![[logo] US EPA](http://www.epa.gov/epafiles/images/logo_epaseal.gif){kind=logo}
Average Fuel Economy Standards Passenger Cars and Light Trucks Model Year 2011
Note: EPA no longer updates this information, but it may be useful as a reference or resource.
PDF Version (262 pp, 6731K, About PDF) [Federal Register: March 30, 2009 (Volume 74, Number 59)] [Rules and Regulations] [Page 14395-14444] From the Federal Register Online via GPO Access [wais.access.gpo.gov] [DOCID:fr30mr09-16] Average Fuel Economy Standards Passenger Cars and Light Trucks Model Year 2011 [[Continued from page 14394]] [[Page 14395]] health effects of reductions in population exposure to criteria air pollutants and MSATs that were projected to result from alternative CAFE standards for MY 2011-15. That analysis suggested that significant reductions in adverse health effects and economic damages caused by exposure to these pollutants (primarily PM2.5, the largest known contributor to adverse health effects) could result if higher CAFE standards were adopted for those model years. (See Sec. 3.3.2.4.2 of the FEIS for a detailed description of NHTSA's approach for developing the quantitative estimates of changes in health effects from exposure to air pollution resulting from alternative CAFE standards for MY 2011-15.) E. Picking the Final Standards 1. Eliminating the Alternatives Facially Inconsistent With EPCA (a) No-Action Alternative Two of the alternatives analyzed by NHTSA are facially inconsistent with EPCA. Regardless of how this alternative is defined, i.e., either in terms of setting no standard or setting the MY 2011 standards at the MY 2010 level, the ``no-action'' or ``baseline'' alternative violates EPCA. Under the former definition, the no-action alternative violates, among other EPCA provisions, subsections 32902(a) and (b)(1) and (2), each of which requires the Secretary to establish CAFE standards for each model year separately. Under the latter definition, the no-action alternative violates subsection 32902(b)(2)(A) which requires the MY 2011-2020 standards to be set high enough to ensure that the industry- wide fleet achieves a combined passenger car/light truck average fuel economy of at least 35 mpg. It also violates the requirement in subsection 32902(b)(2)(B) that the standards for MYs 2011-2020 increase annually and ratably. (b) Technology Exhaustion Alternative Although the technology exhaustion alternative is the environmentally preferable alternative for NEPA purposes, it does not reflect any consideration of economic practicability or technological feasibility. This omission violates subsections 32902(a) and (b), which require setting standards at the maximum feasible level, and subsection 32902(f), which requires that ``(w)hen deciding maximum feasible average fuel economy under this section, the Secretary of Transportation shall consider technological feasibility, economic practicability, the effect of other motor vehicle standards of the Government on fuel economy, and the need of the United States to conserve energy.'' (Emphasis added.) 2. Choosing Among the Remaining Alternatives (a) Difficulty and importance of Achieving a Reasonable Balancing of the Factors Section 1(a) of E.O. 12866 provides that ``(i)n choosing among alternative regulatory approaches, agencies should select those approaches that maximize net benefits (including potential economic, environmental, public health and safety, and other advantages; distributive impacts; and equity), unless a statute requires another regulatory approach.'' The Ninth Circuit ruled in CBD v. NHTSA, 538 F.3d 1172, 1197, that EPCA does not require another regulatory approach. We recognize that the Ninth Circuit coupled that ruling with the following cautionary statement about basing decisions about the stringency of CAFE standards on the principle of maximizing net benefits: (W)e reject only Petitioners' contention that EPCA prohibits NHTSA's use of marginal cost-benefit analysis to set CAFE standards. Whatever method it uses, NHTSA cannot set fuel economy standards that are contrary to Congress's purpose in enacting the EPCA-energy conservation. We must still review whether NHTSA's balancing of the statutory factors is arbitrary and capricious. Additionally, the persuasiveness of the analysis in Public Citizen and Center for Auto Safety is limited by the fact that they were decided two decades ago, when scientific knowledge of climate change and its causes were not as advanced as they are today. * * * The need of the nation to conserve energy is even more pressing today than it was at the time of EPCA's enactment. * * * What was a reasonable balancing of competing statutory priorities twenty years ago may not be a reasonable balancing of those priorities today. (footnotes omitted) 538 F.3d 1172, 1197-98. As discussed below, achieving a reasonable balancing of the factors is critical. While, as the Court suggested, there are risks associated with setting standards that are too low, there are also considerable risks associated with setting standards that are too high. Both types of risks must be part of the balancing process. We recognize that the on-road fleet of passenger cars and light trucks is one of largest consumers of petroleum and emitters of CO2 in the U.S. economy. We recognize too that global CO2 emissions have been exceeding the highest of the IPCC 2007 scenarios. We appreciate that, among the remaining alternatives, the total cost/total benefit alternative is the one that reduces those emissions the most. At the same time, we cannot fail to recognize and fully take into account the very serious conditions of the automobile industry, the national economy, and even the global economy. We understand that some aid has been authorized and appropriated for the automobile industry and that the possibility of other aid has been broached, but the extent to which that aid will mitigate the industry's downward spiral is uncertain. What is certain is that the mere fact substantial aid is even being discussed is a reflection of the unusual and extremely serious conditions we face. (b) The Correct Balancing of the Factors for Setting the MY 2011 Standards Is To Maximize Societal Net Benefits We have discussed above how NHTSA considered and balanced the four statutory factors. This section discusses NHTSA's decision that the final standards are the maximum feasible for MY 2011. Congress left the determination of what levels of CAFE standards are ``maximum feasible'' to NHTSA's discretion, requiring only that NHTSA consider the four statutory factors. 49 U.S.C. 32902. NEPA applies independently to require consideration of environmental factors in the decision-making process. The EPCA factors are in tension and tend to pull in opposite directions in terms of stringency, with technological feasibility and especially the need of the nation to conserve energy pointing toward higher standards and economic practicability pointing toward lower ones. Accordingly, NHTSA has historically considered the factors from the perspective of balancing them, given EPCA's overarching purpose of energy conservation.\440\ Thus, NHTSA determines that standards are the maximum feasible if they represent the proper balancing of the four statutory factors, based on all the information before the agency and the entire record. --------------------------------------------------------------------------- \440\ The Ninth Circuit in CBD agreed that NHTSA has discretion to balance the factors in determining what level of stringency is maximum feasible. CBD, 538 F.3d 1172, 1197 (9th Cir. 2008). --------------------------------------------------------------------------- The ``need of the United States to conserve energy'' primarily functions to encourage NHTSA to set standards ever higher. Many commenters cast the need of the nation to conserve energy in terms of the impact of CAFE standards on global warming, and urged NHTSA to give this factor more weight than the others in its determination of the maximum feasible standards, in order to [[Page 14396]] have the maximum possible beneficial impact. Many of these commenters suggested that if NHTSA gave more weight to the need of the nation to conserve energy, it would set standards at levels substantially higher, for example, than those necessary to raise the industry-wide combined average to 35 mpg by MY 2015, or at the level at which total costs equal total benefits, and so forth. NHTSA recognizes that seriousness of the global warming problem facing the nation and the world today, and that CAFE is one of many actions needed around the world to address that problem. NHTSA also recognizes that the higher CAFE standards are, the less they add to global warming and other environmental impacts (as demonstrated in our FEIS), just as the higher CAFE standards are, the less oil the United States must purchase from abroad, with the corresponding impacts on consumer costs, national balance of payments, and foreign policy objectives. The final standards for MY 2011 push CAFE higher and faster than any set of standards since the earliest years of the program, and, we believe, likely put the agency on track to meet EISA's MY 2020 requirement of an industry-wide combined average of at least 35 mpg several years ahead of time. However, NHTSA reiterates that it is required to consider and balance the other three factors in addition to the need of the