Two administrations, the Biden administration and second Trump administration, reached opposite conclusions about whether vehicle emission standards should be rolled back. These conclusions came from a set of different assumptions, including whether there is a value to society from avoiding air pollution and the degree to which car buyers value the future fuel savings from getting a more efficient vehicle. Both of these two assumptions really matter and the disagreement on either of them accounts for hundreds of billions of dollars in the government's cost-benefit math on vehicle emissions standards — math that, in this year's rollback, also assumes gas prices are lower than what drivers are actually paying at the pump.
That's the finding of a new analysis in Science co-authored by Kenneth Gillingham, the Grinstein Class of 1954 Professor of Environmental and Energy Economics at the Yale School of the Environment, along with nine other economists from institutions across the country. When the Trump administration rescinded the EPA's greenhouse gas standards for vehicles this spring, it backed the move with the cost-benefit analysis that is required of any economically significant federal rule. It concluded that Americans would be $600 billion to $790 billion better off without the standards. Gillingham and his co-authors show that correcting just one of the analysis's assumptions is enough to turn that benefit into a net cost.
"My co-authors and I have been carefully following the regulations affecting vehicle efficiency and fuel type, and we observed enormous swings in the analysis between administrations," Gillingham said. "If the numbers are being calculated correctly, one or both of the analyses must be incorrect."
The authors argue that ascribing a zero value for avoided air pollution, as in the latest Trump administration rulemaking, is obviously incorrect, but the assumption about the valuation of vehicle efficiency is more nuanced. This assumption on valuation sounds narrow, but it isn't: How much are car buyers actually willing to pay for a vehicle that saves them money on gas? Research has long shown that buyers pay less than a dollar today for a dollar's worth of future fuel savings. Economists attribute that gap to two possible causes: consumers who simply don't pay close enough attention to long-run fuel costs, a bias often known as "inattention," and "hidden costs" — attribute trade-offs where automakers save fuel by compromising on other attributes, such as weight, performance, or comfort, that consumers also value.
The issue, the authors show, is that the EPA's analyses have assigned that gap to different causes over time. The 2024 rule attributed the entire gap to the behavioral bias of inattention, which means that consumers could be better off in the long run if they invest in more efficiency. The 2026 rule assumes the opposite, attributing it entirely to hidden costs from attribute trade-offs, so that consumers would gain by paying less at the pump by improving efficiency, but lose out in other attributes that they value. That single modeling choice changes the EPA's estimate by hundreds of billions of dollars.
Drawing on a systematic review of the peer-reviewed literature, the authors propose a corrected breakdown: roughly 55 cents of every dollar in "missing" fuel savings reflects inattention, while about 23 cents reflects genuine attribute trade-offs — based on recent EV studies. The remaining 22 cents represent savings that consumers are aware of and fully value.
"A careful review of the evidence suggests neither administration got it quite right, though the 2024 Biden-era analysis better reflects the academic evidence on consumer inattention and is more internally consistent," said Arthur van Benthem, professor of business economics and public policy at Wharton School of the University of Pennsylvania and a study co-author."
The authors note that part of reason that the 2026 modeling choice is more problematic is that the 2026 analysis contradicts itself: To estimate how much a more fuel-efficient car costs to build, the EPA assumes automakers improve fuel economy without changing anything else about the car — same weight, same power, same comfort. But to explain why consumers undervalue future fuel savings, that same analysis assumes the opposite: that fuel-efficient cars do compromise on weight, power, or comfort, and that consumers are absorbing those hidden costs. Both assumptions cannot be true at once, the authors noted.
There's also a problem with how the analysis treats driving: When a less efficient car costs more to fuel, people drive less. The EPA counted the money drivers saved from those skipped trips as a benefit but never subtracted the value of the trips they gave up — even though forgoing a trip is itself a cost to the driver. Correcting this error, along with the valuation split and the technology-cost inconsistency, reverses the EPA's result entirely: The agency's claimed $790 billion net benefit at a 3% discount rate becomes a $670 billion net cost.
The analysis also relies on asymmetric fuel-price assumptions: The 2026 rule pairs its central forecast with a low-price scenario but leaves out the corresponding high-price scenario — even though actual gas prices have already climbed past the central case. Since lower assumed gas prices mechanically shrink the value of fuel savings, that omission tilts the analysis toward the rollback before the valuation debate even begins.
Combined, the corrections wipe out roughly $1.46 trillion of the administration's claimed benefit, the authors calculate — and that's before restoring the $1.8 trillion in environmental and public health benefits that the 2026 rule excludes entirely by ignoring the climate and air pollution-reduction value of the standards.
"Our findings indicate that the basis for rollback is riddled with inconsistencies and hard-to-justify assumptions, which may put it in legal jeopardy," Gillingham noted.
The stakes go beyond this one regulation. Standards have swung back and forth for years under conflicting readings of the same evidence, making it hard for automakers to plan. Gillingham and his co-authors argue that a consistent, evidence-based approach to consumer valuation, and a consideration of alternatives — like fuel-consumption-based carbon taxes, feebates, or mileage taxes that price fuel use directly rather than regulate vehicle design — could help to stabilize policy regardless of which party controls the White House.
