![[logo] US EPA](http://www.epa.gov/epafiles/images/logo_epaseal.gif){kind=logo}
Proposed Rulemaking To Establish Light-Duty Vehicle Greenhouse Gas Emission Standards and Corporate Average Fuel Economy Standards
Note: EPA no longer updates this information, but it may be useful as a reference or resource.
PDF Version (50 pp, 1854K, About PDF) [Federal Register: September 28, 2009 (Volume 74, Number 186)] [Proposed Rules] [Page 49653-49702] From the Federal Register Online via GPO Access [wais.access.gpo.gov] [DOCID:fr28se09-28] Proposed Rulemaking To Establish Light-Duty Vehicle Greenhouse Gas Emission Standards and Corporate Average Fuel Economy Standards [[Continued from page 49652]] [[Page 49653]] Industry Average............. 0.056 0.056 0.056 ------------------------------------------------------------------------ Table IV.C.1-6b--MY 2011 Final Rule Average Planned MY 2011 Vehicle Power-to-Weight Ratio [hp/lb] ------------------------------------------------------------------------ PC LT Avg. ------------------------------------------------------------------------ Manufacturer 1................... 0.065 0.058 0.060 Manufacturer 2................... 0.061 0.065 0.062 Manufacturer 3................... 0.053 0.059 0.056 Manufacturer 4................... 0.060 0.058 0.059 Manufacturer 5................... 0.060 0.057 0.059 Manufacturer 6................... 0.063 0.065 0.065 Manufacturer 7................... 0.053 0.055 0.053 -------------------------------------- Industry Average............. 0.060 0.059 0.060 ------------------------------------------------------------------------ As discussed above, the agencies' market forecast for MY 2012-2016 holds the performance and other characteristics of individual vehicle models constant, adjusting the size and composition of the fleet from one model year to the next. Refresh and redesign schedules (for application in NHTSA's modeling): Expected model years in which each vehicle model will be redesigned or freshened constitute another important aspect of NHTSA's market forecast. As discussed in Section IV.C.2.c below, NHTSA's analysis supporting the current rulemaking times the addition of nearly all technologies to coincide with either a vehicle redesign or a vehicle freshening. Product plans submitted to NHTSA preceding the MY 2011 final rule contained manufacturers' estimates of vehicle redesign and freshening schedules and NHTSA's estimates of the timing of the five- year redesign cycle and the two- to three-year refresh cycle were made with reference to those plans. In the current baseline, in contrast, estimates of the timing of the refresh and redesign cycles were based on historical dates--i.e., counting forward from known redesigns occurring in or prior to MY 2008 for each vehicle in the fleet and assigning refresh and redesign years accordingly. After applying these estimates, the shares of manufacturers' passenger car and light truck estimated to be redesigned in MY 2011 were as summarized below for the current baseline and the MY 2011 final rule. Table IV.C.1-7 below shows the percentages of each manufacturer's fleets expected to be redesigned in MY 2011 for the current baseline. Table IV.C.1-8 presents corresponding estimates from the market forecast used by NHTSA in the analysis supporting the MY 2011 final rule (again, to protect confidential information, manufacturers are not identified by name). Table IV.C.1-7--Current Baseline, Share of Fleet Redesigned in MY 2011 ------------------------------------------------------------------------ Manufacturer PC LT Avg. ------------------------------------------------------------------------ BMW.............................. 32% 40% 34% Chrysler......................... 0% 11% 8% Ford............................. 12% 7% 10% Subaru........................... 0% 51% 22% General Motors................... 20% 2% 11% Honda............................ 31% 33% 32% Hyundai.......................... 20% 0% 16% Tata............................. 28% 100% 73% Kia.............................. 35% 87% 48% Mazda............................ 0% 0% 0% Daimler.......................... 0% 0% 0% Mitsubishi....................... 0% 56% 7% Nissan........................... 4% 18% 9% Porsche.......................... 0% 100% 41% Suzuki........................... 8% 21% 11% Toyota........................... 4% 24% 12% Volkswagen....................... 23% 0% 18% -------------------------------------- Industry Average............. 15% 17% 15% ------------------------------------------------------------------------ Table IV.C.1-8--MY 2011 Final Rule, Share of Fleet Redesigned in MY 2011 ------------------------------------------------------------------------ PC LT Avg. (percent) (percent) (percent) ------------------------------------------------------------------------ Manufacturer 1................... 19 0 11 [[Page 49654]] Manufacturer 2................... 34 27 29 Manufacturer 3................... 5 0 3 Manufacturer 4................... 7 0 5 Manufacturer 5................... 19 0 11 Manufacturer 6................... 34 28 33 Manufacturer 7................... 27 28 28 -------------------------------------- Overall...................... 20 9 15 ------------------------------------------------------------------------ We continue, therefore, to estimate that manufacturers' redesigns will not be uniformly distributed across model years. This is in keeping with standard industry practices, and reflects what manufacturers actually do--NHTSA has observed that manufacturers in fact do redesign more vehicles in some years than in others. NHTSA staff have closely examined manufacturers' planned redesign schedules, contacting some manufacturers for clarification of some plans, and confirmed that these plans remain unevenly distributed over time. For example, although Table 8 shows that NHTSA expects Company 2 to redesign 34 percent of its passenger car models in MY 2011, current information indicates that this company will then redesign only (a different) 10 percent of its passenger cars in MY 2012. Similarly, although Table 8 shows that NHTSA expects four of the largest seven light truck manufacturers to redesign virtually no light truck models in MY 2011, current information also indicates that these four manufacturers will redesign 21-49 percent of their light trucks in MY 2012. e. How Does Manufacturer Product Plan Data Factor Into the Baseline Used in This Proposal? As discussed in Section II.B.4 above, while the agencies received updated product plans in Spring 2009 in response to NHTSA's request, the baseline data used in this proposal is not informed by these product plans, because they contain confidential business information the agencies are legally required to protect from disclosure, and because the agencies have concluded that, for purposes of this NPRM, a transparent baseline is preferable. However, as also discussed above, NHTSA has conducted a separate analysis that does make use of these product plans, contained in NHTSA's PRIA. NHTSA performed this separate analysis for purposes of comparison only. NHTSA used the publicly available baseline for all analysis related to the development and evaluation of the proposed new CAFE standards. 2. How Were the Technology Inputs Developed? As discussed above in Section II.E, for developing the technology inputs for the MY 2012-2016 CAFE and GHG standards, the agencies primarily began with the technology inputs used in the MY 2011 CAFE final rule and in the July 2008 EPA ANPRM, and then reviewed, as requested by President Obama in his January 26 memorandum, the technology assumptions that NHTSA used in setting the MY 2011 standards and the comments that NHTSA received in response to its May 2008 Notice of Proposed Rulemaking. In addition, the agencies supplemented their review with updated information from more current literature, new product plans and from EPA certification testing. More detail is available regarding how the agencies developed the technology inputs for this NPRM above in Section II.E, in Chapter 3 of the Draft Joint TSD, and in Section V of NHTSA's PRIA. a. What Technologies Does NHTSA Consider? Section II.E.1 above describes the fuel-saving technologies considered by the agencies that manufacturers could use to improve the fuel economy of their vehicles during MYs 2012-2016. The majority of the technologies described in this section are readily available, well known, and could be incorporated into vehicles once production decisions are made. As discussed, the technologies considered fall into five broad categories: Engine technologies, transmission technologies, vehicle technologies, electrification/accessory technologies, and hybrid technologies. Table IV.C.2-1 below lists all the technologies considered and provides the abbreviations used for them in the Volpe model,\464\ as well as their year of availability, which for purposes of NHTSA's analysis means the first model year in the rulemaking period that the Volpe model is allowed to apply a technology to a manufacturer's fleet.\465\ Year of availability recognizes that technologies must achieve a level of technical viability before they can be implemented in the Volpe model, and are thus a means of constraining technology use until such time as it is considered to be technologically feasible. For a more detailed description of each technology and their costs and effectiveness, we refer the reader to Chapter 3 of the joint TSD and Section V of NHTSA's PRIA. --------------------------------------------------------------------------- \464\ The abbreviations are used in this section both for brevity and for the reader's reference if they wish to refer to the expanded decision trees and the model input and output sheets, which are available in Docket No. NHTSA-2009-0059 and on NHTSA's Web site. \465\ A date of 2011 means the technology can be applied in all model years, while a date of 2014 means the technology can only be applied in model years 2014 through 2016. Table IV.C.2-1--List of Technologies in NHTSA's Analysis ------------------------------------------------------------------------ Technology Model abbreviation Year available ------------------------------------------------------------------------ Low Friction Lubricants........... LUB................. 2011 Engine Friction Reduction......... EFR................. 2011 VVT--Coupled Cam Phasing (CCP) on CCPS................ 2011 SOHC. Discrete Variable Valve Lift DVVLS............... 2011 (DVVL) on SOHC. Cylinder Deactivation on SOHC..... DEACS............... 2011 [[Page 49655]] VVT--Intake Cam Phasing (ICP)..... ICP................. 2011 VVT--Dual Cam Phasing (DCP)....... DCP................. 2011 Discrete Variable Valve Lift DVVLD............... 2011 (DVVL) on DOHC. Continuously Variable Valve Lift CVVL................ 2011 (CVVL). Cylinder Deactivation on DOHC..... DEADD............... 2011 Cylinder Deactivation on OHV...... DEACO............... 2011 VVT--Coupled Cam Phasing (CCP) on CCPO................ 2011 OHV. Discrete Variable Valve Lift DVVLO............... 2011 (DVVL) on OHV. Conversion to DOHC with DCP....... CDOHC............... 2011 Stoichiometric Gasoline Direct SGDI................ 2011 Injection (GDI). Combustion Restart................ CBRST............... 2014 Turbocharging and Downsizing...... TRBDS............... 2011 Exhaust Gas Recirculation (EGR) EGRB................ 2013 Boost. Conversion to Diesel following DSLC................ 2011 CBRST. Conversion to Diesel following DSLT................ 2011 TRBDS. 6-Speed Manual/Improved Internals. 6MAN................ 2011 Improved Auto. Trans. Controls/ IATC................ 2011 Externals. Continuously Variable Transmission CVT................. 2011 6/7/8-Speed Auto. Trans with NAUTO............... 2011 Improved Internals. Dual Clutch or Automated Manual DCTAM............... 2011 Transmission. Electric Power Steering........... EPS................. 2011 Improved Accessories.............. IACC................ 2011 12V Micro-Hybrid.................. MHEV................ 2011 Belt Integrated Starter Generator. BISG................ 2011 Crank Integrated Starter Generator CISG................ 2011 Power Split Hybrid................ PSHEV............... 2011 2-Mode Hybrid..................... 2MHEV............... 2011 Plug-in Hybrid.................... PHEV................ 2011 Mass Reduction 1 (1.5%)........... MS1................. 2011 Mass Reduction 2 (3.5%-8.5%)...... MS2................. 