nation to conserve energy in determining the maximum feasible level of the standards. While considering the need of the nation to conserve energy alone might counsel for setting the standards at the levels suggested by proponents of higher standards, NHTSA does not believe that those standards would be consistent with economic practicability or technological feasibility. Manufacturers commented that even standards set at the proposed levels would be above the maximum feasible level because, in their view, NHTSA had overestimated benefits and underestimated costs of the fuel-saving technologies. Conversely, many other commenters argued that the proposed standards were below the maximum feasible level because, in their view, NHTSA had underestimated benefits and overestimated costs of the technologies. To respond to these commenters, and aid in resolving their conflicting views and arguments, NHTSA re-examined all of its technology assumptions, with the assistance of Ricardo, as described in Chapter IV. This effort resulted in the agency's revising the methodology underlying the development of many of its technology assumptions in ways that the agency believes makes its final rule analysis substantially more robust than its NPRM analysis. NHTSA is confident that its revised analysis ensures that the standards adopted in this final rule are technologically feasible. The effect of other motor vehicle standards of the Government on fuel economy is incorporated into the agency's analysis through the baseline and the manufacturers' product plans. Yet the question of economic practicability and what level of stringency would cause manufacturers substantial economic hardship must be considered not only in terms of technological feasibility, but also in terms of the economic situation today and as it is anticipated to be in the period leading up to and including MY 2011. The current economic realities are markedly different from those at the time of the NPRM; just several months later, the national and global economies are in crisis and by all accounts in recession. As the economy contracts and consumers reassess their personal spending priorities, manufacturers are increasingly less able to pass the costs of fuel economy-improving technologies on to consumers. As discussed above in the section on economic practicability, manufacturers have only so much ability to absorb those costs, especially given the financial difficulties of some of the larger manufacturers. NHTSA additionally notes that the agency has the authority under 49 U.S.C. Sec. 32902(c) to amend the standards for a model year to a level that the Secretary decides is the maximum feasible average fuel economy level for that model year. NHTSA has previously used this authority to lower the MY 1986 passenger car standards because they were deemed to be beyond maximum feasible. However, NHTSA believes that the authority to lower CAFE standards in MYs 2011-2020 has been constricted by the EISA requirements that standards increase annually and ratably and result in a combined fleetwide average fuel economy of at least 35 mpg in MY 2020. Thus, being unable to predict the economic situation in MY 2011, NHTSA is particularly mindful of economic practicability in establishing the current standards. For this MY 2011 final rule, in balancing the EPCA factors against one another and carefully considering the environmental impacts associated with the various alternatives evaluated, NHTSA continues to believe that the proper overall balance of all relevant consideration is the point at which social net benefits are maximized, and results in CAFE standards that are the maximum feasible for MY 2011. As mentioned above, in identifying this point for this model year, NHTSA evaluated more than 100 alternative stringency levels, and for each one, calculated net benefits in a manner that explicitly accounted for the need of the nation to conserve energy, and for the benefits of reducing greenhouse gas emissions. EPCA's overarching purpose of energy conservation is met by setting standards at the maximum feasible level--EPCA does not require or even permit that standards be set beyond the maximum feasible level in order to achieve more energy conservation. NEPA's purpose is to integrate environmental considerations into that decision-making process. Setting standards at the point at which social net benefits are maximized in NHTSA's analysis results in standards that still increase higher and faster than any standards since the earliest years of the program, do not require the addition of technologies that the agency does not believe will pay for themselves, and result in measurable environmental benefits. The standards thus fulfill NEPA's objectives and, under EPCA, the need of the nation to conserve energy, while not imposing substantial economic hardship on the industry, while taking into account the feasibility of applying technologies appropriately and consistent with manufacturers' natural cycles, and the other motor vehicle standards of the government which manufacturers have to comply with. NHTSA is exercising its discretion and informed judgment, based upon the entire record and including the FEIS, as to the precise levels of CAFE that are the maximum feasible for MY 2011 passenger cars and light trucks, as mandated by 49 U.S.C. 32902. NHTSA emphasizes that it will continue to evaluate alternative approaches for determining the maximum feasible standards for future CAFE rulemakings, and is deciding no more than that the approach taken for MY 2011 is reasonable under the circumstances surrounding this rulemaking. VIII. Safety A. Summary of NHTSA's Approach in This Final Rule NHTSA has devoted substantial efforts over the years studying the relationship between vehicle weight reductions and vehicle injuries and deaths based upon a broad base of available empirical data. More recently, NHTSA addressed these issues in a 1997 study, which was reviewed by the National Academy in its 2002 report. [[Page 14397]] This 1997 study, led by Dr. Charles Kahane of NHTSA, ``stands alone as a comprehensive, scientific analysis of the vehicle weight and safety issue.'' \441\ --------------------------------------------------------------------------- \441\ Effectiveness and Impact of Corporate Average Fuel Economy (CAFE) Standards (NRC, 2002), at 118. --------------------------------------------------------------------------- Thereafter, in a 2003 study, again led by Dr. Kahane, NHTSA analyzed historical fatality rates in crashes involving MY 1991-1999 vehicles, both passenger cars and light trucks. NHTSA's 2003 study built upon and updated the earlier 1997 study analyzed by the National Academy. Among other things, the 2003 study concluded that there is a ``crossover weight,'' a statistically derived weight above which vehicle weight reductions have a net benefit, instead of a net harm, in terms of reduced vehicle injuries and deaths to society. The 2003 study found that this crossover point occurs somewhere in the range of 4,224 pounds to 6,121 pounds. The 2003 study concluded that the most likely location of the crossover point is 5,085 pounds. Based upon the findings of the 2003 study, in setting fuel economy levels in this final rule, NHTSA did not assume that manufacturers would reduce vehicle weight to improve fuel economy for vehicles of 5,000 pounds or less. NHTSA has taken this approach so that manufacturers are not encouraged to downsize vehicles in a way that would be likely to cause a significant number of deaths and injuries. Conversely, NHTSA has considered reduced vehicle weight in its standard-setting analysis for vehicles above 5,000 pounds, since the data indicates no safety penalty is likely for reducing weight for such vehicles. Nevertheless, the agency will continue to consider whether it should set future CAFE standards in a manner that assumes manufacturers may, without compromising highway safety, reduce the mass of vehicles below 5,000 pounds. B. Background As the courts have recognized, ``NHTSA has always examined the safety consequences of the CAFE standards in its overall consideration of relevant factors since its earliest rulemaking under the CAFE program.'' Competitive Enterprise Institute v. NHTSA, 901 F.2d 107, 120 n. 11 (D.C. Cir. 1990) (``CEI I'') (citing 42 FR 33534, 33551 (June 30, 1977)). The courts have consistently upheld NHTSA's implementation of EPCA in this manner. See, e.g., Competitive Enterprise Institute v. NHTSA, 956 F.2d 321, 322 (D.C. Cir. 1992) (``CEI II'') (in determining the maximum feasible fuel economy standard, ``NHTSA has always taken passenger safety into account.'') (citing CEI I, 901 F.2d at 120 n. 11); Competitive Enterprise Institute v. NHTSA, 45 F.3d 481, 482-83 (D.C. Cir. 1995) (``CEI III'') (same); Center for Biological Diversity v. NHTSA, 538 F.3d 1172, 1203-04 (9th Cir. 2008) (upholding NHTSA's analysis of vehicle safety issues associated with weight in connection with the MY 2008-11 light truck CAFE rule). As early as 1974, before Congress even enacted EPCA, the Department of Transportation and EPA warned Congress of potential adverse safety effects associated with increasing fuel economy requirements for vehicles. See CEI I, 901 F.2d at 120 n. 11 (citing 53 FR 39275, 39294 (1988), in turn citing a report from the Department of Transportation and EPA, ``Potential for Motor Vehicle Fuel Economy Improvements: Report to the Congress,'' (Oct. 24, 1974), which discussed ``the possible trade offs in the areas of improved fuel economy, lower emissions, and increased occupant safety,'' noting that ``a sustained or increased shift to small cars * * * would likely lead to an increase in the rate of highway deaths and serious injuries''). The relationship of vehicle weight to safety has been a contentious issue for many years. This contentiousness arises, at least in part, from the difficulty of isolating vehicle weight from other confounding factors (e.g., driver factors, such as age and gender, other vehicle factors, such as engine size and wheelbase, and environmental factors, such as rural/urban). In addition, several vehicle factors are closely related, such as vehicle mass, wheelbase, track width, and structural integrity. (Historically, as vehicles got longer and wider, they also got heavier). The papers that were initially published addressing vehicle size and safety did not attempt to fully address all of these factors. 