CR finds DOT and EPA made questionable choices in trying to justify the rollback of federal fuel economy and vehicle emissions standards.

Recently, the Department of Transportation (DOT) and the Environmental Protection Agency (EPA) finalized a plan to weaken federal gas mileage requirements, even though the new rules acknowledge that they will lead to higher fuel costs, increased climate pollution, and fewer auto industry jobs.

Consumer Reports has reviewed thousands of pages of documents DOT and EPA published in support of the rollback, known as the “SAFE Rule,” and found that DOT and EPA made several irregular, and at times, even unexplainable, choices in its calculations, with the effect of making the already bad ‘SAFE Rule’ look just a little bit less worse than it really is. Even with these fundamental flaws, the ‘SAFE Rule’ projects overall cost increases for consumers, drastic increases in greenhouse gas emissions, and the loss of thousands of auto industry jobs.

Below, CR describes the dubious decisions, and then estimates how much money DOT and EPA were off by in their calculations.

  1. Technology Markup. DOT and EPA inflated all technology costs (cost to the manufacturer by a factor of 1.5. They said they considered another methodology — “indirect cost multipliers,’ which are based upon automotive industry experts’ knowledge on the actual costs to the industry to install technology — but then chose not to use it. Failing to use the more sophisticated industry expert approach inflated technology costs by 22 percent.
    • Estimated increase in cost of rollback: $23B
    • Where to look: FRIA p 374
  2. Rebound Effect. DOT and EPA doubled the “Rebound Effect”.  When the original rule was put in place EPA and NHTSA’s analysis used a value of 10% for the rebound effect, but recent research shows that it may actually be even lower than 10%. This rollback used a 20% rebound effect. The rebound effect captures the effect on driving when consumers have more or less money due to changes in fuel economy. Under the rollback, people will have less money, so they won’t engage in as much economic activity, like going out to dinner or the movies, or shopping online. Some of these activities involve driving, so people will drive less, but nowhere near as much as the agencies assumed.
    • Estimated increase in cost of rollback: at least $22B (based on adjusting to 10%)
    • Where to look: Table I-74, FRIA p.48; p962-972
  3. Gas Tax Paid as a Benefit. The SAFE Rule calls an increase in gas tax revenue a negative “private cost,” i.e. a benefit of the rollback.  This is completely out of touch with consumers who would see paying more gas taxes a cost under the rule.
    • Estimated increase in cost of rollback: $32B
    • Where to look: Table I-74, FRIA p.48
  4. High Battery Costs. The analysis by DOT and EPA uses battery costs that start higher than current costs paid by the industry, and then applies a very low learning rate (assume battery costs decrease by 4.5% per year). In reality, battery costs for EVs have been dropping at rates that far exceed 10% per year and such rates of improvement are expected to continue. The analysis includes a sensitivity scenario with lower battery cost with higher learning rate (still only 6%). This scenario shows a $21B increase in the costs of the rollback. In reality the effect of the battery cost assumptions are probably much greater than this. Especially since the high battery costs are also subject to the 1.5 retail price markup.
    • Estimated increase in cost of rollback: at least $21B
    • Where to look: Table VII-482 p.1804, Low battery cost and faster learning scenario; Table VI-101, p595 of the preamble for comparison of battery costs.
  5. Sales Model Elasticity. The analysis increased the elasticity (how much sales change with a change in price) from -0.2 to -0.3 from the proposed rule to -1.0 in the final rule. This has the effect of quadrupling the sales effect, which affects things like job losses and safety impacts. DOT and EPA also claimed that consumers only value 2.5 years of fuel economy. When CR did its analysis assuming consumers value 6 years of fuel economy at a 7% discount rate and more reasonable technology costs we found that sales actually decreased under all rollback scenarios. The analysis includes a table citing 3 recent studies all showing that consumers value between 50%-100% of future fuel savings, undercutting their rationale for only including 2.5 years of fuel savings in their model.
    • Estimated increase in cost of rollback: Up to $16B or more (if sales effect negative)
    • Where to look: FRIA p861; Table VI-151, p760 of preamble for studies showing higher consumer willingness to pay for fuel economy. CR analysis p. 19
  6. Electric Vehicle (EV) Sales Constraint. The analysis puts a strict constraint on the number of EVs that the model can use to meet standards. This artificially constricts EV deployment by the model. However, DOT and EPA include a sensitivity study removing this constraint, which indicates that their own model finds EVs to be a cost-effective solution, but blocked it from use.
    • Estimated increase in cost of rollback: $13B
    • Where to look: Table VII-482 p.1804 of FRIA, Unconstrained BEV adoption sensitivity
  7. Light Truck Sales. The analysis uses a “dynamic fleet share model to estimate that light trucks will make up between 47%-44% of sales. However, in 2018, 52% of sales were light trucks, according to EPA. Since the consumer losses shown by their own analysis per light truck ($903) are greater than the losses per car ($577), a lower percentage of sales of light trucks lowered the apparent net losses of the new rule. All of this despite the fact that one of the key excuses used to justify the rollback is that consumers are buying more light trucks. The reality is that this rule will hurt buyers of crossovers, SUVs and pickups the most. 
    • Estimated increase in cost of rollback: Unquantified, but fixing would show higher cost due to rollback
    • Where to look: See Table I-5 p. 15 of preamble for costs per vehicle; For fleet makeup see Table VI-192 on p. 882 of FRIA
  8. Inconsistent Fatalities Math. There appear to be different numbers used to value fatalities due to crashes vs. fatalities due to air pollution. A value of $10.4m per life appears to be used for crashes and a number of $8.7m per life appears to be used for air pollution impacts, a difference of 20%.
    • Estimated increase in cost of rollback: Unquantified, but fixing would show higher cost due to rollback
    • Where to look: See p1183 of FRIA for crash VSL; See p 1639 for air pollution VSL; See table VII-374 p1636 for estimated pollution impacts.
  9. Other Inexplicable Numbers. In summary table I-5, the numbers don’t seem to add up. For example, the per-vehicle price increase and the lifetime fuel savings don’t add up to the per vehicle savings which are higher than they would be if a simple difference is taken. This may be due to a failure to use consistent time periods and discount rates, which make reductions in technology costs look larger, and makes lost fuel savings look not quite as bad. Also, the light truck and car “consumer per-vehicle savings” are both higher than the combined fleet number, which doesn’t seem possible. You can’t take a weighted average of two numbers and get a number that’s lower than both of the numbers that make up the average. 
    • Estimated increase in cost of rollback: Unquantified
    • Where to look: See Table I-5 p. 15 of preamble