2014 Low Rolling Resistance Tires...... ROLL................ 2011 Low Drag Brakes................... LDB................. 2011 Secondary Axle Disconnect 4WD..... SAX................. 2011 Aero Drag Reduction............... AERO................ 2011 ------------------------------------------------------------------------ For purposes of this NPRM and as discussed in greater detail in the joint TSD, NHTSA and EPA carefully reviewed the list of technologies used in the agency's analysis for the MY 2011 final rule. Given the relatively short amount of time, from a technology-development perspective, that has elapsed since March 2009 and this NPRM, NHTSA and EPA concluded that the considerable majority of technologies were correctly defined and continued to be appropriate for use in the analysis supporting the proposed standards. However, some refinements were made as discussed below. Specific to its modeling, NHTSA has revised eight of the technologies used in the current analysis from those considered in the MY 2011 final rule. Specifically, two technologies which were previously unavailable in the MY 2011 time frame are now available (in the extended MY 2012-2016 period); one technology has been combined with another; one is newly introduced; three have revised names and/or definitions; and one has been deleted entirely. These changes are discussed further below, and NHTSA seeks comment on both these changes and the validation of the unchanged technology assumptions and estimates. Availability: In the MY 2011 final rule, two of the engine technologies--EGR boost and combustion restart--were unavailable because they were not considered technologically feasible until beyond that rulemaking time frame. While both were described and discussed in the MY 2011 final rule, neither was applied in the modeling process that supported those standards.\466\ In this analysis, EGR boost becomes available in MY 2013, and combustion restart in MY 2014, so both are being applied by the Volpe model, as needed, in this analysis. --------------------------------------------------------------------------- \466\ As an additional note, since combustion restart was unavailable in the MY 2011 time frame, the technology titled diesel following combustion restart (DSLC), which as the name indicates was only applied after combustion restart, was also unavailable. Accordingly, DSLC, which was described and discussed in the MY 2011 final rule, is now available in the current analysis. --------------------------------------------------------------------------- Merging of technologies: In the MY 2011 final rule, higher voltage and improved alternator (HVIA) was used to represent changes in the design of the alternator, effectively optimizing it for higher efficiency (instead of for low cost as is typically done). For purposes of this analysis, the HVIA technology is no longer represented individually, but instead has been incorporated into a new-to-this- analysis technology called belt integrated starter generator, or BISG, as discussed next. New technology: In the MY 2011 final rule, two levels of mild hybrid technology were defined: A 12 volt micro-hybrid (MHEV) system, which utilized a belt-driven starter generator operating at 12 volts, and the more capable integrated starter generator technology (ISG) operating at higher voltages (100 volts). ISG envisioned both belt and crank configured starter generator systems. In the current proposal, and in an effort to offer a broader spectrum of more diversified mild hybrid technologies for the modeling process to choose from, NHTSA has added the BISG technology to the electrification decision tree, and redefined the ISG technology to be a crank mounted version of ISG, accordingly renamed to CISG. The BISG technology is a belt-coupled system like the 12-volt MHEV, but it operates at a higher voltage (e.g., 42 volts) and thus can make use of regenerative braking, as well as [[Page 49656]] potentially adding some limited motive power. Since BISG is a higher voltage system, optimization of the alternator occurs as part of the BISG technology application; hence the HVIA technology is no longer needed as a separate technology. Additionally, the CISG technology is now defined as a crank mounted system that operates at higher voltages (100 volts) than BISG, yet at lower voltage than the strong hybrids (300 volts) that make greater use of regenerative braking and provide greater motive power capability. Thus, three levels of mild hybrid technology exist in the current proposal, as opposed to the two levels offered in the MY 2011 final rule. Revisions and Deletions: The Mass Reduction/Material Substitution technologies have been revised for the current proposal. In the MY 2011 final rule, the Volpe model used three levels of material substitution technologies, referred to as MS1, MS2, and MS5, which were progressively applied to vehicles with curb weights in excess of 5,000 pounds (2,268 kg) so as to reduce weight by up to a 5 percent maximum. In keeping with the agency's desire to limit potential negative impacts to safety performance as a result of vehicle weight reduction, material substitution was not applied to vehicles with curb weights below 5,000 pounds. In contrast, in the current analysis, and in keeping with some manufacturers' stated plans to decrease overall fleet weights regardless of subclass or curb weight, NHTSA now defines two Mass Reduction/Material Substitution technologies as follows: The Mass Reduction 1.5 percent (MS1) represents a 1.5 percent weight decrease through material substitution applicable to all vehicle subclasses, regardless of curb weight, that can be applied throughout the rulemaking period (and at refresh or redesign cycle times). This technology is similar to the MS1 technology used in the prior analysis in terms of the scale of the weight reduction (1 versus 1.5 percent), the methods and techniques manufacturers are anticipated to use in achieving the reductions, and when in the product cycle the model applies it (at refresh or redesign). A second technology, Mass Reduction 3.5-8.5 percent (MS2), has also been defined. The MS2 technology is unavailable until MY 2014, and can only be applied by the Volpe model at a product redesign cycle. MS2 assumes a 3.5 to 8.5 percent weight reduction dependent on subclass (with the smaller/lighter subclasses receiving the lowest amounts of reduction, 3.5 percent, and the larger/heavier vehicles the 8.5 percent) via the types of more intrusive and complex mass reduction associated with a complete vehicle redesign.\467\ MS2 is cumulative to MS1, as it is only applied after MS1, therefore the maximum weight reduction that can occur for smaller subclass vehicles is 5 percent, while large cars, truck, and SUVs could experience 10 percent weight reductions. Restricting weight reduction on smaller vehicle to lower limits, and vice versa for larger vehicles, is intended to mitigate or minimize the potential safety consequences from the modeled weight reductions. Postponing the availability of the technology until MY 2014 recognizes the lead time required to implement platform redesigns that would be necessary for these levels of weight reduction and mass reduction. NHTSA seeks comment on the agency's use of a two-step process, with the higher applications of MS in MYs 2014 and beyond, and the process of applying smaller mass reductions to lighter vehicles and higher reductions to heavier vehicles for the purpose of maintaining safety neutrality. --------------------------------------------------------------------------- \467\ Examples of such weight savings associated with new platform introductions have been provided in confidential product plan information provided by some manufacturers. --------------------------------------------------------------------------- The MS5 technology used in the MY 2011 final rule is deleted. Additionally, for purposes of this NPRM, NHTSA has revised the applicability of the diesel technologies to restrict it to vehicles with engines of 6 cylinders or more. NHTSA seeks comment on its decision not to apply diesel technologies to vehicles with 4-cylinder engines. NHTSA also seeks comment on the revised costing methodology for diesel technologies. Besides these, all other technologies considered in this analysis were also considered in the analysis for the MY 2011 final rule, and although the costs and effectiveness estimates may have been revised as discussed further below, the other technologies remain otherwise unchanged for the purposes of this analysis in terms of their definition, functionality, and configuration. Thus, with this catalog of technologies as a starting point, NHTSA could then review and consider effectiveness and cost estimates for each technology, and, through the Volpe model analysis, how a manufacturer might feasibly apply these technologies to their MY 2012-2016 vehicles in order to achieve compliance with the proposed standards. b. How Did NHTSA Determine the Costs and Effectiveness of Each of These Technologies for Use in Its Modeling Analysis? Building on NHTSA's estimates developed for the MY 2011 CAFE final rule and EPA's Advanced Notice of Proposed Rulemaking, which relied on the 2008 Staff Technical Report,\468\ the agencies took a fresh look at technology cost and effectiveness values for purposes of the joint proposal under the National Program. This joint work is reflected in Chapter 3 of the Draft Joint TSD and in Section II of this preamble, which is summarized below. For more detailed information on the effectiveness and cost of fuel-saving technologies, please refer to Chapter 3 of the joint TSD and Section V of NHTSA's PRIA. --------------------------------------------------------------------------- \468\ EPA Staff Technical Report: Cost and Effectiveness Estimates of Technologies Used to Reduce Light-Duty Vehicle Carbon Dioxide Emissions. EPA420-R-08-008, March 2008. --------------------------------------------------------------------------- Generally speaking, while NHTSA and EPA found that much of the cost information used in NHTSA's MY 2011 final rule and EPA's 2008 staff report was consistent to a great extent, the agencies, in reconsidering information from many sources, revised several component costs of several major technologies: turbocharging/downsizing, mild and strong hybrids, diesels, SGDI, and Valve Train Lift Technologies. These are discussed at length in the joint TSD and in NHTSA's PRIA. Additionally, most effectiveness estimates used in both the MY 2011 final rule and the 2008 EPA staff report were determined to be accurate and were carried forward without significant change into this rulemaking. When NHTSA and EPA's estimates for effectiveness diverged slightly due to differences in how agencies apply technologies to vehicles in their respective models, we report the ranges for the effectiveness values used in each model. For much more information on the costs and effectiveness of individual technologies, we refer the reader to Chapter 3 of the joint TSD and Section V of NHTSA's PRIA. NHTSA notes that, in developing technology cost and effectiveness estimates, the agencies have made every effort to hold constant aspects of vehicle performance and utility typically valued by consumers, such as horsepower, carrying capacity, and towing and hauling capacity. For example, NHTSA includes in its analysis technology cost and effectiveness estimates that are specific to performance passenger cars (i.e., sports cars), as compared to non-performance passenger cars. When [[Page 49657]] commenting on the agencies' technology cost and effectiveness estimates, NHTSA urges commenters either to place any related comments within the same context, or explain any assumptions or estimates regarding increases or decreases in vehicle performance or utility. Additionally, NHTSA seeks comment on the extent to which commenters believe that the agencies have been successful in holding constant these elements of vehicle performance and utility in developing the technology cost and effectiveness estimates. Additionally, NHTSA notes that the technology costs included in this NPRM take into account only those associated with the initial build of the vehicle. The agencies seek comments on the additional lifetime costs, if any, associated with the implementation of advanced technologies, including warranty, maintenance and replacement costs, such as the replacement costs for low rolling resistance tires, low friction lubricants, and hybrid batteries, and maintenance costs for diesel aftertreatment components. While the agencies believe that the ideal estimates for the final rule would be based on tear down studies or BOM approach and subjected to a transparent peer-reviewed process, NHTSA and EPA are confident that the thorough review conducted, led to the best available conclusion regarding technology costs and effectiveness estimates for the current rulemaking and resulted in excellent consistency between the agencies' respective analyses for developing the CAFE and CO2 standards. NHTSA seeks comment on the incremental cost and effectiveness estimates employed by the agency in the Volpe modeling analysis for this NPRM, examples of which are provided in table form below. These example Tables present effectiveness and cost estimates which are incremental in nature, according to the decision trees used in the Volpe modeling analysis. Thus, the effectiveness and cost estimates are not absolute to a single baseline vehicle, but are incremental to the technology that precedes it. BILLING CODE 6560-50-C [[Page 49658]] [GRAPHIC] [TIFF OMITTED] TP28SE09.029 [[Page 49659]] [GRAPHIC] [TIFF OMITTED] TP28SE09.030 BILLING CODE 6560-50-C c. How Does NHTSA Use These Assumptions in Its Modeling Analysis? NHTSA's analysis, using the Volpe model, relies on several inputs and data files to conduct the compliance analysis, as discussed further below and in Section V of the PRIA. For the purposes of applying technologies, the Volpe model primarily uses three data files, one that contains data on the vehicles expected to be manufactured in the model years covered by the rulemaking, one that identifies the appropriate stage within the vehicle's life-cycle for the technology to be applied, and one that contains data/parameters regarding the available technologies the model can apply. These inputs are discussed below. As discussed above, the Volpe model begins with an initial state of the domestic vehicle market, which in this case is the market for passenger cars and light trucks to be sold during the period covered by the proposed standards. The [[Page 49660]] vehicle market is defined on a model-by-model, engine-by-engine, and transmission-by-transmission basis, such that each defined vehicle model refers to a separately defined engine and a separately defined transmission. For the current proposal, which covers MYs 2012-2016, the light vehicle (passenger car and light truck) market forecast was developed jointly by NHTSA and EPA staff using MY 2008 CAFE compliance data. The MY 2008 compliance data includes about 1,100 vehicle models, about 400 specific engines, and about 200 specific transmissions, which is a somewhat lower level of detail in the representation of the vehicle market than that used by NHTSA in recent CAFE analyses.