1. NHTSA's Early Studies It was important for NHTSA to help move the debate forward with more serious analyses. After all, NHTSA must understand the relationship between vehicle factors and safety, both for establishing our safety standards and for establishing our CAFE standards. In July 1991, NHTSA published a study of the effects of passenger car downsizing during 1970-1982 titled Effect of Car Size on Fatality and Injury Risk. In this report, NHTSA concluded that changes in the size and weight composition of the new car fleet from 1970 to 1982 resulted in increases of nearly 2,000 deaths and 20,000 serious injuries per year over the number of deaths and serious injuries that would have occurred absent this downsizing. Parties reviewing NHTSA's 1991 report identified a number of areas that could be improved. Suggestions included extending the analyses to include light trucks and vans, examining finer gradations to distinguish the relative impacts of weight reduction for the heavier cars from the lighter cars, analyzing all crash modes, and doing more to isolate the effects of vehicle mass from behavioral and environmental variables. NHTSA agreed that these suggestions would make the study more useful as a tool for NHTSA decisions on safety and fuel economy standards. Accordingly, NHTSA developed a more comprehensive analytic model to encompass all light vehicles, and to allow a finer look at safety impacts in different segments of the light vehicle population. This study was NHTSA's first effort to estimate the effect of a 100- pound weight reduction in each of the important crash modes, and to do this separately for cars and light trucks. NHTSA recognized that the findings, whatever they were, would likely be controversial, so the agency chose to have the draft report peer-reviewed by the National Academy of Sciences before publishing the document. The Academy published its review on June 12, 1996.\442\ The report expressed concerns about the methods used in the analyses and concluded, in part, ``the Committee finds itself unable to endorse the qualitative conclusions in the reports about projected highway fatalities and injuries because of large uncertainties associated with the results * * *.'' These reservations were principally concerned with the question of whether the NHTSA analyses had adequately controlled for confounding factors, such as driver age, gender, and aggressiveness. --------------------------------------------------------------------------- \442\ Transportation Research Board, Letter Report--Committee to Review Federal Estimates of the Relationship of Vehicle Weight to Fatality and Injury Risk, Accession Number 00723787. See http://onlinepubs.trb.org/onlinepubs/reports/letrept.html(last accessed Nov. 11, 2008). --------------------------------------------------------------------------- NHTSA responded at length to the committee report, and revised its report to address the committee recommendations. The revised report was published as a finished document in 1997,\443\ with a new Appendix F titled ``Summary and Response to TRB's Recommendations on the Draft Report.'' --------------------------------------------------------------------------- \443\ Kahane, C. J., 1997. Relationships Between Vehicle Size and Fatality Risk in Model Year 1985-93 Passenger Cars and Light Trucks, NHTSA Technical Report, DOT HS 808 570. Springfield, VA: National Technical Information Services. --------------------------------------------------------------------------- [[Page 14398]] In this 1997 report, NHTSA concluded that, calibrated from 1985-93 cars and light trucks involved in crashes in calendar years 1989-1993, there was little overall effect for a 100-pound weight reduction in light trucks and vans, because increased fatalities of truck occupants were offset by a reduction of fatalities in the vehicles that collided with the lighter trucks, whereas a 100-pound reduction in cars was associated with an increase of about 300 fatalities per year. Based on this analysis and subsequent activities, the safety consequences of weight reduction have been considered by NHTSA in deciding upon the appropriate stringency of each of the new safety and fuel economy requirements since that time. NHTSA's 1997 report did not end the public discussion of this issue. NHTSA followed its standard practice of publishing a notice announcing the report and inviting public comment on the 1997 report.\444\ In addition to comments to NHTSA's docket, other papers analyzing the relationship of vehicle weight and safety were published. For instance, Dr. David L. Greene of the U.S. Department of Energy's Oak Ridge National Laboratory published a report titled Why CAFE Worked soon after NHTSA's 1997 report was released.\445\ In section 5.2 of this report, Dr. Greene's introductory paragraph reads as follows: --------------------------------------------------------------------------- \444\ See 62 FR 34491 (June 26, 1997). \445\ Dr. Greene's report is available online at http://www.osti.gov/bridge/servlets/purl/625225-KPQDOu/webviewable/625225.pdf (last accessed October 28, 2008). Vehicle weight significantly affects the safety of the vehicle's occupants. Enough credible work has been done on this subject that this assertion cannot be seriously questioned (citations omitted). On the other hand, the nature of the trade-off between vehicle mass and safety is often misunderstood, and the implications for fuel economy regulations are generally misinterpreted. The relationship between fuel economy, mass, and public safety is complex, yet it is probably reasonable to conclude that reducing vehicle mass to improve fuel economy will require some trade-off with safety. The rational person will realize that individuals, manufacturers, and governments are constantly making trade-offs between safety and cost, safety and other vehicle attributes, safety and convenience, etc. (citation omitted). An essential feature of a rational economic consumer is the willingness to trade-off risk for money and, since --------------------------------------------------------------------------- fuel economy saves money, to trade-off safety for fuel economy. David L. Greene, 1997, Why CAFE Worked, ORNL/CP-94482, Oak Ridge National Laboratory, Oak Ridge, Tennessee, at 22 (Emphases added). It is noteworthy that Dr. Greene's published work explicitly acknowledges the vehicle weight-safety trade-off documented by NHTSA's studies of the real world crash data. As to Dr. Greene's concerns that the trade-off will be misunderstood, NHTSA has been clear on this point. NHTSA wants to ensure that the public, manufacturers, and governments are aware of the empirical data that demonstrate that there is a trade-off between vehicle mass and safety. Parties must understand this trade-off exists and the size of the trade-off should be quantified as accurately as possible, so it can be considered as part of the decision on average fuel economy standards. 2. The 2002 National Academy of Sciences Study The next significant event in the vehicle weight and safety discussion began in October 2000, when the Department of Transportation's Appropriations Act for fiscal year 2001 was signed into law. That appropriations law included a provision directing DOT to fund a National Academy of Sciences (NAS) study on the effectiveness and impacts of CAFE standards. NAS released its final study in January 2002 (hereafter, the 2002 NAS Report).\446\ --------------------------------------------------------------------------- \446\ Effectiveness and Impact of Corporate Average Fuel Economy (CAFE) Standards (NRC, 2002). --------------------------------------------------------------------------- As part of a comprehensive look at the impacts of CAFE standards, it was necessary for the 2002 NAS Report to address the safety impacts of CAFE standards. In Chapter 2 of the study, NAS looked back at the safety impacts of past CAFE standards. Among other observations, NAS recognized that much of the increase in fuel economy between 1975 and 1988 was due to reductions in the size and weight of vehicles, which led to increased safety risks.