Facebook Twitter Tumblr Close Skip to main content
A Project of The Annenberg Public Policy Center

The Facts on Fuel Economy Standards


The Trump administration is in the process of rolling back Obama-era fuel economy standards, which were originally set to hit an average of 54.5 miles per gallon for passenger cars and trucks by 2025.

While not yet finalized, the Trump administration’s proposed rule freezes fuel economy standards at 2020 levels through 2026 — a move the government says will save lives and money.

But many experts, and even an official inside the Environmental Protection Agency, are skeptical of those new conclusions. Researchers have identified numerous errors and faulty assumptions in the administration’s latest analyses, some of which violate the basic principles of economics.

For example, Mark Jacobsen, an economist at the University of California, San Diego, said the administration implausibly assumes that the U.S. will have about 6 million fewer cars on the road by 2029 if the rollback happens. The Trump administration has cited Jacobsen’s research in pushing for the rollback, but he said his work is being misapplied, resulting in misleading conclusions about fatalities and pollution.

William Charmley, an official with the EPA’s Office of Transportation and Air Quality, wrote in an internal staff memo that the proposed standards would increase the fatality rate and “are detrimental to safety.”

Under President Donald Trump, the EPA also has said it will remove California’s unique ability to set stricter emissions standards for vehicles. California currently accepts federal standards, but doesn’t want to continue to accept them if they are frozen at the 2020 level. If EPA revokes California’s special power, that decision would likely be challenged in court and might lead to a patchwork of regulations across the country — an outcome many automakers would prefer to avoid.

We’ll explain how U.S. fuel economy regulations work, go over what President Barack Obama implemented, and take a look at the Trump administration’s justification for the proposed rollback.

Two Fuel Economy Programs

The idea behind a fuel economy standard is to push automakers to produce vehicles that travel further on the same amount of fuel, thereby reducing the need for gasoline and decreasing pollution.

Today, the U.S. fuel economy program technically includes two standards that are overseen by different agencies under different laws.

The Corporate Average Fuel Economy, or CAFE, standard is set by the Department of Transportation’s National Highway Traffic Safety Administration, or NHTSA. The standard is set in miles per gallon, and was first used on model year 1978 cars and 1979 trucks following the 1973-74 oil embargo, which had drastically raised gas prices. At the time, the motivation for the standard was to make the nation less dependent on foreign oil.

Following a 2007 Supreme Court decision that carbon dioxide and other greenhouse gases are air pollutants under the Clean Air Act, the EPA concluded in 2009 that greenhouse gas emissions from vehicles endanger public health and welfare. This allowed the EPA to create a greenhouse gas emissions standard for cars and trucks, which is set in grams of emitted carbon dioxide, or the CO2-equivalent of another greenhouse gas, per mile traveled.

Under Obama, NHTSA and EPA worked together to set standards that would be essentially equivalent to each other. The agencies also collaborated with the state of California, which had been allowed to set a stricter standard under the Clean Air Act, so that the state would accept the federal standards, thereby creating a so-called “national program.”

Obama first set new standards in this way for model years 2012-2016, and also included the United Auto Workers union and major automakers when brokering the deal. In 2011, he then announced a second phase of fuel economy improvements for model years 2017-2025. It’s the latter half of this second phase of improvements that the Trump administration has proposed to eliminate.

How the Standards Work

While the standards are often summarized by a single number, such as Obama’s 2025 goal of 54.5 mpg as an average across all vehicles, that is simply an estimate of what the agency expects under the program.

As a Congressional Research Service FAQ document on the topic says, “the 54.5 number is not a requirement for every — or for any specific — vehicle or manufacturer.”

Each vehicle has a specified fuel economy target, which is based on its size and class. But what matters is the average target across all of the cars that are sold in a given fleet. This means manufacturers can sell underperforming cars and make up the difference by selling other cars that outperform their targets.

Companies face financial penalties if their vehicle fleets do not meet their individually calculated standards. But there are built-in flexibilities that mean automakers still have ways of avoiding penalties if they miss their marks in a given year.