\469\ However, within the limitations of information that can be made available to the public, it provides the foundation for a realistic analysis of manufacturer-specific costs and the analysis of attribute-based CAFE standards, and is much greater than the level of detail used by many other models and analyses relevant to light vehicle fuel economy.\470\ --------------------------------------------------------------------------- \469\ The market file for the MY 2011 final rule, which included data for MYs 2011-2015, had 5500 records, or rows, about 5 times what we are using in this analysis of the MY 2008 certification data. However, both market files had the same number of fields, or rows. \470\ Because CAFE standards apply to the average performance of each manufacturer's fleet of cars and light trucks, the impact of potential standards on individual manufacturers cannot be credibly estimated without analysis of fleets manufacturers can be expected to produce in the future. Furthermore, because required CAFE levels under an attribute-based CAFE standard depend on manufacturers' fleet composition, the stringency of an attribute-based standard cannot be predicted without performing analysis at this level of detail. --------------------------------------------------------------------------- In addition to containing data about each vehicle, engine, and transmission, this file contains information for each technology under consideration as it pertains to the specific vehicle (whether the vehicle is equipped with it or not), the model year the vehicle is undergoing redesign, and information about the vehicle's subclass for purposes of technology application. In essence, the model considers whether it is appropriate to apply a technology to a vehicle. Is a vehicle already equipped, or can it not be equipped, with a particular technology? The market forecast file provides NHTSA the ability to identify, on a technology by technology basis, which technologies may already be present (manufactured) on a particular vehicle, engine, or transmission, or which technologies are not applicable (due to technical considerations) to a particular vehicle, engine, or transmission. These identifications are made on a model-by-model, engine-by-engine, and transmission-by-transmission basis. For example, if the market forecast file indicates that Manufacturer X's Vehicle Y is manufactured with Technology Z, then for this vehicle Technology Z will be shown as used. Additionally, NHTSA has determined that some technologies are only suitable or unsuitable when certain vehicle, engine, or transmission conditions exist. For example, secondary axle disconnect is only suitable for 4WD vehicles, and cylinder deactivation is unsuitable for any engine with fewer than 6 cylinders, while CVTs can only be applied to unibody vehicles. Similarly, comments received to the 2008 NPRM indicated that cylinder deactivation could not be applied to vehicles equipped with manual transmissions, due primarily to driveability and NVH concerns. The Volpe model employs ``engineering constraints'' to address issues like these, which are a programmatic method of controlling technology application that is independent of other constraints. Thus, the market forecast file would indicate that the technology in question should not be applied to the particular vehicle/engine/transmission (i.e., is unavailable). Since multiple vehicle models may be equipped with an engine or transmission, this may affect multiple models. In using this aspect of the market forecast file, NHTSA ensures the Volpe model only applies technologies in an appropriate manner, since before any application of a technology can occur, the model checks the market forecast to see if it is either already present or unavailable. NHTSA seeks comment on whether this approach is reasonable and ensures that technologies are applied in an appropriate manner. Is a vehicle being redesigned or refreshed? Manufacturers typically plan vehicle changes to coincide with certain stages of a vehicle's life cycle that are appropriate for the change, or in this case the technology being applied. In the automobile industry there are two terms that describe when technology changes to vehicles occur: redesign and refresh (i.e., freshening). Vehicle redesign usually refers to significant changes to a vehicle's appearance, shape, dimensions, and powertrain. Redesign is traditionally associated with the introduction of ``new'' vehicles into the market, often characterized as the ``next generation'' of a vehicle, or a new platform. Vehicle refresh usually refers to less extensive vehicle modifications, such as minor changes to a vehicle's appearance, a moderate upgrade to a powertrain system, or small changes to the vehicle's feature or safety equipment content. Refresh is traditionally associated with mid-cycle cosmetic changes to a vehicle, within its current generation, to make it appear ``fresh.'' Vehicle refresh generally occurs no earlier than two years after a vehicle redesign, or at least two years before a scheduled redesign. For the majority of technologies discussed today, manufacturers will only be able to apply them at a refresh or redesign, because their application would be significant enough to involve some level of engineering, testing, and calibration work.\471\ --------------------------------------------------------------------------- \471\ For example, applying material substitution through weight reduction, or even something as simple as low rolling-resistance tires, to a vehicle will likely require some level of validation and testing to ensure that the vehicle may continue to be certified as compliant with NHTSA's Federal Motor Vehicle Safety Standards (FMVSS). Weight reduction might affect a vehicle's crashworthiness; low rolling-resistance tires might change vehicle's braking characteristics or how it performs in crash avoidance tests. --------------------------------------------------------------------------- Some technologies (e.g., those that require significant revision) are nearly always applied only when the vehicle is expected to be redesigned, like turbocharging and engine downsizing, or conversion to diesel or hybridization. Other technologies, like cylinder deactivation, electric power steering, and aerodynamic drag reduction can be applied either when the vehicle is expected to be refreshed or when it is expected to be redesigned, while a few others, like low friction lubricants, can be applied at any time, regardless of whether a refresh or redesign event is conducted. Accordingly, the model will only apply a technology at the particular point deemed suitable. These constraints are intended to produce results consistent with manufacturers' technology application practices. For each technology under consideration, NHTSA stipulates whether it can be applied any time, at refresh/redesign, or only at redesign. The data forms another input to the Volpe model. NHTSA develops redesign and refresh schedules for each of a manufacturer's vehicles included in the analysis, essentially based on the last known redesign year for each vehicle and projected forward in a 5-year redesign and a 2-3 year refresh cycle, and this data is also stored in the market forecast file. We note that this approach is different than NHTSA has employed previously for determining redesign and refresh schedules, where NHTSA included the redesign and refresh dates in the market forecast file as provided by manufacturers in confidential product plans. The new approach is necessary [[Page 49661]] given the nature of the new baseline which as a single year of data does not contain its own refresh and redesign cycle cues for future model years, and to ensure the complete transparency of the agency's analysis. Vehicle redesign/refresh assumptions are discussed in more detail in Section V of the PRIA and in Chapter 3 of the TSD. NHTSA seeks comment on its application for this proposal of refresh and redesign schedules to manufacturers' vehicles counting from the last known redesign in or prior to the baseline fleet, as compared to its approach in the MY 2011 final rule. Once the model has concluded that a technology should be applied to a vehicle, the model must evaluate which technology should be applied. This will depend on the vehicle subclass to which the vehicle is assigned; what technologies have already been applied to the vehicle (i.e., where in the ``decision tree'' the vehicle is); when the technology is first available (i.e., year of availability); whether the technology is still available (i.e., ``phase-in caps''); and the costs and effectiveness of the technologies being considered. Technology costs may be reduced, in turn, by learning effects, while technology effectiveness may be increased or reduced by synergistic effects between technologies. In the technology input file, NHTSA has developed a separate set of technology data variables for each of the twelve vehicle subclasses. Each set of variables is referred to as an ``input sheet,'' so for example, the subcompact input sheet holds the technology data that is appropriate for the subcompact subclass. Each input sheet contains a list of technologies available for members of the particular vehicle subclass. The following items are provided for each technology: the name of the technology, its abbreviation, the decision tree with which it is associated, the (first) year in which it is available, the upper and lower cost and effectiveness (fuel consumption reduction) estimates, the learning type and rate, the cost basis, its applicability, and the phase-in values. To which vehicle subclass is the vehicle assigned? As part of its consideration of technological feasibility, the agency evaluates whether each technology could be implemented on all types and sizes of vehicles, and whether some differentiation is necessary in applying certain technologies to certain types and sizes of vehicles, and with respect to the cost incurred and fuel consumption and CO2 emissions reduction achieved when doing so. The 2002 NAS Report differentiated technology application using ten vehicle ``classes'' (4 cars classes and 6 truck classes),\472\ but did not determine how cost and effectiveness values differ from class to class. NAS's purpose in separating vehicles into these classes was to create groups of ``like'' vehicles, i.e., vehicles similar in size, powertrain configuration, weight, and consumer use, and for which similar technologies are applicable. NHTSA similarly differentiates vehicles by ``subclass'' for the purpose of applying technologies to vehicles and assessing their incremental costs and effectiveness. NHTSA assigns each vehicle manufactured in the rulemaking period to one of 12 subclasses: for passenger cars, Subcompact, Subcompact Performance, Compact, Compact Performance, Midsize, Midsize Performance, Large, and Large Performance; and for light trucks, Small SUV/Pickup/Van, Midsize SUV/ Pickup/Van, Large SUV/Pickup/Van, and Minivan. --------------------------------------------------------------------------- \472\ The NAS classes included subcompact cars, compact cars, midsize cars, large cars, small SUVs, midsize SUVs, large SUVs, small pickups, large pickups, and minivans. --------------------------------------------------------------------------- For this NPRM as for the MY 2011 final rule, NHTSA divides the vehicle fleet into subclasses based on model inputs, and applies subclass-specific estimates, also from model inputs, of the applicability, cost, and effectiveness of each fuel-saving technology. Therefore, the model's estimates of the cost to improve the fuel economy of each vehicle model depend upon the subclass to which the vehicle model is assigned. Each vehicle's subclass is stored in the market forecast file. When conducting a compliance analysis, if the Volpe model seeks to apply technology to a particular vehicle, it checks the market forecast to see if the technology is available and if the refresh/redesign criteria are met. If these conditions are satisfied, the model determines the vehicle's subclass from the market data file, which it then uses to reference another input called the technology input file. NHTSA reviewed its methodology for dividing vehicles into subclasses for purposes of technology application that it used in the MY 2011 final rule, and concluded that the same methodology would be appropriate for this NPRM for MYs 2012-2016, but the agency invites comments on the method of assigning vehicles to subclasses for the purposes of technology application in the CAFE model, and on the issue of technology-application subclasses generally. The subclasses and the methodology for dividing vehicles among them are discussed in more detail in Section V of the PRIA and in Chapter 3 of the TSD. For the reader's reference, the subclasses and example vehicles from the market forecast file are provided in the tables below. Passenger Car Subclasses Example (MY 2008) Vehicles ------------------------------------------------------------------------ Class Example vehicles ------------------------------------------------------------------------ Subcompact............................. Chevy Aveo, Honda Civic. Subcompact Performance................. Mazda Miata, Saturn Sky. Compact................................ Chevy Cobalt, Nissan Sentra and Altima. Compact Performance.................... Audi S4 Quattro, Mazda RX8. Midsize................................ Chevy Camaro (V6), Toyota Camry, Honda Accord, Hyundai Azera. Midsize Performance.................... Chevy Corvette, Ford Mustang (V8), Nissan G37 Coupe. Large.................................. Audi A8, Cadillac CTS and DTS. Large Performance...................... Bentley Arnage, Daimler CL600. ------------------------------------------------------------------------ Light Truck Subclasses Example (MY 2008) Vehicles ------------------------------------------------------------------------ Class Example vehicles ------------------------------------------------------------------------ Minivans............................... Dodge Caravan, Toyota Sienna. [[Page 49662]] Small SUV/Pickup/Van................... Ford Escape & Ranger, Nissan Rogue. Midsize SUV/Pickup/Van................. Chevy Colorado, Jeep Wrangler 4- door, Volvo XC70, Toyota Tacoma. Large SUV/Pickup/Van................... Chevy Silverado, Ford Econoline, Toyota Sequoia. ------------------------------------------------------------------------ What technologies have already been applied to the vehicle (i.e., where in the ``decision trees'' is it)? NHTSA's methodology for technology application analysis developed out of the approach taken by NAS in the 2002 Report, and evaluates the application of individual technologies and their incremental costs and effectiveness. Incremental costs and effectiveness of individual technologies are relative to the prior technology state, which means that it is crucial to understand what technologies are already present on a vehicle in order to determine correct incremental cost and effectiveness values. The benefit of the incremental approach is transparency in accounting, insofar as when individual technologies are added incrementally to individual vehicles, it is clear and easy to determine how costs and effectiveness adds up as technology levels increase. To keep track of incremental costs and effectiveness and to know which technology to apply and in which order, the Volpe model's architecture uses a logical sequence, which NHTSA refers to as ``decision trees,'' for applying fuel economy-improving technologies to individual vehicles. In the MY 2011 final rule, NHTSA worked with Ricardo to modify previously-employed decision trees in order to allow for a much more accurate application of technologies to vehicles. For purposes of the NPRM, NHTSA reviewed the technology sequencing architecture and updated, as appropriate, the decision trees used in the analysis reported in the final rule for MY 2011. In general, and as described in great detail in the MY 2011 final rule and in Section V of the current PRIA, each technology is assigned to one of the five following categories based on the system it affects or impacts: engine, transmission, electrification/accessory, hybrid or vehicle. Each of these categories has its own decision tree that the Volpe model uses to apply technologies sequentially during the compliance analysis. The decision trees were designed and configured to allow the Volpe model to apply technologies in a cost-effective, logical order that also considers ease of implementation. For example, software or control logic changes are implemented before replacing a component or system with a completely redesigned one, which is typically a much more expensive option. In some cases, and as appropriate, the model may combine the sequential technologies shown on a decision tree and apply them simultaneously, effectively developing dynamic technology packages on an as-needed basis. For example, if compliance demands indicate, the model may elect to apply LUB, EFR, and ICP on a dual overhead cam engine, if they are not already present, in one single step. An example simplified decision tree for engine technologies is provided below; the other simplified decision trees may be found in Chapter 3 of the joint TSD and in the PRIA. Expanded decision trees are available in the docket for this NPRM. BILLING CODE 6560-50-C [[Page 49663]] [GRAPHIC] [TIFF OMITTED] TP28SE09.032 BILLING CODE 6560-50-C Each technology within the decision trees has an incremental cost and an incremental effectiveness estimate associated with it, and estimates are specific to a particular vehicle subclass (see the tables in Section V of the PRIA). Each technology's incremental estimate takes into account its position in the decision tree path. If a technology is located further down the decision tree, the estimates for the costs and effectiveness values attributed to that technology are influenced by the incremental estimates of costs and effectiveness values for prior technology applications. In essence, this approach accounts for ``in- path'' effectiveness synergies, as well as cost effects that occur between the technologies in the same path. When comparing cost and effectiveness estimates from various sources and those provided by commenters in the previous CAFE [[Page 49664]] rulemakings, it is important that the estimates evaluated are analyzed in the proper context, especially as concerns their likely position in the decision trees and other technologies that may be present or missing. Not all estimates available in the public domain or offered for the agencies' consideration during the comment period can be evaluated in an ``apples-to-apples'' comparison with those used by the Volpe model, since in some cases the order of application, or included technology content, is inconsistent with that assumed in the decision tree. The MY 2011 final rule discussed in detail the revisions and improvements made to the Volpe model and decision trees during that rulemaking process, including the improved handling and accuracy of valve train technology application and the development and implementation of a method for accounting path-dependent correction factors in order to ensure that technologies are evaluated within the proper context. The reader should consult the MY 2011 final rule documents for further information on these modeling techniques, all of which continued to be utilized in developing this proposal.\473\ To the extent that the decision trees have changed for purposes of this NPRM, it was due not to revisions in the order of technology application, but rather to redefinitions of technologies or addition or subtraction of technologies. NHTSA seeks comment on the decision trees described here and in the PRIA. --------------------------------------------------------------------------- \473\ See, e.g., 74 FR 14238-46 (Mar. 30, 2009) for a full discussion of the decision trees in NHTSA's MY 2011 final rule, and Docket No. NHTSA-2009-0062-0003.1 for an expanded decision tree used in that rulemaking. --------------------------------------------------------------------------- Is the next technology available in this model year? As discussed above, the majority of technologies considered are available on vehicles today, and thus will be available for application in the rulemaking time frame. Some technologies, however, will not become available for purposes of NHTSA's analysis until later in the rulemaking time frame. When the model is considering whether to add a technology to a vehicle, it checks its year of availability--if the technology is available, it may be added; if it is not available, the model will consider whether to switch to a different decision tree to look for another technology, or will skip to the next vehicle in a manufacturer's fleet. The year of availability for each technology is provided above in Table IV.C.2-1. Has the technology reached the phase-in cap for this model year? Besides the refresh/redesign cycles used in the Volpe model, which constrain the rate of technology application at the vehicle level so as to ensure a period of stability following any modeled technology applications, the other constraint on technology application employed in NHTSA's analysis is ``phase-in caps.'' Unlike vehicle-level cycle settings, phase-in caps constrain technology application at the vehicle manufacturer level.\474\ They are intended to reflect a manufacturer's overall resource capacity available for implementing new technologies (such as engineering and development personnel and financial resources), thereby ensuring that resource capacity is accounted for in the modeling process. At a high level, phase-in caps and refresh/ redesign cycles work in conjunction with one another to avoid the modeling process out-pacing an OEM's limited pool of available resources during the rulemaking time frame, especially in years where many models may be scheduled for refresh or redesign. This helps to ensure technological feasibility and economic practicability in determining the stringency of the standards. --------------------------------------------------------------------------- \474\ While phase-in caps are expressed as specific percentages of a manufacturer's fleet to which a technology may be applied in a given model year, phase-in caps cannot always be applied as precise limits, and the Volpe model in fact allows ``override'' of a cap in certain circumstances. When only a small portion of a phase-in cap limit remains, or when the cap is set to a very low value, or when a manufacturer has a very limited product line, the cap might prevent the technology from being applied at all since any application would cause the cap to be exceeded. Therefore, the Volpe model evaluates and enforces each phase-in cap constraint after it has been exceeded by the application of the technology (as opposed to evaluating it before application), which can result in the described overriding of the cap. --------------------------------------------------------------------------- NHTSA has been developing the concept of phase-in caps over the course of the last several CAFE rulemakings, as discussed in greater detail in the MY 2011 final rule,\475\ and in Section V of the PRIA and Chapter 3 of the joint TSD. The MY 2011 final rule employed non-linear phase-in caps (that is, caps that varied from year to year) that were designed to respond to comments raising lead-time concerns in reference to the agency's proposed MY 2011-2015 standards, but because the final rule covered only one model year, many phase-in caps for that model year were lower than had originally been proposed. NHTSA emphasized that the MY 2011 phase-in caps were based on assumptions for the full five year period of the proposal (2011-2015), and stated that it would reconsider the phase-in settings for all years beyond 2011 in a future rulemaking analysis. --------------------------------------------------------------------------- \475\ 74 FR 14268-14271 (Mar. 30, 2009). --------------------------------------------------------------------------- For purposes of the current proposal for MYs 2012-2016, as in the MY 2011 final rule, NHTSA combines phase-in caps for some groups of similar technologies, such as valve phasing technologies that are applicable to different forms of engine design (SOHC, DOHC, OHV), since they are very similar from an engineering and implementation standpoint. When the phase-in caps for two technologies are combined, the maximum total application of either or both to any manufacturer's fleet is limited to the value of the cap.