\447\ In fact, NAS noted that ``the preponderance of evidence indicates that this downsizing of the vehicle fleet resulted in a hidden safety cost, namely travel safety would have improved even more had vehicles not been downsized.'' \448\ --------------------------------------------------------------------------- \447\ Id., at 24. \448\ Id., at 69-70. --------------------------------------------------------------------------- The committee then focused its analysis on the 1997 NHTSA analysis led by Dr. Kahane. Since there are many published papers on this subject in the literature, the question must be asked, ``Why did the National Academy of Sciences choose the NHTSA analyses out of all the published papers?'' The NAS committee clearly and unequivocally answered this in its report, where it found that ``NHTSA's fatality analyses are still the most complete available in that they accounted for all crash types in which vehicles might be involved, for all involved road users, and for changes in crash likelihood as well as crashworthiness.'' \449\ The NAS committee went on to find that ``The April 1997 NHTSA analyses allow the committee to reestimate the approximate effect of downsizing the fleet between the mid-1970s and 1993.'' In other words, a committee of the National Academy of Sciences found that NHTSA's analyses were the most thorough of all the published papers, and that NHTSA's analyses were sufficiently persuasive and rigorous to permit a reasonable estimate of the safety penalty associated with downsizing the fleet. In the committee's words: --------------------------------------------------------------------------- \449\ Id., at 27. Thus, the majority of this committee believes that the evidence is clear that past downweighting and downsizing of the light-duty vehicle fleet, while resulting in significant fuel savings, has also resulted in a safety penalty. In 1993, it would appear that the safety penalty included between 1,300 and 2,600 motor vehicle crash deaths that would not have occurred had vehicles been as large and heavy as in 1976.\450\ --------------------------------------------------------------------------- \450\ Id., at 28. While this look back is informative, the greater challenge is to use this understanding of the past to guide future actions. Again the NAS --------------------------------------------------------------------------- committee offered clear guidance in this regard. The NAS Report said: In summary, the majority of the committee finds that the downsizing and weight reduction that occurred in the late 1970s and early 1980s most likely produced between 1,300 and 2,600 crash fatalities and between 13,000 and 26,000 serious injuries in 1993. The proportion of these casualties attributable to CAFE standards is uncertain. It is not clear that significant weight reduction can be achieved in the future without some downsizing, and similar downsizing would be expected to produce similar results. Even if weight reduction occurred without any downsizing, casualties would be expected to increase. Thus, any increase in CAFE as currently structured could produce additional road casualties, unless it is specifically targeted at the largest, heaviest light trucks. For fuel economy regulations not to have an adverse impact on safety, they must be implemented using more fuel-efficient technology. Current CAFE requirements are neutral with regard to whether fuel economy is improved by increasing efficiency or by decreasing vehicle weight. One way to reduce the adverse impact on safety would be to establish fuel economy requirements as a function of vehicle attributes, particularly vehicle weight (see Chapter 5). * * * [[Page 14399]] If an increase in fuel economy is effected by a system that encourages either downweighting or the production and sale of more small cars, some additional traffic fatalities would be expected. Without a thoughtful restructuring of the program, that would be the trade-off that must be made if CAFE standards are increased by any significant amount.\451\ --------------------------------------------------------------------------- \451\ Id., at 77. This discussion by the NAS committee was an impetus for NHTSA to use its existing statutory authority to reform its light truck CAFE program. This involved moving away from the single flat standard for light trucks, because those standards' neutrality with regard to decreasing vehicle weight, in lieu of increasing efficiency to improve fuel economy, means they necessarily have a potential safety trade-off. In place of the single flat standard, NHTSA established an attribute- based standard that is a function of the vehicle's footprint. Under this attribute-based standard, the fuel economy target for a vehicle increases as the vehicle is downsized. As long as vehicle manufacturers have to expend the same levels of advanced technology for each footprint size, there is no incentive to change the vehicle to get a less-demanding fuel economy target. Thus, the necessary safety trade- off under the single flat standard system does not arise under an attribute-based system. That is not to suggest there are no safety consequences if vehicle mass is reduced--there are, as documented by NHTSA and explained by the National Academy of Sciences. However, the standards are no longer structured to confer an advantage to a manufacturer that makes downsizing trade-offs. This is a key feature of the attribute-based fuel economy program NHTSA implemented for light trucks. Two of the 13 NAS committee members dissented on the safety issues.\452\ The dissent acknowledges that, ``Despite these limitations, Kahane's analysis is far and away the most comprehensive and thorough analysis'' of the safety issue.\453\ The dissent's primary disagreement with the other 11 committee members centers on the large uncertainties associated with NHTSA's analyses. The dissent acknowledges NHTSA's efforts in the study led by Dr. Kahane to quantify the safety penalty, but concludes that the number of factors in real world crashes is so large and the controls used by the analytical models introduce so much uncertainty that it is not possible to definitively make any statements about a safety penalty.\454\ --------------------------------------------------------------------------- \452\ One of the two dissenters was Dr. David Greene, the author of the 1997 report Why CAFE Worked, discussed supra. \453\ Effectiveness and Impact of Corporate Average Fuel Economy (CAFE) Standards, at 118. \454\ 2002 NAS Report, at Appendix A. --------------------------------------------------------------------------- It should also be noted that the majority of the committee responded to the dissent by saying: However, the committee does not agree that these concerns should prevent the use of NHTSA's careful analyses to provide some understanding of the likely effects of future improvements in fuel economy, if those improvements involve vehicle downsizing. The committee notes that many of the points raised in the dissent (for example, the dependence of the NHTSA results on specific estimates of age, sex, aggressive driving and urban vs. rural location) have been explicitly addressed in Kahane's response to the [NAS] review and were reflected in the final 1997 report. The estimated relationship between mass and safety were (sic) remarkably robust in response to changes in the estimated effects of these parameters. The committee also notes that the most recent NHTSA analyses yield results that are consistent with the agency's own prior estimates of the effect of vehicle downsizing (citations omitted) and with other studies of the likely effects of weight and size changes in the vehicle fleet (citation omitted). The consistency over time and methodology provides further evidence of the robustness of the adverse safety effects of vehicle size and weight reduction.\455\ --------------------------------------------------------------------------- \455\ Id., at 27-28. In addition, the NAS Committee unanimously agreed that NHTSA should undertake additional research on the subject of fuel economy and safety, ``including (but not limited to) a replication, using current field data, of its 1997 analysis of the relationship between vehicle size and fatality risk.'' \456\ NHTSA concurred with this recommendation, and thereafter, NHTSA undertook a replication of the 1997 study, using the additional field data that had become available: NHTSA's 2003 study, led again by Dr. Kahane. --------------------------------------------------------------------------- \456\ Id., at 6. --------------------------------------------------------------------------- As Congress was developing the bill that ultimately became EISA, Congress considered NHTSA's reformed light truck CAFE program established under existing NHTSA authority in deciding what additional CAFE authority NHTSA should be given and what constraints should be put on that authority. Ultimately, EISA was enacted, which mandates that NHTSA establish an attribute-based CAFE system for cars and light trucks. 3. NHTSA's Updated 2003 Study In October 2003, NHTSA published its updated study.