Both NHTSA’s CAFE program and EPA’s greenhouse gas program operate on credit systems that allow companies to stockpile credits if they overperform; those credits can then be used in the future or even applied retroactively. Automakers can also transfer credits between their fleets or trade credits to other manufacturers. As a result, even a company that fails to meet its standard one year could be in compliance because of previous good performance or by credit swapping within the industry.

Finally, it’s worth noting that the standards sound more stringent than they are in reality. As the EPA explains, real-world greenhouse emissions are about 25 percent higher than the CO2 standard, and on-the-road fuel economy is about 20 percent lower than the listed CAFE standard.

Obama’s initial 2025 goal of 54.5 mpg is a mpg version of the CO2 standard. To put that in perspective, the EPA estimates that the gas mileage is closer to 40 mpg. But even that assumes that all of the greenhouse gas improvements come from fuel economy, which is unlikely to be the case, since some are expected to come from other areas, such as air conditioning. The real-world fuel economy will be even lower.

Obama’s Standards

Under Obama’s proposal for 2017-2025, emissions standards would tighten by 3.5 percent every year for the first five model years, and by 5 percent every year for the last four years.

In 2012, the EPA finalized Obama’s proposed increases to the fuel economy standards for all model years, and NHTSA finalized the rules through model year 2021, since the agency could not legally finalize standards for more than five years at a time. Given the long lead time, the EPA also agreed to a process in which the 2022-2025 model years would be reevaluated in a midterm evaluation to see whether they were still appropriate.

In late November 2016, the EPA issued a proposed determination that the rules remained appropriate, and did not need to be made more or less stringent. Based on a technical assessment report and other updated information, including public comments on that report, the EPA found that, if anything, the cost of installing fuel economy technology was lower than projected in 2012.

On Jan. 12, 2017, a week before Trump took office, Obama’s EPA head, Gina McCarthy, issued a final determination that the 2022-25 standards were appropriate.

Although the standards had not changed, the agency’s updated fleet information shifted the average emissions and fuel economy projections slightly. Instead of 163 grams of CO2 per mile, or the equivalent of 54.5 mpg in 2025, the agency now expected 173 grams per mile, or the equivalent of 51.4 mpg. This level of fuel efficiency, the EPA explained, would be met by a car getting about 36 mpg in the real world — some 10 mpg higher than a typical 2016 vehicle.

Trump’s Proposal

Just a few months after Trump took office, his administration announced its intent to challenge the 2022-2025 fuel economy rules.

In April 2017, Scott Pruitt, the EPA administrator at the time, formally withdrew the final determination McCarthy had issued in January, arguing that the standards were “based on outdated information, and that more recent information suggests that the current standards may be too stringent.”

Then, in August 2018, the EPA and NHTSA released a new plan for fuel efficiency called the Safer Affordable Fuel Efficient Vehicles rule, or SAFE, which would apply to model year 2021-2026 cars and trucks.

In the notice of proposed rulemaking for SAFE, the agencies outlined nine different alternatives to the Obama rules, but stated that the preferred option was the most complete rollback: freezing the standards at the 2020 level through model year 2026.

The proposed rule also revokes California’s ability to set its own, higher fuel efficiency standards.

Justification for the Rollback

The gist of the administration’s justification for the rollback is that the Obama-era standard is too expensive to do and is no longer feasible or reasonable for automakers to achieve.

In the notice of proposed rulemaking, which includes a new analysis, EPA and NHTSA said that SAFE would save more than $500 billion in costs to society, for a net gain of about $177 billion.

Freezing the standards at 2020 levels, the agencies said, would also result in 12,700 fewer fatalities from car accidents — a tally that includes deaths over the lifetime of vehicles through model year 2029.

At the same time, the administration concluded the rollback would lead the U.S. to use about half a million more barrels of oil every day — an increase of 2 to 3 percent — and would result in an extra 7.4 billion metric tons of carbon dioxide being released by 2100.

The administration argues that this figure is minimal, since it corresponds to 3/1000ths of a degree Celsius in warming. But it’s equivalent to about four years’ worth of greenhouse gas emissions from the nation’s entire transportation sector.

