What the NRC report on LDV technologies does and doesn't add to the debate in the run up to the US CAFE 2025 standard midterm evaluation
The National Research Council’s latest vehicle technology assessment, Cost, Effectiveness and Deployment of Fuel Economy Technologies for Light-Duty Vehicles, released last week, presents a massive amount of detailed technical information relevant to the mid-term evaluation of the U.S. 2025 passenger vehicle CAFE/GHG regulation. The synthesis report provides a good summary of how the use of full-vehicle simulation modeling in combination with lumped parameter modeling has improved the regulatory agencies’ estimation of fuel economy impacts, and how tear-down studies have improved cost estimates. The relationships between footprint, mass, and safety are clearly surveyed, along with the safety benefits of advanced materials, vehicle design, and crash-avoidance technology. The recommendation to gather data on real-world fuel economy to assess the gap between real-world fuel economy and certification values is something we’ve been finding elsewhere. The suggestion that vehicle tests use actual weight instead of weight classes is sound.
Sifting through the technical details, a few take-away points emerge.
The CAFE/GHG standards get a vote of confidence. Press reports and other comments (for example, here and here and here) have noted that the report supports a conclusion that automakers will be able to meet the 2025 fuel economy targets more easily and more cheaply than some were beginning to claim—without resorting to a slew of expensive electric vehicles, but through development of “conventional” technologies and designs that the regulation itself is driving.
More efficiency technology is on the way. As the NRC points out, “There are also new technologies not considered by EPA and NHTSA that might provide additional fuel consumption reductions for SI engines, or provide alternative approaches by 2025 and beyond. These technologies include higher compression ratio, exhaust scavenging, lean burn, and electrically assisted supercharger approaches.” In other words: the NRC is indicating that the story will only get better with all the new technologies under development. The ability of manufacturers to exceed projections is illustrated by the graph below of fuel economy improvements in some of the most popular mid-size sedans from 2010 to 2014. The fuel economy is plotted against the footprint-based targets, so that changes in size are also reflected. Despite the fact that all of the 2010 4-cylinder cars substantially exceeded their 2010 targets, redesigns that occurred through 2014 exceeded their 2016 targets by 2 to 8 mpg. And this is just showing model year 2014.
This is useful and encouraging, as far as it goes. The point I’d like to make sure we don’t lose sight of over the next couple of years is that the report should have and could have pushed its conclusions further. It takes a conservative approach on important assumptions about costs and technology innovation. That’s understandable, and maybe unavoidable given the NRC’s mandate. But there are good empirical reasons to anticipate that we’ll look back on this report and realize that it was too conservative — just as earlier NRC vehicle technology assessments have been (and undoubtedly for the same reasons). Here’s what I mean.
The report’s new conservative set of technology costs aren’t likely. The committee appeared to be of two minds, with two sets of “most likely” cost estimates. The “most likely high” estimate is, in essence, an inventory of what might go wrong. These high cost estimates routinely reflect only factors that would increase cost compared to NHTSA’s estimates in its 2017–2025 light-duty vehicle CAFE rulemaking. While it is legitimate to estimate the high end of the range of costs, this is not a “most likely” estimate and should not be labeled as such. Thus, while the report contains two sets of cost estimates, only the “most likely low” are actually estimates of what costs are likely to be. And these are not “low” estimates, as in general the report did not consider reasons why costs might be lower in the future than the agency’s cost estimates. Instead, if we look to past experience, we should expect, continuous vehicle technology improvements to cut costs; vehicle regulations tend to OVER-estimate rather than underestimate costs (e.g., see the NRC 2006 report). It is also inconsistent with the committee’s own finding in comparing 2014 MY vehicles to the 2008 MY baseline (p. 8-33), which states “that the actual fuel consumption reductions based on EPA certification test data meet, and in some cases exceed, the aggregation of NHTSA technology effectiveness estimates.”
The primary problem is that the report doesn’t consider how future innovation influences costs. (The report did include annual cost reductions due to “learning,” but this is different from innovation. Learning reflects well-established trends in cost reduction of existing technology. Innovation refers to technology breakthroughs that go beyond just taking cost out of an existing technology.)
Here’s an illustration of how regulatory cost estimates are almost always much higher than final costs: the 2002 NAS study “Effectiveness and Impact of Corporate Average Fuel Economy (CAFE) Standards.” This detailed, comprehensive, well-regarded, widely cited study of vehicle technology was used by NHTSA as the analytical basis of its light-truck CAFE standards in the 2000s. But the decade-plus since its publication has seen a number of developments that the report completely failed to anticipate and that enabled dramatic improvements in vehicle efficiency and cost. These include such now commonplace vehicle technologies and designs as dual-clutch transmissions, higher-torque continuously variable transmissions, innovative automatic transmission designs with addition gears, the pairing of gasoline direct-injection with turbocharging, improvements in turbocharger materials, and low-cost high-strength steel. In fact, every technology projection in the 2002 NRC report has been either matched or exceeded.
Lightweighting costs are a topic to keep a particular eye on. The NRC report does recognize recent developments in lightweight materials and design, and forecast higher rates of lightweight material use than were included by the agencies in the 2017–2025 rule. But it also states that manufacturers will reduce light-truck mass by 20% despite very high cost ($1,617–$2,343 for a 5,550 pound truck) because “implementation of mass reduction techniques can provide several benefits that might be attractive to an OEM” (p. 6-10). Not only are these cost estimates far too high for reasons that would take too long to explain here (but see here for several rigorous studies that indicate lower lightweighting cost), but these other benefits have substantial value to customers, such as better performance, ride, handling, and braking and higher towing and payload capacity —that is, it’s incorrect to assign all mass-reduction costs to fuel consumption when mass is being reduced primarily for other reasons.
This NRC report is an early input into a years-long process. There will be many dozens of additional rigorous studies completed over the next two years. There will also be dozens of more product launches that deploy new powertrain technology, utilize more lightweighting, and achieve even greater fuel economy. The agencies’ Technical Assessment Report will be done in 2016. Additional analyses and studies will be conducted after this for the Notice of Proposed Rulemaking in 2017, and then more analyses and studies will be done to respond to comments on the proposed rule before the Final Rule is published in 2018. EPA, NHTSA, and CARB have very impressive research plans, with many technical reports, including on technology benchmarking, new vehicle simulation modeling of advanced powertrains, additional cost teardown analysis, and research on consumer and economic issues.
Because it arrives so early — and it’s the National Research Council — the NRC report may get some extra attention. So, while I don’t intend my comments to be unduly critical (it’s a good report and a constructive contribution to the evidence base for the 2025 CAFE standard mid-term evaluation), I do think it’s important to note right away that there’s a good likelihood, based on both the internal evidence of the report and on experience, that its assessments of costs and innovations will turn out to have been too cautious. All the more reason, then, to be encouraged that it already supports the agencies’ initial analysis so convincingly.