\457\ NHTSA's update again used regression models to calibrate crash fatality rates per billion miles for model year 1991-1999 passenger cars, pickup trucks, SUVs, and vans during calendar years 1995-2000. These rates were calibrated separately by vehicle weight, vehicle type, driver age and gender, urban/rural and other vehicle, driver, and environmental factors. One major point of note is that, as the analyses get more sophisticated and able to differentiate the safety trade-off among different types of vehicles, each analysis NHTSA has ever conducted continues to show that there is a safety trade-off for the existing light vehicle fleet as vehicle mass is reduced. --------------------------------------------------------------------------- \457\ Charles J. Kahane, ``Vehicle Weight, Fatality Risk, and Crash Compatibility of Model Year 1991-99 Passenger Cars and Light Trucks,'' DOT HS 809 662, October 2003. This report is available online at http://www.nhtsa.dot.gov/cars/rules/regrev/evaluate/pdf/809662.pdf (last accessed Oct. 28, 2008). --------------------------------------------------------------------------- After controlling for vehicle, driver and environmental factors, the new study found that: • The association between vehicle weight and overall crash fatality rates in the heavier 1991-1999 light trucks and vans was not significant. Thus, there was no safety penalty for reducing weight in these vehicles. • In the other three groups of 1991-1999 vehicles--the lighter light trucks and vans, the heavier cars, and especially the lighter cars--fatality rates increased as weights decreased. [cir] Lighter light trucks and vans would have an increase of 234 fatalities per year per 100-pound weight reduction. [cir] Heavier cars would have an increase of 216 fatalities per year per 100-pound weight reduction. [cir] Lighter cars would have an increase of 597 fatalities per year per 100-pound weight reduction. • There is a crossover weight, above which crash fatality rates increase for heavier light trucks and vans, because the added harm for other road users from the additional weight exceeds any benefits for the occupants of the vehicles. This occurs in the interval of 4,224 pounds to 6,121 pounds, with the most likely single point being 5,085 pounds. The fatality rate changes by less than ±1 percent per 100-pound weight increase over this range. The draft report was reviewed before publication by experts in statistical analysis of crash data and related vehicle weight and safety issues: Drs. James H. Hedlund, Adrian K. Lund, and Donald W. Reinfurt. The review process is on record--the comments on the draft are available in Docket NHTSA-2003-16318-0004. Consistent with NHTSA's standard practice, NHTSA published its analysis and sought public comment on it.\458\ NHTSA then docketed a response [[Page 14400]] to the public comments on November 9, 2004.\459\ There were three principal criticisms of NHTSA's updated study, which are summarized below together with NHTSA's response. --------------------------------------------------------------------------- \458\ See 68 FR 66153 (Nov. 5, 2003). \459\ Docket No. NHTSA-2003-16318-0016. --------------------------------------------------------------------------- (1) The analyses only considered the relationship of vehicle mass to fatality risk. It did not consider other attributes of vehicle size, such as track width and wheelbase. Dynamic Research Inc. (DRI) presented analyses that included all three of these variables, and its analysis indicated that mass was harmful (i.e., reducing it would be positive for safety) while track width and wheelbase were beneficial. If true, this meant that weight reduction would benefit safety if track width and wheelbase were maintained. Agency response: The DRI results were strongly biased as a consequence of including 2-door cars in the analysis. Two-door muscle and sports cars stand apart from all other groups of cars by having a short wheelbase relative to their weight. They also have by far the highest fatality rates of all cars, for reasons mostly related to the drivers. The regression analysis immediately identifies short wheelbase with high weight as a disastrous combination. Being a regression, it tells you that you can make any car safer, including 4-door cars, by increasing wheelbase and/or reducing weight. This bias is amplified by treating highly correlated size attributes as independent factors in the model. To clarify this latter concern, NHTSA's analyses are calibrating the historical relationship of vehicle mass and fatality risk. In this type of analysis, ``vehicle mass'' incorporates not only the effects of vehicle mass per se, but also the effects of many other size attributes that are historically and/or causally related to mass, such as wheelbase, track width, and structural integrity. If historical relationships between mass and these other size attributes continue, future changes in mass will continue to be associated with similar changes in fatality risk. If the historical relationships change, one will be able to analyze the mass and size attributes independently, but it will take some years to get such data. However, as a check of DRI's suggestion that mass was not as significant as track width and wheelbase, NHTSA ran both its 1997 and 2003 analyses of 4-door cars only with mass, track width, and wheelbase as separate variables. When we did this, we saw that mass continued to have a substantial effect, even independent of track width and wheelbase in all crash modes except rollovers. In fact, only curb weight had a consistent, significant effect in both the data sets used in NHTSA's 1997 analyses and his 2003 analyses. This was publicly reported over four years ago, in NHTSA's November 2004 response to the comments on his 2003 analyses. After considering the DRI submission, NHTSA made no change to the findings in its 2003 report. (2) Marc Ross, of the University of Michigan, and Tom Wenzel, of Lawrence Berkeley National Laboratory, commented that vehicle ``quality'' has a much stronger relationship with fatality risk than vehicle mass. They suggest that lighter cars have a higher fatality risk on average because they are usually the least expensive cars and, in many cases, the ``poorest quality'' cars. If true, weight reduction is fairly harmless, as long as the lighter cars are of the same ``quality'' as the heavier cars they replace. Agency response: In their analyses, Ross and Wenzel did not adjust their rates for driver age and gender. Absent those adjustments, the analysis mingles the effects of what sort of people buy and drive the car with the intrinsic safety of the car, making its conclusions about the intrinsic safety of the car suspect, at best. On average, and considering all crash modes as well as both weight groups of cars, controlling for price has little effect on the weight-safety coefficients in NHTSA's analyses. As a final check, NHTSA ran an analysis of head-on collisions of two 1991-99 cars, since this is a pure measure of the vehicle's performance. The results were that the more expensive vehicle's driver had a slightly higher fatality risk than the less expensive vehicle's driver, although the difference was not statistically significant. This indicates that the lower fatality rates for more expensive cars in Ross and Wenzel's study are not due to expensive cars' superior performance in crashes. Accordingly, NHTSA the Ross and Wenzel comment did not warrant a change in NHTSA's report. (3) The Alliance of Automobile Manufacturers, DaimlerChrysler, William E. Wecker Associates, and Environmental Defense all question the accuracy and robustness of the report's calculation of a ``crossover weight,'' above which weight reductions have a net benefit, instead of harm. NHTSA's report said that this crossover point occurs somewhere in the range of 4,224 pounds to 6,121 pounds (this is the ``interval estimate''); with the most likely location of the crossover point at 5,085 pounds (this is the ``point estimate''). Wecker suggested that NHTSA's interval estimate of from 4,224 to 6,121 pounds only takes sampling error into account. Wecker identified additional factors that make this estimate not robust, and suggests that the interval estimate should be wider. The Alliance and DaimlerChrysler suggested that the crossover weight could be substantially greater than 5,085 pounds, in which case weight reductions for light trucks and vans in the 5-6,000 pound range would have detrimental net effects on safety. Conversely, Environmental Defense believes the crossover weight is well below 5,085 pounds, in which case there would be opportunities to reduce vehicle mass in many light trucks and vans without any safety penalty. Agency response: While NHTSA's report estimates the crossover weight, the report expressly acknowledged the uncertainty about the exact location of the crossover weight. That is why the report highlighted the interval estimate, instead of the point estimate. It is important to note that the net weight-safety relationship remains close to zero for many hundreds of pounds above and below the point estimate for the crossover weight. As shown on pages 163-166 of NHTSA's 2003 report, the crash fatality rate changes by less than ±1 percent per 100-pound weight increase over a 1,200 pound range on either side of the point estimate for the crossover weight. The data and analysis in the report will not show a statistically significant relationship, in either direction, between weight and safety for the heavier light trucks and vans. That is the important information the report puts in front of the decision maker--that the robust relationship between weight and safety that exists for most vehicles does not exist for the heavier light trucks and vans. With the available data, one cannot develop a precise point estimate for this crossover weight. Thus, NHTSA determined that its report did not require changes in response to these comments. 