Criticisms of the Trump Analysis

Numerous economists have come forward to point out flaws and errors in the Trump analysis, suggesting that the costs of the standards are not as high as the administration estimates and questioning the validity of the safety claims.

Flawed modeling: The biggest error, said Mark Jacobsen, an economist at the University of California, San Diego, is that the administration inappropriately assumes that the U.S. will have about 6 million fewer cars on the road by 2029 if the rollback happens.

The assumption runs counter to everything economists know about loosening regulations and lowering the price of vehicles. If costs fall, people will buy more cars, not fewer.

“It’s not plausible,” Jacobsen told us of a fleet size decrease under the rollback. “It can’t happen.”

Jacobsen is one of 11 authors on a policy forum paper published in the journal Science last December that concluded the Trump analysis “has fundamental flaws and inconsistencies, is at odds with basic economic theory and empirical studies” and is “misleading.”

This particular error, Jacobsen said, is a result of a faulty fleet turnover model that attempts to incorporate scrappage, or how often people choose to scrap, or junk, their cars.

Jacobsen has studied scrappage, and the Trump administration has cited his work in pushing for the rollback. While he’s encouraged that scrappage is entering into the government’s thinking — it wasn’t part of the Obama analysis — he said his work is being misapplied and taken to a level “that’s not even feasible economically.”

The reason why a smaller fleet size under the rollback is such a big deal is because it changes all of the subsequent calculations.

As the Science paper explains, fewer cars mean fewer miles traveled, which means less gas consumed, less traffic congestion and noise, less pollution and fewer car accidents. In other words, the Trump analysis isn’t counting any of the negatives for at least 6 million cars, making the rollback appear more beneficial than it is.

Doubled rebound effect: Another point of contention comes from the rebound effect, or the adjustment made to account for the fact that when people have more fuel efficient cars that are cheaper to run, they tend to drive them more.

The 2016 Obama analysis used a figure of 10 percent, meaning that for every 10 percent drop in fuel costs, people would drive 1 percent more. It’s essentially how much of the fuel savings will be lost to the extra driving — in this case, roughly 0.5 percent of the 5 percent increase in fuel efficiency.

The 2018 Trump analysis doubled the rebound effect to 20 percent. But Yale economist and fellow Science paper co-author Kenneth Gillingham, who has worked for the California Air Resources Board and was on the Council of Economic Advisers under Presidents Obama and George W. Bush, said that the best evidence doesn’t support such an increase.

There are studies that identify 20 percent or larger rebound effects, he said, but those studies tend to be decades older and survey-based, or from Europe, where gas prices are higher and public transportation systems are more robust.

Studies within the past decade in America using multiple odometer readings, which are more reliable than people’s memories, have turned up much smaller rebound effects. 

According to Gillingham’s review of the literature, which he shared with the agencies in response to the SAFE proposal, the average rebound estimate from these recent odometer studies is 8.1 percent, with some identifying no rebound effect at all. Many of these studies weren’t included in NHTSA and EPA’s review.

“They’re relying on effects from Europe, old estimates and a cherry-picked set of estimates from more recent times,” Gillingham said of the agencies in an interview. “It’s pretty hard if you look at the more recent work to argue for 20 percent.”

The change to the rebound effect doesn’t alter the cost-benefit analysis figures all that much, the Science paper says, but it’s a big part of how the agencies are able to arrive at their estimate of 12,700 additional deaths under the Obama standards. As we’ll explain in more detail below, that conclusion is also likely faulty.

A Sketchy Safety Argument

One of the key ways the Trump administration justifies the rollback is by claiming that it would save lives.

Lower car costs, the agencies argue, would allow more people to buy new, safer cars. Without the standards, automakers also wouldn’t tinker with vehicle weight, and there wouldn’t be the extra miles from people driving more fuel efficient vehicles.

The researchers we spoke with, however, weren’t convinced by this line of thinking.

Michael Anderson, a health and environmental economist at the University of California, Berkeley who was not involved with the Science paper, said that while faster fleet turnover might help somewhat with safety, many factors, such as driver age and wealth, correlate with newer vehicles, so it’s “hard to know exactly how much that matters.”

