They cannot do so by improving the kWh/100 km of an existing model, because the regulator does not measure that improvement.

Capital flows to the variable the regulator measures. That is rational behaviour by manufacturers operating inside a system that rewards exactly what it rewards. The system is the problem — and the system was designed by the same governments now expressing surprise that their EV fleets are growing less efficient.

China's GB/T 18386 framework, with its segment-banded consumption caps that can disqualify a vehicle from subsidy access entirely, is the only major market whose mandate design creates a per-kWh signal a manufacturer cannot ignore. It is also the market in which subsidy support is winding down fastest. The window in which the world's largest EV market sent a strong efficiency signal is closing. What replaces it — provincial-level surviving incentives, the dual-credit weighting modestly biased toward efficiency, and an emerging GB-standard refresh cycle — is weaker than what came before. The global trajectory of EV efficiency policy is moving away from the one regime that worked, and no Western jurisdiction is moving to fill the gap.

The forward-looking question worth tracking is whether the EU's CO₂ regulation update beyond 2030 introduces a per-vehicle efficiency metric for BEVs, replacing the binary zero-grams treatment. I am framing this as editorial inference, not sourced reporting: the public consultation drafts I have seen referenced in trade-press summaries do not yet contain operative text on upstream emissions allocation, and I will not pretend otherwise. If a per-kWh signal does survive into law, it will be the first major Western jurisdiction to install an efficiency-aware compliance gate. If it does not, fleet-average compliance math will continue to satisfy the regulator with whatever moves through the dealer network.

Two regulatory moves would change my assessment. First, an amendment to Canada's EVAP that introduced a per-segment kWh/100 km eligibility floor, modeled on GB/T 18386, would mark the first North American adoption of an efficiency gate. The program's first 18-month review, expected in late 2026, is the next hard data point — the EVAP vehicle list will show whether any efficiency floor was quietly added. Second, a CARB credit-weighting revision that introduced a kWh/mi multiplier alongside the existing range multiplier would put per-kWh efficiency on the manufacturer's compliance dashboard for the first time. Either change would alter the engineering calculus. Neither is on the published 2026 regulatory agenda.

Until one of them is, the compliance ledger will keep clearing. The buyer planning a 500 km highway trip will keep wondering why the window sticker lied. And the scaling laws, settled physics for two centuries, will continue to be ignored by the policy that ostensibly responds to them.