4. Summary of Studies Prior to This Rulemaking Several important observations can be made based on the various studies performed in the years preceding this rulemaking on the relationship between safety and vehicle weight in the context of fuel economy: 1. The question of the effect of weight on vehicle safety is a complex question that poses serious analytic challenges. The issue has been addressed in the literature for more than two decades. 2. NHTSA has been actively engaged in this discussion. 3. All of NHTSA's analyses have found that there is a strong correlation [[Page 14401]] between vehicle mass and vehicle safety for cars and light trucks, up to a certain weight range. a. Given the historic fact that vehicles have been made primarily of steel, there are a number of other parameters that are highly correlated with vehicle mass. These factors include vehicle size (e.g., track width and wheelbase). b. The precise weight point at which the safety penalty ends is difficult to pinpoint, because the fatality rate curve is so flat at that point. NHTSA can say with high confidence that the crossover point is in the range of 4,224 to 6,121 pounds. There are safety penalties for reductions of weight below this crossover weight. There is no reduced societal safety for reducing weight on vehicles that weigh more than this crossover point, because the reduced risk for other road users would exceed any reduced benefits for the occupants of the heavy vehicle. 4. The National Academy of Sciences has twice peer-reviewed NHTSA's work in this area. The 2002 NAS Report found that there was a safety penalty for reducing weight in all but the heaviest light trucks. The study stated that ``the downsizing and weight reduction that occurred in the late 1970s and early 1980s most likely produced between 1,300 and 2,600 crash fatalities in 1993.'' a. Neither the Academy nor NHTSA is suggesting that all of the downsizing and weight reduction were a direct response to the CAFE standards. It is difficult to objectively quantify what amount of downsizing was a response to CAFE standards, and what was a response to other real or perceived market forces. However, the Academy stated that some of the downsizing was in response to CAFE standards. b. NHTSA does not accord the safety dissent, which represented the views of two of the 13 committee members, the same stature as the views expressed in the body of the report, which represents the views of 11 of the 13 committee members. 5. In response to the National Academy's unanimous 2002 recommendation, NHTSA updated its previous work on weight and safety in 2003 to reflect the most recent data. This update found that the trends were similar, and if anything the safety penalty was now higher for reducing weight in small cars. This update also found that there is a crossover weight, which occurs somewhere between 4,264 and 6,121 pounds, with a point estimate at 5,085 pounds, above which there is no safety penalty for reducing vehicle weight. This is because the added harm for other road users from the additional weight exceeds any benefits for the occupants of the vehicles. NHTSA embodied this finding in its CAFE rulemaking by restricting materials substitution in its development of stringency levels to vehicles over 5,000 pounds. 6. NHTSA published its update and asked for public comments on the updated document. 7. In response to the request for comments, NHTSA received two recent studies to review. After reviewing these studies, NHTSA concluded that both studies had inadvertently introduced significant biases in their analyses. NHTSA made public its review of these studies in November 2004. a. One of these studies was a 2002 study by DRI that purported to analyze mass, track width, and wheelbase as independent variables. DRI's 2002 paper indicated that reducing mass would be beneficial, while reducing track width and wheelbase would be harmful. If true, this meant that weight reduction would benefit safety if track width and wheelbase were maintained. As discussed above, NHTSA concluded that the DRI results were strongly biased as a consequence of including 2- door cars in the analysis and explained why this was so.\460\ --------------------------------------------------------------------------- \460\ As discussed below, DRI acknowledged this observation to be accurate and submitted a new 2005 analysis that excludes 2-door cars in response to NHTSA's suggestions. --------------------------------------------------------------------------- b. The other of these studies was a 2002 analysis by Ross and Wenzel that suggested that lighter cars have a higher fatality risk because they are the least expensive and, in many cases, the poorest quality cars. The implication of this analysis was that weight reduction is fairly harmless, as long as the lighter cars are of the same ``quality'' as the heavier cars they replace. NHTSA noted that the Ross and Wenzel analyses did not adjust for driver age and gender. Absent those adjustments, the analysis mingles the effects of what sort of people buy and drive the car with the intrinsic safety of the car, making its conclusions about the intrinsic safety of the car suspect, at best. B. Response to Comments in This Rulemaking on Safety and Vehicle Weight With this background, NHTSA will now address the comments it received on safety in response to its NPRM. First, however, it is important to understand how NHTSA has embodied the accumulated knowledge and expertise from the studies explained above in this final rule. The rule is a performance standard that does not dictate the way manufacturers satisfy the standard. It does not preclude manufacturers from reducing the weight of future vehicles. Instead, in calculating its stringency standards, NHTSA has not considered weight-reducing materials substitution as a methodology for improving fuel economy of vehicles of 5,000 pounds or less. NHTSA has done so based on available data in order not to encourage downsizing of vehicles in a way that would be likely to cause a significant number of deaths and injuries. At the same time, for vehicles above 5,000 pounds, where the data indicate no safety penalty is likely for reducing weight, NHTSA has considered materials substitution in its standard-setting analysis. The effect of this is to encourage weight reductions to improve fuel economy where doing so is not likely to endanger lives. We believe this careful drawing of a data-based line in our analysis is the best way to serve both safety and fuel economy. As an overview, many commenters questioned the continuing validity of the 2002 NAS Report, the 2003 NHTSA study led by Dr. Kahane, or both. NHTSA notes both these reports were based on considerable empirical data and thoroughly peer-reviewed. More recent studies will need to be of a very high quality for NHTSA to adopt them in lieu of the the 2002 NAS Report and the 2003 NHTSA analyses. 1. Views of Other Government Agencies After our proposed rule was published and after the comment period had closed for the proposal, EPA published an Advance Notice of Proposed Rulemaking (ANPRM) on regulating greenhouse gas emissions under the Clean Air Act.\461\ The ANPRM was accompanied by a Vehicle Technical Support Document--Mobile Source.\462\ The Technical Support Document contains a discussion on pp. 15-17 of the safety issues. EPA provided a brief summary of the issues involved and cited no new work in that area. --------------------------------------------------------------------------- \461\ 73 FR 44354 (July 30, 2008). \462\ Docket No. EPA-HQ-OAR-2008-0318-0084. --------------------------------------------------------------------------- Agency response: The work cited by EPA has already been addressed by NHTSA within the discussion of the 2002 NAS study and within NHTSA's responses to other comments to the NPRM docket regarding the Wenzel and Ross study. CARB also commented on the relationship between vehicle weight and safety. CARB stated that the NHTSA study led by Dr. Kahane ``assumed that weight and size are completely correlated,'' and argued that NHTSA should have focused more closely on [[Page 14402]] the DRI reports and other recent studies, which it said concluded that ``safety is primarily a design issue, not a weight issue.'' CARB included with its comments an ``expert report by David Greene,'' which it said concluded after reviewing the existing research that ``there has been no relationship between fuel economy and traffic fatalities and that there should be none in the future.'' CARB also commented that it believed that NHTSA was inconsistent by restricting materials substitution in its analysis to only vehicles over 5,000 pounds, but also stating in the NPRM that footprint-based standards would facilitate the use of lightweight materials that are not yet cost-effective, which could eventually improve both safety and fuel economy. CARB argued that ``NHTSA should expand the applicability of weight reduction technologies to vehicles under 5,000 pounds,'' because weight reduction can be ``a viable technology if accompanied by proper vehicle design to assure vehicle safety is not compromised.'' Agency response: The available empirical data are derived from vehicles that are in use on the public roads, and weight and size are highly correlated in those vehicles. Underlying