The entire premise of faster turnover is also predicated on the idea that under the Obama standards, cars will cost much more and new vehicle sales will fall. But Gillingham uncovered a simple math error when the agencies failed to convert a quarterly figure into an annual one. When corrected, the estimated number of lost sales — or new cars that would be sold under the rollback, but not under the Obama rules — drops by 70 percent.

“That right off the bat cuts the number of fatalities,” said Gillingham, who added that fixing the single error would reduce the 12,700 deaths by a third. “It’s huge. That is a really important fact that they did not get right.”

The Trump administration also predicts that automakers would downsize vehicles to meet the standards, and says those weight adjustments would have a negative effect on safety (see page 43106).

But Anderson, who has studied the effect of vehicle weight on traffic fatalities, told us that the relationship between safety and car mass is complex. While adding 100 pounds to one car makes that car safer, it makes other collisions less safe. The best thing for safety is to have vehicles that are the same size. But it’s not obvious, he said, what the standards would do to vehicle size.

Finally, the administration’s tables (see tables 11-71, 11-73, 11-75 and 11-77) show that about half or more of the lives come from the rebound effect — which, as we explained, was doubled in the Trump analysis, and refers to when people drive more often if their vehicles are more fuel efficient. A small number of lives saved are from weight changes.

“The preponderance of evidence would suggest,” Anderson said of the standards, “that whatever the safety changes are — they might be negative, they might be positive — they’re not going to be very large in magnitude.”

The 12,700 tally, he said, is “not based on the best evidence or data that we have on this subject.”

Internal documents reveal EPA staff also doubted the high fatality estimates before the publication of the SAFE proposal in August 2018.

In June 2018, William Charmley, an official with the EPA’s Office of Transportation and Air Quality, shared his concerns about NHTSA’s modeling in a memo that became public in a government docket.

In it, he specifically called out the scrappage model for inappropriately inflating the number of vehicle miles driven and thereby increasing the number of fatalities, at one point stating that the result is “clearly wrong” and “driving incorrect fatality estimates.” The memo also documents code revision work EPA staff did in an attempt to fix some of the obvious errors, which found that a rollback would increase the fatality rate and lead to 17 more deaths each year between 2036 and 2045, not including rebound effects.

“The Proposed standards are detrimental to safety, rather than beneficial,” the memo reads.

What Happens Next?

The Trump administration hasn’t yet finalized its fuel economy rollback. Initially, the finalized rule was expected at the end of March, but Andrew Wheeler, the head of the EPA, has since stated that the rule would hopefully come out later in the spring or early this summer.

The New York Times reported in April, according to anonymous sources familiar with the administration’s plan, that rather than completely eliminating the previous administration’s 5 percent increases in CO2 emissions stringency, the agency might opt for a 1 percent improvement each year. Wheeler said in an April 11 interview with Reuters that the final rule “is not going to be the same” as the proposed rule, but wouldn’t elaborate any further because EPA was “still reviewing the data and information.”

What is also unclear is whether the agencies will fix the flaws identified by outside scientists and by members of its own teams. On April 2, Wheeler said that career staff had been consulted on the proposed rule and would continue to be for the finalized version (see 27 minutes into video). In the Reuters interview, he referred to the feedback the agency had received, saying, “We have taken constructive comments, criticisms and concerns from a whole host of different interest groups, and I hope that our final regulation is something that everyone can get behind and support.”

The other big question is what will happen with California. Government officials announced at the end of February that talks with the California Air Resources Board had fallen apart and the agencies would proceed on the rulemaking without the state’s approval.

That result makes some automakers nervous because while many had pushed Trump to weaken the standards, they still wanted buy-in from California so that there would be a single standard across the U.S. Around a dozen other states and a third of the American market follow the Golden State’s standards. EPA argues that it is providing a single federal standard by removing California’s ability to set a higher one. But California is expected to sue in response, sending the issue to the courts and into a protracted legal limbo — a result that fails to provide regulatory certainty.

California has already announced that it is suing EPA and NHTSA for failing to provide additional information explaining how the agencies arrived at their decision to overturn the Obama rules. According to the state, that information had been requested under the Freedom of Information Act in September.