\476\ In contrast to the phase-in caps used in the MY 2011 final rule, NHTSA has increased the phase-in caps for most of the technologies, as discussed below. --------------------------------------------------------------------------- \476\ See 74 FR 14270 (Mar 30, 2009) for further discussion and examples. --------------------------------------------------------------------------- In developing phase-in cap values for purposes of the current proposal, NHTSA initially considered the fact that many of the technologies commonly applied by the model, those placed near the top of the decision trees, such as low friction lubes, valve phasing, electric power steering, improved automatic transmission controls, and others, have been commonly available to manufacturers for several years now. Many technologies, in fact, precede the 2002 NAS Report, which estimated that such technologies would take 4 to 8 years to penetrate the fleet. Since the current proposal would take effect in MY 2012, nearly 10 years beyond the NAS report, and extends to MY 2016, and in the interest of harmonization with EPA's proposal, NHTSA tentatively determined that higher phase-in caps were likely justified. Additionally, NHTSA considered the fact that manufacturers, as part of the agreements supporting the National Program, appear to be anticipating higher technology application rates than those used in the MY 2011 final rule. This also supported higher phase-in caps for purposes of the proposal. Thus, while phase-in caps for the MY 2011 final rule reached a maximum of 50 percent for a couple of technologies and generally fell in the range between 0 and 20 percent, phase-in caps for this NPRM for the majority of technologies are set to reach 85 or 100 percent by MY 2016, although more advanced technologies like diesels and strong hybrids reach only 15 percent by MY 2016. Theoretically, significantly higher phase-in caps, such as those used in the current proposal as compared to those used in the MY 2011 final rule, should [[Page 49665]] result in higher levels of technology penetration in the modeling results. Reviewing the modeling output does not, however, indicate unreasonable levels of technology penetration for the proposed standards.\477\ NHTSA believes that this is due to the interaction of the various changes in methodology for the current proposal--changes to phase-in caps are but one of a number of revisions to the Volpe model and its inputs that could potentially impact the rate at which technologies are applied in this proposal as compared to prior rulemakings. Other revisions that could impact application rates include the use of transparent CAFE certification data in baseline fleet formulation and the use of other data for projecting it forward,\478\ or the use of a multi-year planning programming technique to apply technology retroactively to earlier-MY vehicles, both of which may have a direct impact on the modeling process. Conversely the model and inputs remain unchanged in other areas that also could impact technology application, such as in the refresh/redesign cycle settings, estimates used for the technologies, both of which remain largely unchanged from the MY 2011 final rule. These changes together make it difficult to predict how phase-in caps should be expected to function in the new modeling process. --------------------------------------------------------------------------- \477\ The modeling output for the analysis underlying these proposed standards is available on NHTSA's Web site. \478\ The baseline fleet sets the starting point, from a technology point of view, for where the model begins the technology application process, so changes have a direct impact on the net application of technology. --------------------------------------------------------------------------- Thus, after reviewing the output files, NHTSA tentatively concludes that the higher phase-in caps, and the resulting technology application rates produced by the Volpe model, at both the industry and manufacturer level, are appropriate for this proposal, achieving a suitable level of stringency without requiring unrealistic or unachievable penetration rates. However, the agency will consider comments received on this approach in determining what phase-in caps to employ in the analysis for the final rule, and may change the caps in response to comments and/or further analysis. One additional question the agency has, which may be primarily academic at this point, is what impact lower phase-in caps, such as those used in earlier rulemakings, would have on compliance costs (and whether they might counter- intuitively increase costs by forcing more expensive technologies). NHTSA seeks comment on the revised phase-in caps as compared to the MY 2011 final rule, and particularly on whether, combined with the refresh and redesign assumptions, they help to ensure sufficient lead time for manufacturers to make the technology changes required by the proposed standards. Readers are invited to review and assess the phase-in caps listed and described more fully in Section V of the PRIA, along with the application and penetration rates found in the Volpe model's output files, and after making their own assessment, provide comment and recommendations to the agency as appropriate. Is the technology less expensive due to learning effects? Historically, NHTSA did not explicitly account for the cost reductions a manufacturer might realize through learning achieved from experience in actually applying a technology. Since working with EPA to develop the 2008 NPRM for MYs 2011-2015, and with Ricardo to refine the concept for the March 2009 MY 2011 final rule, NHTSA has accounted for these cost reductions through two kinds of mutually exclusive learning, ``volume-based'' and ``time-based'' which it continues to use in this proposal, as discussed below. In the 2008 NPRM, NHTSA applied learning factors to technology costs for the first time. These learning factors were developed using the parameters of learning threshold, learning rate, and the initial cost, and were based on the ``experience curve'' concept which describes reductions in production costs as a function of accumulated production volume. The typical curve shows a relatively steep initial decline in cost which flattens out to a gentle downwardly sloping line as the volume increase to large values. In the NPRM, NHTSA applied a learning rate discount of 20 percent for each successive doubling of production volume (on a per manufacturer basis), and a learning threshold of 25,000 units was assumed (thus a technology was viewed as being fully learned out at 100,000 units). The factor was only applied to certain technologies that were considered emerging or newly implemented on the basis that significant cost improvements would be achieved as economies of scale were realized (i.e., the technologies were on the steep part of the curve). In the MY 2011 final rule, NHTSA continued to use this learning factor, referring to it as volume-based learning since the cost reductions were determined by production volume increases, and again only applied it to emerging technologies. However, and in response to comments, NHTSA revised its assumptions on learning threshold, basing them instead on an industry-wide production basis, and increasing the threshold to 300,000 units annually. Commenters to the 2008 NPRM also described another type of learning factor which NHTSA decided to adopt and implement in the MY 2011 final rule. Commenters described a relatively small negotiated cost decrease that occurred on an annual basis through contractual agreements with first tier component and systems suppliers for readily available, high volume technologies commonly in use by multiple OEMs. Based on the same experience curve principal, however at production volumes that were on the flatter part of the curve (and thus the types of volumes that represent annual industry volumes), NHTSA adopted this type learning and referred to it as time-based learning. An annual cost reduction of 3 percent in the second and each subsequent year, which was consistent with estimates from commenters and supported by work Ricardo conducted for NHTSA, was used in the final rule. In developing this proposal, NHTSA and EPA have reviewed both types of learning factors, and the thresholds (300,000) and reduction rates (20 percent for volume,\479\ 3 percent for time-based) they rely on, and as implemented in the MY 2011 final rule, and agreed that both factors continue to be accurate and appropriate; each agency has thus implemented time- and volume-based learning in their analyses. Noting that only one type of learning can be applied to any single technology, if any learning is applied at all, the agencies reviewed each to determine which learning factor was appropriate. Volume-based learning is applied to the higher complexity hybrid technologies, while no learning is applied to technologies likely to be affected by commodity costs (LUB, ROLL) or that have loosely-defined BOMs (EFR, LDB), as was the case in the MY 2011 final rule. Chapter 3 of the joint TSD shows the specific learning factors that NHTSA has applied in this analysis for each technology, and discusses learning factors and each agencies' use of them further. NHTSA seeks comment on its use of learning factors, including the types, the thresholds, and the reduction rates proposed, and particularly on the revisions to the learning (time- and volume-based) logic as compared to the MY 2011 final rule. --------------------------------------------------------------------------- \479\ NHTSA will conduct a sensitivity analysis on the volume- based learning value of 20 percent for the final rule. --------------------------------------------------------------------------- Is the technology more or less effective due to synergistic effects? When two or more technologies are added to a particular vehicle model to [[Page 49666]] improve its fuel efficiency and reduce CO2 emissions, the resultant fuel consumption reduction may sometimes be higher or lower than the product of the individual effectiveness values for those items.\480\ This may occur because one or more technologies applied to the same vehicle partially address the same source (or sources) of engine, drivetrain or vehicle losses. Alternately, this effect may be seen when one technology shifts the engine operating points, and therefore increases or reduces the fuel consumption reduction achieved by another technology or set of technologies. The difference between the observed fuel consumption reduction associated with a set of technologies and the product of the individual effectiveness values in that set is referred to for purposes of this rulemaking as a ``synergy.'' Synergies may be positive (increased fuel consumption reduction compared to the product of the individual effects) or negative (decreased fuel consumption reduction). An example of a positive synergy might be a vehicle technology that reduces road loads at highway speeds (e.g., lower aerodynamic drag or low rolling resistance tires), that could extend the vehicle operating range over which cylinder deactivation may be employed. An example of a negative synergy might be a variable valvetrain system technology, which reduces pumping losses by altering the profile of the engine speed/load map, and a six-speed automatic transmission, which shifts the engine operating points to a portion of the engine speed/load map where pumping losses are less significant. As the complexity of the technology combinations is increased, and the number of interacting technologies grows accordingly, it becomes increasingly important to account for these synergies. --------------------------------------------------------------------------- \480\ More specifically, the products of the differences between one and the technology-specific levels of effectiveness in reducing fuel consumption. For example, not accounting for interactions, if technologies A and B are estimated to reduce fuel consumption by 10% (i.e., 0.1) and 20% (i.e., 0.2) respectively, the ``product of the individual effectiveness values'' would be 1-0.1 times 1-0.2, or 0.9 times 0.8, which equals 0.72, corresponding to a combined effectiveness of 28% rather than the 30% obtained by adding 10% to 20%. The ``synergy factors'' discussed in this section further adjust these multiplicatively combined effectiveness values. --------------------------------------------------------------------------- NHTSA and EPA determined synergistic impacts for this rulemaking using EPA's ``lumped parameter'' analysis tool, which EPA described at length in its March 2008 Staff Technical Report.\481\ The lumped parameter tool is a spreadsheet model that represents energy consumption in terms of average performance over the fuel economy test procedure, rather than explicitly analyzing specific drive cycles. The tool begins with an apportionment of fuel consumption across several loss mechanisms and accounts for the average extent to which different technologies affect these loss mechanisms using estimates of engine, drivetrain and vehicle characteristics that are averaged over the EPA fuel economy drive cycle. Results of this analysis were generally consistent with those of full-scale vehicle simulation modeling performed in 2007 by Ricardo, Inc. --------------------------------------------------------------------------- \481\ EPA Staff Technical Report: Cost and Effectiveness Estimates of Technologies Used to Reduce Light-duty Vehicle Carbon Dioxide Emissions; EPA420-R-08-008, March 2008. --------------------------------------------------------------------------- For the current rulemaking, NHTSA used the lumped parameter tool as modified in the MY 2011 CAFE final rule. NHTSA modified the lumped parameter tool from the version described in the EPA Staff Technical Report in response to public comments received in its rulemaking. The modifications included updating the list of technologies and their associated effectiveness values to match the updated list of technologies used in the final rule. NHTSA also expanded the list of synergy pairings based on further consideration of the technologies for which a competition for losses would be expected. These losses are described in more detail in Section V of the PRIA. NHTSA and EPA incorporate synergistic impacts in their analyses in slightly different manners. Because NHTSA applies technologies individually in its modeling analysis, NHTSA incorporates synergistic effects between pairings of individual technologies. The use of discrete technology pair incremental synergies is similar to that in DOE's National Energy Modeling System (NEMS).