this, larger vehicles contain more steel and weigh more. NHTSA has not and is not now claiming that weight and size are completely correlated. Thus, for any given curb weight, there may not be some variations in the track widths and wheelbases of vehicle make-models at that curb weight. However, these variations are not random--they are nearly always correlated with the vehicle's market class or design group. NHTSA agrees that, conceptually, substitution of strong, lightweight materials should be a less harmful way to downweight than reducing the size of the vehicle. CARB has not supported its concept by presenting information on how this would be achieved or the consequences on the feasibility and practicability of doing so. There is not yet sufficient empirical evidence to conclude that material substitution is harmless, let alone beneficial to safety. NHTSA is proceeding cautiously and erring on the side of the safety of the public until there is more convincing evidence that requiring investments by vehicle makers in greater fuel efficiency through use of lightweight materials will not have the significant unintended consequence of simultaneously reducing the safety protection afforded to the American people, and attendant deaths as have occurred in the past. As for the DRI reports, NHTSA reviewed its 2002 report and publicly responded in 2004 that the DRI results were strongly biased as a result of including 2-door cars in the analysis. To DRI's credit, they reviewed their report and agreed that this flaw needed to be corrected. DRI submitted a new study which, they say, limited some of their analyses to 4-door cars excluding police cars. DRI further claimed that it could now mimic NHTSA's logistic regression approach for an analysis of model year 1991-98 4-door cars in calendar year 1995-1999 crashes. DRI claims that its new analysis still shows results directionally similar to its earlier work--increased risk for lower track width and wheelbase, reduced risk for lower mass--although DRI acknowledges that the wheelbase and mass effects are no longer statistically significant after removing the 2-door cars from the analysis. NHTSA does not accept the updated DRI analysis because it contains results that are inconsistent with results NHTSA has seen and, in light of this, DRI has not justified its results. For example in MY 1991- 1998, the average car weighing x + 100 pounds had a track width that was 0.34 inches larger and a wheelbase that was 1.01 inch longer. Thus, we could say that a ``historical'' 100-pound weight reduction would have been accompanied by a 0.34 inch track width reduction and a 1.01 inch wheelbase reduction. However, using a reasonable check, if one dissociates weight, track width, and wheelbase and treats them as independent parameters, DRI's logistic regression of model year 1991- 1998 4-door cars excluding police cars attributes the following effects: [GRAPHIC] [TIFF OMITTED] TR30MR09.077 Now if we apply NHTSA's logistic regression analyses to NHTSA's database, exactly as described in the agency's response to comments on its 2003 report, except for limiting the data to model years 1991-98, instead of 1991-99, the results are not at all like DRI's. For NHTSA, mass still has the largest effect, exceeding track width, and it moves in the expected direction. [GRAPHIC] [TIFF OMITTED] TR30MR09.078 [[Page 14403]] NHTSA obtains its estimates by adding the results from 12 individual logistic regressions: six types of crashes multiplied by two car-weight groups (less than 2,950 pounds; 2,950 pounds or more).\463\ DRI has apparently not followed the same procedures, based on the widely differing results. --------------------------------------------------------------------------- \463\ See, e.g., Kahane (2003), Table 2 on P. xi. --------------------------------------------------------------------------- Based on the evidence before us now, NHTSA is not persuaded by the DRI analysis. Even though NHTSA's analyses continue to attribute a much larger effect for mass than for track width or wheelbase in small cars, NHTSA has never said that mass alone is the single factor that increases or decreases fatality risk. There may not be a single factor, but rather it may be that mass and some of the other factors that are historically correlated with mass, such as wheelbase and track width, together are the factors. We can say that NHTSA's analyses do not corroborate the 2005 DRI analysis, suggesting that mass can be reduced without safety harm and perhaps with safety benefit. We would note that comparatively, it would seem the least harmful way to reduce mass would be from materials substitution, where one replaces a heavy material with a lighter one that delivers the same performance, or other designs that reduce mass while maintaining wheelbase and track width. There is an absence of supporting data for the thrust of the 2005 DRI analysis. We cannot analyze data on that yet, because those changes have not happened to any substantial number of vehicles. We do know that mass has historically been correlated with wheelbase and track width, and that reductions in mass have also reduced those other factors. Until there is a more credible analysis than the 2005 DRI study that demonstrates that mass does not matter for safety, NHTSA concludes it should be guided by the decades' worth of studies suggesting that mass is the most important of the related factors. The report by Dr. David Greene that was submitted by CARB as part of its comments is a document submitted by Dr. Greene when he was an expert witness in a lawsuit.\464\ We note that Dr. Greene was one of the two dissenters to the 2002 NAS report. Dr. Greene reiterates the arguments in his dissent to the 2002 NAS Report; namely, mass alone should not have any safety effect except in crashes where two vehicles collide with each other (which undisputedly occurs, with fatal results). In light of this view, all the empirical data showing higher fatality rates for lighter vehicles in single-vehicle crashes and elsewhere are due to something other than mass. Therefore, we conclude mass may be reduced without harming safety. But, as explained above, mass has been historically correlated with other factors, such as size and structural integrity. Unless NHTSA can determine based on data what the significant parameters are and demonstrate ways to reduce mass without affecting the significant parameters, NHTSA cannot simply ignore the empirical data showing higher fatality rates for lighter vehicles. --------------------------------------------------------------------------- \464\ This is the same Dr. Greene who concluded in his 1997 report, cited above, that ``it is probably reasonable to conclude that reducing vehicle mass to improve vehicle economy will require some trade-off with safety.'' --------------------------------------------------------------------------- Dr. Greene's expert report refers to the Ross and Wenzel and DRI studies, which have been discussed at length above. Dr. Greene also refers to a study titled ``The Effect of Fuel Economy on Automobile Safety: A Reexamination.'' \465\ This report is a long-term (1966-2002) time-series analysis of the annual number of crash fatalities in the United States, the average fuel economy of the vehicles on the road that year, and some other factors such as the price of fuel, the national speed limit, population, and annual vehicle miles traveled. The conclusion is that national fatalities did not increase, in fact tended to decrease, from the early 1970s forward, while fuel economy improved. Therefore, fuel economy has not had an adverse effect on safety. Suffice it to say that this is an exceedingly ``macro'' level to examine the relationships between fuel economy and fatality risk. Long-term time-series analyses are unlikely to separate the effects of downsizing for the other demographic, economic, and technological trends that have had an impact on fatality rates over the period. For instance, seat belt use has risen from 14 percent to 82 percent, many life-saving safety features (e.g., front and side airbags) have been added to vehicles, impaired driving is not as accepted, and so forth. It is general knowledge that traffic fatalities are now lower than 1970, primarily as a result of the major safety advances just mentioned. The reexamination ignores the effects of these variables and leaps to the conclusion that fuel economy did not have an adverse effect on safety--a conclusion that is at odds with the 2002 NAS study. But the relevant question in the safety/fuel economy context is, ``Would fatalities have been even lower if cars had not been downsized?'' To analyze that relationship accurately, it would be necessary to compare the fatality risk of small and large vehicles, not just the trend in total fatalities, over this long period. --------------------------------------------------------------------------- \465\ Sanjana Ahmad and David L. Greene, 2005, ``Effect of Fuel Economy on Automobile Safety: A Reexamination,'' Transportation Research Record 1941, Transportation Research Board of the National Academy of Sciences. --------------------------------------------------------------------------- With respect to CARB's suggestion that NHTSA expand the applicability of weight reduction technologies to vehicles under 5,000 pounds, because weight reduction can be accompanied by proper vehicle design to assure vehicle safety is not compromised, the agency repeats its general view that there may be possibilities in the use materials substitution and other processes to reduce weight without reducing vehicle safety. This should be explored. However, there are no data or analyses that show this to be true today. NHTSA specifically does not find either the 2002 or 2005 DRI analyses to be demonstrative, since the former study was strongly biased by including 2-door cars and the latter study says it mimicked NHTSA's database and NHTSA's analysis method, but got results that are substantially different. Until NHTSA can see thorough evidence using a significant and valid empirical data set, which is yet to be presented, that weight reduction can be accomplished without safety trade-offs, the agency will continue to set its CAFE standards at levels that do not encourage weight reduction in vehicles that weigh less than the safety crossover identified in NHTSA's 2003 analyses. We recognize that given the lives at stake, this reflects caution, but we believe it is also prudent. We also note that the California CO2 emissions standards for which California requested a waiver under the Clean Air Act sets up a program that uses the same ``flat standards'' approach for its standards that the 2002 NAS Report found gives rise to the safety concerns identified in that report. The consequences of this structure for the program have been identified by 2002 report: ``If an increase in fuel economy is effected by a system that encourages either downweighting or the production and sale of more small cars, some additional traffic fatalities would be expected. Without a thoughtful restructuring of the program, that would be the trade-off that must be made if CAFE standards are increased by any significant amount.'' \466\ --------------------------------------------------------------------------- \466\ 2002 NAS Report at 77. --------------------------------------------------------------------------- 2. Comments From Other Parties Several comments were received from parties other than government agencies on the weight-safety issue. NRDC argued that NHTSA should not have relied on [[Page 14404]] only on its 2003 study led by Dr. Kahane, because Wenzel and Ross had commented to NHTSA's 2005 light truck CAFE NPRM that ``the relationship between car weight and safety is tenuous at best,'' and because Dr. Kahane himself stated that his study ``does not claim that mass per se is the specific factor that increases or decreases fatality risk* * *'' ``In that sense, it is irrelevant whether mass, wheelbase, track width or some other attribute is the principal causal factor on fatality risk. If you decrease mass, you will also tend to reduce wheelbase, track width and other dimensions of size.'' NRDC stated that this may no longer be correct for future vehicle designs, and argued that NHTSA had recognized as much in the NPRM by stating that high-strength, light-weight materials may help manufacturers reduce vehicle weight without reducing size or safety. NRDC further argued that vehicle design, ``which could in fact be enhanced with lightweight materials,'' is much more relevant to safety. Thus, NRDC concluded that NHTSA should apply material substitution to lighter vehicles in its analysis. The comments received from Wenzel and Ross stand in direct contradiction to the 2002 NAS Report, which said, ``Thus, the majority of this committee believes that the evidence is clear that past downweighting and downsizing of the light-duty vehicle fleet, while resulting in significant fuel savings, has also resulted in a safety penalty.'' The Wenzel and Ross comment was also based on their study, discussed earlier, which NHTSA said in 2004 is flawed, since it did not control for driver age and gender. Thus, the findings of Wenzel and Ross are not helpful since they mingle the effects of what sort of people buy and drive the car with the intrinsic safety of the car, making its conclusions about the intrinsic safety of the car suspect, at best. NRDC is correct insofar as NHTSA has not claimed that mass alone is the single factor that is entirely responsible for the safety factor, and in the future there may be demonstrations that weight (the amount has not been identified) can be removed without adversely affecting safety. However, as we said in response to the same point from CARB, when setting CAFE standards, NHTSA will continue to limit its consideration of weight reduction to vehicles over 5,000 pounds until there is convincing empirical evidence that there are no negative safety consequences from removing weight from lighter vehicles. Sierra Club et al. also commented that vehicle design is more important than weight to vehicle safety. This is largely the same point made by other commenters. The point is very general, and there are no analyses that demonstrate this proposition is true. Sierra Club also argued that NHTSA should not use its retrospective 2003 study to analyze future standards, because of the design improvements and because ``[s]ubstitution of light weight, high strength materials such as low alloy steels and aluminum will decrease both primary and secondary vehicle weight while maintaining vehicle size and increasing crashworthiness.'' NHTSA believes that it would be irresponsible to set standards by ignoring the available data, based on the hope that a promising development will come to fruition. The available data indicate that there is a safety penalty for weight reductions in vehicles under a certain weight. Sierra Club et al. also stated that ``The industry's long history of consistent opposition to the CAFE law has relied on a flawed size/ safety argument,'' which it suggested also affected Congress' action in establishing EISA. Sierra Club argued, however, that that argument was disproven by the fact that manufacturers can obviously build vehicles that ``demonstrate size, safety, and fuel economy performance'' such as the Prius or the hybrid Escape. These vehicles tend to be cited for use of hybrid propulsion systems. They often have heavy battery systems but lighter engines. In any event, manufacturers continue to offer a full range of vehicles, and they strive to deliver safety, fuel economy, and value in all of their vehicles. However, the available data at the level of the entire fleet demonstrate that, below a certain weight range, there has been a safety penalty from downweighting vehicles. The introduction of new vehicle models does nothing to change that historical record and it is unknown how the new models will affect the fleet wide fatality risk in future years. Sierra Club additionally repeated the oft-stated assertion that smaller cars continue to become safer as manufacturers ``apply side airbags, design vehicles to better protect occupants, and utilize light weight materials that enhance safety.'' It is of course true that, with the advent of important safety features like side air bags and Electronic Stability Control, combined with higher levels of seat belt use, today's small vehicles should have a better safety record than those produced a decade ago. However, that is not really the question that is being considered in deciding on the safety penalty for weight reduction--the question is whether today's small vehicles have a safety penalty compared to today's vehicles that weigh 100 pounds more. Unless there are some safety technologies that are offered only on small cars, or that are more effective on small cars, the additional safety technologies will not affect the relative safety performance between vehicles with a 100-pound weight difference. It is proper to compare vehicles of the same time period, not a light vehicle today with air bags and a heavy vehicle of years ago without air bags. If offered today, the heavy vehicle would have air bags and better safety performance. Sierra Club also argued that a study by the Center for Auto Safety and UCS ``found that applying existing fuel-saving and safety technology to a conventional Ford Explorer would result in a 71 percent improvement in fuel economy and 2,900 fewer traffic fatalities if all SUVs met equivalent safety standards,'' while ``At the same time, the redesigned vehicle resulted in greater consumer savings and lower global warming emissions as a result of the improved fuel economy.'' \467\ The document generated by the Center for Auto Safety and UCS does not address the safety penalty as weight is reduced. This document asserts that if several safety and fuel-savings technologies were used on a 2001 Ford Explorer, it would achieve greater fuel economy and have a better safety record. The safety and fuel savings benefits, along with the costs, are extrapolated from different sources. The paper does state that the redesign would reduce the test weight of the vehicle by 10 percent, to 4100 pounds (p. 10). However, the question of the safety consequences of reducing the vehicle mass by 400 pounds is not answered by any data, since the redesigned vehicle does not exist. As such, this document is not persuasive. --------------------------------------------------------------------------- \467\ Sierra Club et al. cited ``Building a Better SUV: A Blueprint for Saving Lives, Money and Gasoline,'' by CAS and UCS. This 2003 pamphlet is accessible online at http://www.ucsusa.org/ assets/documents/clean_vehicles/building_a_better_suv_web.pdf