\482\ Inputs to the Volpe model incorporate NEMS-identified pairs, as well as additional pairs from the set of technologies considered in the Volpe model. --------------------------------------------------------------------------- \482\ U.S. Department of Energy, Energy Information Administration, Transportation Sector Module of the National Energy Modeling System: Model Documentation 2007, May 2007, Washington, DC, DOE/EIAM070(2007), at 29-30. Available at http://tonto.eia.doe.gov/ ftproot/modeldoc/m070(2007) (last accessed Jul. 6, 2009). --------------------------------------------------------------------------- NHTSA notes that synergies that occur within a decision tree are already addressed within the incremental values assigned and therefore do not require a synergy pair to address. For example, all engine technologies take into account incremental synergy factors of preceding engine technologies, and all transmission technologies take into account incremental synergy factors of preceding transmission technologies. These factors are expressed in the fuel consumption improvement factors in the input files used by the Volpe model. For applying incremental synergy factors in separate path technologies, the Volpe model uses an input table (see the tables in Chapter 3 of the TSD and in the PRIA) which lists technology pairings and incremental synergy factors associated with those pairings, most of which are between engine technologies and transmission/electrification/ hybrid technologies. When a technology is applied to a vehicle by the Volpe model, all instances of that technology in the incremental synergy table which match technologies already applied to the vehicle (either pre-existing or previously applied by the Volpe model) are summed and applied to the fuel consumption improvement factor of the technology being applied. Synergies for the strong hybrid technology fuel consumption reductions are included in the incremental value for the specific hybrid technology block since the model applies technologies in the order of the most effectiveness for least cost and also applies all available electrification and transmission technologies before applying strong hybrid technologies. NHTSA seeks comment on whether the synergistic effects presented are accurate, and whether there are other synergies that the agency may have overlooked. d. Where Can Readers Find More Detailed Information About NHTSA's Technology Analysis? Much more detailed information is provided in Section V of the PRIA, and a discussion of how NHTSA and EPA jointly reviewed and updated technology assumptions for purposes of this NPRM is available in Chapter 3 of the TSD. Additionally, all of NHTSA's model input and output files are now public and available for the reader's review and consideration. The technology input files can be found in the docket for this NPRM, Docket No. NHTSA-2009-0059, and on NHTSA's Web site. And finally, because much of NHTSA's technology analysis for purposes of this NPRM builds on the work that was done for the MY 2011 final rule, we refer readers to that document as well for background information concerning how NHTSA's methodology for technology application analysis has evolved over the past several rulemakings, both in response to comments and as a result of the agency's [[Page 49667]] growing experience with this type of analysis.\483\ --------------------------------------------------------------------------- \483\ 74 FR 14233-308 (Mar. 30, 2009). --------------------------------------------------------------------------- 3. How Did NHTSA Develop the Economic Assumption Inputs? NHTSA's preliminary analysis of alternative CAFE standards for the model years covered by this proposed rulemaking relies on a range of forecast variables, economic assumptions, and parameter values. This section describes the proposed sources of these forecasts, the rationale underlying each assumption, and the agency's preliminary choices of specific parameter values. These proposed economic values play a significant role in determining the benefits of alternative CAFE standards, as they have for the last several CAFE rulemakings. Under those alternatives where standards would be established by reference to their costs and benefits, these economic values also affect the levels of the CAFE standards themselves. Some of these variables have more important effects on the level of CAFE standards and the benefits from requiring alternative increases in fuel economy than do others. In reviewing these variables and the agency's estimates of their values for purposes of this NPRM, NHTSA reconsidered previous comments it had received and reviewed newly available literature. As a consequence, the agency elected to revise some of its economic assumptions and parameter estimates, while retaining others. Some of the most important changes, which are discussed in greater detail below, as well as in Chapter 4 of the joint TSD and in Chapter VIII of the PRIA, include significant revisions to the markup factors for technology costs; reducing the rebound effect from 15 to 10 percent; and revising the value of reducing CO2 emissions based on recent interagency efforts to develop estimates of this value for government-wide use. For the reader's reference, Table IV.C.3-1 below summarizes the values used to calculate the economic benefits from each alternative. The agency seeks comment on the economic assumptions presented in the table and discussed below. Table IV.C.3-1--Economic Values for Benefits Computations (2007$) ------------------------------------------------------------------------ ------------------------------------------------------------------------ Fuel Economy Rebound Effect............................. 10% ``Gap'' between test and on-road MPG.................... 20% Value of refueling time per ($ per vehicle-hour)........ $ 24.64 Annual growth in average vehicle use.................... 1.1% Fuel Prices (2012-50 average, $/gallon) .............. Retail gasoline price............................... $3.77 Pre-tax gasoline price.............................. $3.40 Economic Benefits from Reducing Oil Imports ($/gallon) .............. ``Monopsony'' Component............................. $ 0.00 Price Shock Component............................... $ 0.17 Military Security Component......................... $ 0.00 --------------- Total Economic Costs ($/gallon)................. $ 0.17 Emission Damage Costs (2020, $/ton or $/metric ton) .............. Carbon monoxide..................................... $ 0 Volatile organic compounds (VOC).................... $ 1,283 Nitrogen oxides (NOx)--vehicle use.................. $ 5,116 Nitrogen oxides (NOx)--fuel production and $ 5,339 distribution....................................... Particulate matter (PM2.5)--vehicle use............. $ 238,432 Particulate matter (PM2.5)--fuel production and $ 292,180 distribution....................................... Sulfur dioxide (SO2)................................ $ 30,896 Carbon dioxide (CO2)................................ $ 20 Annual Increase in CO2 Damage Cost.................. 3% External Costs from Additional Automobile Use ($/vehicle- .............. mile) Congestion.......................................... $ 0.054 Accidents........................................... $ 0.023 Noise............................................... $ 0.001 --------------- Total External Costs............................ $ 0.078 External Costs from Additional Light Truck Use ($/ .............. vehicle-mile) Congestion.......................................... $0.048 Accidents........................................... $0.026 Noise............................................... $0.001 --------------- Total External Costs............................ $0.075 Discount Rate Applied to Future Benefits................ 3% ------------------------------------------------------------------------ a. Costs of Fuel Economy-Improving Technologies We developed detailed estimates of the costs of applying fuel economy-improving technologies to vehicle models jointly with EPA for use in analyzing the impacts of alternative standards considered in this rulemaking. The estimates were based on those reported by the 2002 NAS Report analyzing costs for increasing fuel economy, but were modified for purposes of this analysis as a result of extensive consultations among engineers from NHTSA, EPA, and the Volpe Center. As part of this process, the agency also developed varying cost estimates for applying certain fuel economy technologies to vehicles of different sizes and body styles. We may adjust these cost estimates based on comments received to this NPRM. The technology cost estimates used in this analysis are intended to represent [[Page 49668]] manufacturers' direct costs for high-volume production of vehicles with these technologies and sufficient experience with their application so that all remaining cost reductions due to ``learning curve'' effects have been fully realized. However, NHTSA recognizes that manufacturers' actual costs for employing these technologies include additional outlays for accompanying design or engineering changes to models that use them, development and testing of prototype versions, recalibrating engine operating parameters, and integrating the technology with other attributes of the vehicle. Manufacturers' indirect costs for employing these technologies also include expenses for product development and integration, modifying assembly processes and training assembly workers to install them, increased expenses for operation and maintaining assembly lines, higher initial warranty costs for new technologies, any added expenses for selling and distributing vehicles that use these technologies, and manufacturer and dealer profit. In previous CAFE rulemakings and in NHTSA's safety rulemakings, the agency has accounted for these additional costs by using a Retail Price Equivalent (RPE) multiplier of 1.5. For purposes of this rulemaking, based on recent work by EPA, NHTSA has applied indirect cost multipliers ranging from 1.11 to 1.64 to the estimates of vehicle manufacturers' direct costs for producing or acquiring each technology to improve fuel economy.\484\ These multipliers vary with the complexity of each technology and the time frame over which costs are estimated. More complex technologies are associated with higher multipliers because of the larger increases in manufacturers' indirect costs for developing, producing (or procuring), and deploying these more complex technologies. The appropriate multipliers decline over time for technologies of all complexity levels, since increased familiarity and experience with their application is assumed to reduce manufacturers' indirect costs for employing them. NHTSA seeks comment regarding the new indirect cost multiplier approach to technology costs estimates. We note additionally that this issue will be addressed in the upcoming revised NAS report. --------------------------------------------------------------------------- \484\ NHTSA notes that in addition to the technology cost analysis employing this ``ICM'' approach, the PRIA contains a sensitivity analysis using a technology cost multiplier of 1.5. --------------------------------------------------------------------------- b. Potential Opportunity Costs of Improved Fuel Economy An important concern is whether achieving the fuel economy improvements required by alternative CAFE standards would require manufacturers to compromise the performance, carrying capacity, safety, or comfort of their vehicle models. To the extent that it does so, the resulting sacrifice in the value of these attributes to consumers represents an additional cost of achieving the required improvements in fuel economy. While exact dollar values of these attributes to consumers are difficult to infer, differences in vehicle purchase prices and buyers' choices among competing models that feature different combinations of these characteristics clearly demonstrate that changing vehicle attributes clearly affect the utility and economic value that vehicles provide to potential buyers.\485\ --------------------------------------------------------------------------- \485\ See, e.g., Kleit A.N., 1990. ``The Effect of Annual Changes in Automobile Fuel Economy Standards.'' Journal of Regulatory Economics 2: 151-172; Berry, Steven, James Levinsohn, and Ariel Pakes, 1995. ``Automobile Prices in Market Equilibrium,'' Econometrica 63(4): 841-940; McCarthy, Patrick S., 1996. ``Market Price and Income Elasticities of New Vehicle Demands.'' Review of Economics and Statistics 78: 543-547; and Goldberg, Pinelopi K., 1998. ``The Effects of the Corporate Average Fuel Efficiency Standards in the US,'' Journal of Industrial Economics 46(1): 1-33. --------------------------------------------------------------------------- NHTSA and EPA have approached this potential problem by developing cost estimates for fuel economy-improving technologies that include any additional manufacturing costs that would be necessary to maintain the originally planned levels of performance, comfort, carrying capacity, and safety of any light-duty vehicle model to which those technologies are applied. In doing so, the agencies followed the precedent established by the 2002 NAS Report, which estimated ``constant performance and utility'' costs for fuel economy technologies. NHTSA has used these as the basis for its continuing efforts to refine the technology costs it uses to analyze manufacturer's costs for complying with alternative passenger car and light truck CAFE standards for MYs 2012-2016. Although the agency has revised its estimates of manufacturers' costs for some technologies significantly for use in this rulemaking, these revised estimates are still intended to represent costs that would allow manufacturers to maintain the performance, carrying capacity, and utility of vehicle models while improving their fuel economy. Although we believe that our tentative cost estimates for fuel economy-improving technologies should be generally sufficient to prevent significant reductions in consumer welfare provided by vehicle models to which manufacturers apply those technologies, it is possible that they do not include adequate allowance for the necessary efforts by manufacturers to prevent sacrifices in these attributes on all vehicle models. If this is the case, the true economic costs of achieving higher fuel economy should include the opportunity costs to vehicle owners of any sacrifices in vehicles' performance, carrying capacity, and utility and the agency's estimated technology costs would underestimate the true economic costs of improving fuel economy. Recognizing this possibility, it may be preferable for NHTSA to estimate explicitly the changes in vehicle buyers' welfare from the combination of higher prices for new vehicle models, increases in their fuel economy, and any accompanying changes in vehicle attributes such as performance, passenger- and cargo-carrying capacity, or other dimensions of utility. The net change in buyer's welfare that results from the combination of these changes would provide a more accurate estimate of the true economic costs for improving fuel economy. The agency seeks comment on this or other possible ways to deal with this extremely important issue. c. The On-Road Fuel Economy ``Gap'' Actual fuel economy levels achieved by light-duty vehicles in on- road driving fall somewhat short of their levels measured under the laboratory-like test conditions used by EPA to establish its published fuel economy ratings for different models. In analyzing the fuel savings from alternative CAFE standards, NHTSA has previously adjusted the actual fuel economy performance of each light truck model downward from its rated value to reflect the expected size of this on-road fuel economy ``gap.'' On December 27, 2006, EPA adopted changes to its regulations on fuel economy labeling, which were intended to bring vehicles' rated fuel economy levels closer to their actual on-road fuel economy levels.\486\ --------------------------------------------------------------------------- \486\ 71 FR 77871 (Dec. 27, 2006). --------------------------------------------------------------------------- In its Final Rule, EPA estimated that actual on-road fuel economy for light-duty vehicles averages 20 percent lower than published fuel economy levels. For example, if the overall EPA fuel economy rating of a light truck is 20 mpg, the on-road fuel economy actually achieved by a typical driver of that vehicle is expected to be 16 mpg (20*.80). NHTSA employed EPA's revised estimate of this on-road fuel economy gap in its analysis of the fuel [[Page 49669]] savings resulting from alternative CAFE standards evaluated in the MY 2011 final rule. For purposes of this NPRM, NHTSA conducted additional analysis of this issue. The agency used data on the number of passenger cars and light trucks of each model year that were registered for use during calendar years 2000 through 2006, average fuel economy for passenger cars and light trucks produced during each model year, and estimates of average miles driven per year by cars and light trucks of different ages. These data were combined to develop estimates of the average fuel economy that the U.S. passenger car and light truck fleets would have achieved from 2000 through 2006 under test conditions. NHTSA compared these estimates to the Federal Highway Administration's (FHWA) published values of actual on-road fuel economy for passenger cars and light trucks during each of those years.\487\ FHWA's estimates of actual fuel economy for passenger cars averaged 22 percent lower than NHTSA's estimates of its fleet-wide average value under test conditions over this period, while FHWA's estimates for light trucks averaged 17 lower than NHTSA's estimates of average light truck fuel economy under test conditions. These results appear to confirm that the 20 percent on-road fuel economy discount or gap represents a reasonable estimate for use in evaluating the fuel savings likely to result from alternative CAFE standards for MY 2012-2016 vehicles. --------------------------------------------------------------------------- \487\ Federal Highway Administration, Highway Statistics, 2000 through 2006 editions, Table VM-1; see http://www.fhwa.dot.gov/ policy/ohpi/hss/hsspubs.cfm (last accessed July 27, 2009). --------------------------------------------------------------------------- d. Fuel Prices and the Value of Saving Fuel Projected future fuel prices are a critical input into the preliminary economic analysis of alternative CAFE standards, because they determine the value of fuel savings both to new vehicle buyers and to society. NHTSA relied on the most recent fuel price projections from the U.S. Energy Information Administration's (EIA) Annual Energy Outlook (AEO) for this analysis. Specifically, we used the AEO 2009 (April 2009 release) Reference Case forecasts of inflation-adjusted (constant-dollar) retail gasoline and diesel fuel prices, which represent the EIA's most up-to-date estimate of the most likely course of future prices for petroleum products.\488\ --------------------------------------------------------------------------- \488\ Energy Information Administration, Annual Energy Outlook 2009, Revised Updated Reference Case (April 2009), Table 12. Available at http://www.eia.doe.gov/oiaf/servicerpt/stimulus/excel/ aeostimtab_12.xls(last accessed July 26, 2009). EIA's Updated Reference Case reflects the effects of the American Reinvestment and Recovery Act of 2009, as well as the most recent revisions to the U.S. and global economic outlook. --------------------------------------------------------------------------- While NHTSA relied on the forecasts of fuel prices presented in AEO 2008 High Price Case in the MY 2011 final rule, we noted at the time that we were relying on that estimate primarily because volatility in the oil market appeared to have overtaken the Reference Case, and that we anticipated that the Reference Case forecast would be significantly higher in the next AEO. In fact, EIA's AEO 2009 Reference Case forecast projects higher retail fuel prices in most future years than those forecast in the High Price Case from AEO 2008. NHTSA is thus confident that the AEO 2009 Reference Case is an appropriate forecast for projected future fuel prices. Measured in constant 2007 dollars, the Reference Case forecast of retail gasoline prices during calendar year 2020 is $3.62 per gallon, rising gradually to $3.82 by the year 2030 (these values include Federal, State and local taxes). To obtain fuel price forecasts for the years 2031 through 2050, the agency assumes that retail fuel prices will continue to increase after 2030 at the average annual rates projected for 2020-2030 in the AEO 2009 Revised Reference Case.\489\ This assumption results in a projected retail price of gasoline that reaches $4.25 in 2007 dollars by the year 2050. --------------------------------------------------------------------------- \489\ This projection uses the rate of increase in fuel prices for 2020-2030 rather than that over the complete forecast period (2009-2030) because there is extreme volatility in the forecasts for the years 2009 through approximately 2020. Using the average rate of change over the complete 2009-2030 forecast period would result in projections of declining fuel prices after 2030. --------------------------------------------------------------------------- The value of fuel savings resulting from improved fuel economy to buyers of light-duty vehicles is determined by the retail price of fuel, which includes Federal, State, and any local taxes imposed on fuel sales. Total taxes on gasoline, including Federal, State, and local levies averaged $0.42 per gallon during 2006, while those levied on diesel averaged $0.50. Because fuel taxes represent transfers of resources from fuel buyers to government agencies, however, rather than real resources that are consumed in the process of supplying or using fuel, their value must be deducted from retail fuel prices to determine the value of fuel savings resulting from more stringent CAFE standards to the U.S. economy as a whole. NHTSA follows the assumptions used by EIA in AEO 2009 that State and local gasoline taxes will keep pace with inflation in nominal terms, and thus remain constant when expressed in constant 2007 dollars. In contrast, EIA assumes that Federal gasoline taxes will remain unchanged in nominal terms, and thus decline throughout the forecast period when expressed in constant 2007 dollars. These differing assumptions about the likely future behavior of Federal and State/local fuel taxes are consistent with recent historical experience, which reflects the fact that Federal as well as most State motor fuel taxes are specified on a cents-per-gallon basis, and typically require legislation to change. The projected value of total taxes is deducted from each future year's forecast of retail gasoline and diesel prices reported in AEO 2009 to determine the economic value of each gallon of fuel saved during that year as a result of improved fuel economy. Subtracting fuel taxes results in a projected value for saving gasoline of $3.22 per gallon during 2020, rising to $3.45 per gallon by the year 2030. EIA includes ``High Price Case'' and ``Low Price Case'' forecasts in each AEO, which reflect uncertainties regarding future levels of oil production and demand. These alternative scenarios project retail gasoline prices that range from a low of $2.02 to a high of $5.04 per gallon during 2020, and from $2.04 to $5.47 per gallon during 2030. In conjunction with our assumption that fuel taxes will remain constant in real or inflation-adjusted terms over this period, these forecasts imply pre-tax values of saving fuel ranging from $1.63 to $4.65 per gallon during 2020, and from $1.67 to $5.10 per gallon in 2030. In conducting the preliminary analysis of uncertainty in benefits and costs from alternative CAFE standards required by OMB, NHTSA evaluated the sensitivity of its benefits estimates to these alternative forecasts of future fuel prices. The results of this sensitivity analysis can be found in the PRIA. e. Consumer Valuation of Fuel Economy and Payback Period In estimating the value of fuel economy improvements that would result from alternative CAFE standards to potential vehicle buyers, NHTSA assumes, as in the MY 2011 final rule, that buyers value the resulting fuel savings over only part of the expected lifetime of the vehicles they purchase. Specifically, we assume that buyers value fuel savings over the first five years of a new vehicle's lifetime, and discount the value of these future fuel savings at a 3 percent annual rate. The five-year figure represents [[Page 49670]] approximately the current average term of consumer loans to finance the purchase of new vehicles. We recognize that the period over which individual buyers finance new vehicle purchases may not correspond exactly to the time horizons they apply in valuing fuel savings from higher fuel economy. The agency deducts the discounted present value of fuel savings over the first five years of a vehicle model's lifetime from the technology costs incurred by its manufacturer to improve that model's fuel economy to determine the increase in its ``effective price'' to buyers. The Volpe model uses these estimates of effective costs for increasing the fuel economy of each vehicle model to identify the order in which manufacturers would be likely to select models for the application of fuel economy-improving technologies in order to comply with stricter standards. The average value of the resulting increase in effective cost from each manufacturer's simulated compliance strategy is also used to estimate the impact of alternative standards on its total sales for future model years. However, it is important to recognize that NHTSA estimates the aggregate value to the U.S. economy of fuel savings resulting from alternative standards--or their ``social'' value--over the entire expected lifetimes of vehicles manufactured under those standards, rather than over this shorter ``payback period'' we assume for their buyers. The procedure the agency uses for doing so is discussed in detail in the following section. f. Vehicle Survival and Use Assumptions NHTSA's first step in estimating lifetime fuel consumption by vehicles produced during a model year is to calculate the number expected to remain in service during each year following their production and sale.\490\ This is calculated by multiplying the number of vehicles originally produced during a model year by the proportion typically expected to remain in service at their age during each later year, often referred to as a ``survival rate.'' --------------------------------------------------------------------------- \490\ Vehicles are defined to be of age 1 during the calendar year corresponding to the model year in which they are produced; thus for example, model year 2000 vehicles are considered to be of age 1 during calendar year 2000, age 1 during calendar year 2001, and to reach their maximum age of 26 years during calendar year 2025. NHTSA considers the maximum lifetime of vehicles to be the age after which less than 2 percent of the vehicles originally produced during a model year remain in service. Applying these conventions to vehicle registration data indicates that passenger cars have a maximum age of 26 years, while light trucks have a maximum lifetime of 36 years. See Lu, S., NHTSA, Regulatory Analysis and Evaluation Division, ``Vehicle Survivability and Travel Mileage Schedules,'' DOT HS 809 952, 8-11 (January 2006). Available at http://www- nrd.nhtsa.dot.gov/Pubs/809952.pdf (last accessed July 27, 2009). --------------------------------------------------------------------------- To estimate production volumes of passenger cars and light trucks for individual manufacturers, NHTSA relied on a baseline market forecast constructed by EPA staff beginning with MY 2008 CAFE certification data. After constructing a MY 2008 baseline, EPA used projected car and truck volumes for this period from Energy Information Administration's (EIA's) 2009 Annual Energy Outlook (AEO).\491\ However, AEO projects sales only at the car and truck level, not at the manufacturer and model-specific level, which are needed in order to estimate the effects new standards will have on individual manufacturers.\492\ Therefore, EPA purchased data from CSM-Worldwide and used their projections of the number of vehicles of each type predicted to be sold by manufacturers in 2011-2015.\493\ This provided the year-by-year percentages of cars and trucks sold by each manufacturer as well as the percentages of each vehicle segment. Although it was thus necessary to assume the same manufacturer and segment shares in 2016 as in 2015, 2016 estimates from CSM should be available for the final rule. Using these percentages normalized to the AEO projected volumes then provided the manufacturer-specific market share and model-specific sales for model years 2011-2016. --------------------------------------------------------------------------- \491\ Available at http://www.eia.doe.gov/oiaf/aeo/index.html. NHTSA and EPA made the simplifying assumption that projected sales of cars and light trucks during each calendar year from 2012 through 2016 represented the likely production volumes for the corresponding model year. The agency did not attempt to establish the exact correspondence between projected sales during individual calendar years and production volumes for specific model years. \492\ Because AEO 2009's ``car'' and ``truck'' classes did not reflect NHTSA's recent reclassification (in March 2009 for enforcement beginning MY 2011) of many two wheel drive SUVs from the nonpassenger (i.e., light truck) fleet to the passenger car fleet, EPA staff made adjustments to account for such vehicles in the baseline. \493\ EPA also considered other sources of similar information, such as J.D. Powers, and concluded that CSM was better able to provide forecasts at the requisite level of detail for most of the model years of interest. --------------------------------------------------------------------------- To estimate the number of passenger cars and light trucks originally produced during model years 2012 through 2016 that will remain in use during each subsequent year the agency applied age- specific survival rates for cars and light trucks to these adjusted forecasts of passenger car and light truck sales. In 2008, NHTSA updated its previous estimates of car and light truck survival rates using the most current registration data for vehicles produced during recent model years, in order to ensure that they reflected recent increases in the durability and expected life spans of cars and light trucks.\494\ --------------------------------------------------------------------------- \494\ Lu, S., NHTSA, Regulatory Analysis and Evaluation Division, ``Vehicle Survivability and Travel Mileage Schedules,'' DOT HS 809 952, 8-11 (January 2006). Available at http://www- nrd.nhtsa.dot.gov/Pubs/809952.pdf (last accessed August 9, 2009). These updated survival rates suggest that the expected lifetimes of recent-model passenger cars and light trucks are 13.8 and 14.5 years. --------------------------------------------------------------------------- The next step in estimating fuel use is to calculate the total number of miles that model year 2012-2016 cars and light trucks remaining in use will be driven each year. To estimate total miles driven, the number projected to remain in use during each future year is multiplied by the average number of miles they are expected to be driven at the age they will reach in that year. The agency estimated annual usage of cars and light trucks of each age using data from the Federal Highway Administration's 2001 National Household Transportation Survey (NHTS).\495\ Because these estimates reflect the historically low gasoline prices that prevailed at the time the 2001 NHTS was conducted, however, NHTSA adjusted them to account for the effect on vehicle use of subsequent increases in fuel prices. Details of this adjustment are provided in Chapter VIII of the PRIA and Chapter of the draft joint TSD. --------------------------------------------------------------------------- \495\ For a description of the Survey, see http://nhts.ornl.gov/ quickStart.shtml (last accessed August 9, 2009). --------------------------------------------------------------------------- Increases in average annual use of cars and light trucks have been an important source of historical growth in the total number of miles they are driven each year. To estimate future growth in their average annual use for purposes of this rulemaking, NHTSA calculated the rate of growth in the adjusted mileage schedules derived from the 2001 NHTS necessary for total car and light truck travel to increase at the rate forecast in the AEO 2009 Reference Case.\496\ This rate was calculated to be consistent with future changes in the overall size and age distributions of the U.S. passenger car and light truck fleets that result from the agency's forecasts of total car and light truck sales and updated survival rates. The resulting growth rate in average annual car and light truck use of approximately 1.1 percent per year was [[Page 49671]] applied to the mileage figures derived from the 2001 NHTS to estimate annual mileage during each year of the expected lifetimes of MY 2012- 2016 cars and light trucks.\497\ --------------------------------------------------------------------------- \496\ This approach differs from that used in the MY 2011 final rule, where it was assumed that future growth in the total number of cars and light trucks in use resulting from projected sales of new vehicles was adequate by itself to account for growth in total vehicle use, without assuming continuing growth in average vehicle use. \497\ While the adjustment for future fuel prices reduces average mileage at each age from the values derived from the 2001 NHTS, the adjustment for expected future growth in average vehicle use increases it. The net effect of these two adjustments is to increase expected lifetime mileage by about 18 percent significantly for both passenger cars and about 16 percent for light trucks. --------------------------------------------------------------------------- Finally, the agency estimated total fuel consumption by passenger cars and light trucks remaining in use each year by dividing the total number of miles surviving vehicles are driven by the fuel economy they are expected to achieve under each alternative CAFE standard. Each model year's total lifetime fuel consumption is the sum of fuel use by the cars or light trucks produced during that model year during each year of their life spans. In turn, the savings in a model year's lifetime fuel use that will result from each alternative CAFE standard is the difference between its lifetime fuel use at the fuel economy level it attains under the Baseline alternative, and its lifetime fuel use at the higher fuel economy level it is projected to achieve under that alternative standard.\498\ --------------------------------------------------------------------------- \498\ To illustrate these calculations, the agency's adjustment of the AEO 2009 Revised Reference Case forecast indicates that 9.26 million passenger cars will be produced during 2012, and the agency's updated survival rates show that 83 percent of these vehicles, or 7.64 million, are projected to remain in service during the year 2022, when they will have reached an age of 10 years. At that age, passenger achieving the fuel economy level they are projected to achieve under the Baseline alternative are driven an average of about 800 miles, so surviving model year 2012 passenger cars will be driven a total of 82.5 billion miles (= 7.64 million surviving vehicles x 10,800 miles per vehicle) during 2022. Summing the results of similar calculations for each year of their 26-year maximum lifetime, model year 2012 passenger cars will be driven a total of 1,395 billion miles under the Baseline alternative. Under that alternative, they are projected to achieve a test fuel economy level of 32.4 mpg, which corresponds to actual on-road fuel economy of 25.9 mpg (= 32.4 mpg x 80 percent). Thus their lifetime fuel use under the Baseline alternative is projected to be 53.9 billion gallons (= 1,395 billion miles divided by 25.9 miles per gallon). --------------------------------------------------------------------------- g. Accounting for the Rebound Effect of Higher Fuel Economy The fuel economy rebound effect refers to the fraction of fuel savings expected to result from an increase in vehicle fuel economy-- particularly an increase required by the adoption of higher CAFE standards--that is offset by additional vehicle use. The increase in vehicle use occurs because higher fuel economy reduces the fuel cost of driving, typically the largest single component of the monetary cost of operating a vehicle, and vehicle owners respond to this reduction in operating costs by driving slightly more. By lowering the marginal cost of vehicle use, improved fuel economy may lead to an increase in the number of miles vehicles are driven each year and over their lifetimes. Even with their higher fuel economy, this additional driving consumes some fuel, so the rebound effect reduces the net fuel savings that result when new CAFE standards require manufacturers to improve fuel economy. The magnitude of the rebound effect is an important determinant of the actual fuel savings that are likely to result from adopting stricter CAFE standards. Research on the magnitude of the rebound effect in light-duty vehicle use dates to the early 1980s, and generally concludes that a statistically significant rebound effect occurs when vehicle fuel efficiency improves.\499\ The agency reviewed studies of the rebound effect it had previously relied upon, considered more recently published estimates, and developed new estimates of its magnitude for purposes of this NPRM.\500\ Recent studies provide some evidence that the rebound effect has been declining over time, and may decline further over the immediate future if incomes rise faster than gasoline prices. This result appears plausible, because the responsiveness of vehicle use to variation in fuel costs is expected to decline as they account for a smaller proportion of the total monetary cost of driving, which has been the case until very recently. At the same time, rising personal incomes would be expected to reduce the sensitivity of vehicle use to fuel costs as the time component of driving costs--which is likely to be related to income levels--accounts for a larger fraction the total cost of automobile travel. NHTSA developed new estimates of the rebound effect by using national data on light-duty vehicle travel over the period from 1950 through 2006 to estimate various econometric models of the relationship between vehicle miles-traveled and factors likely to influence it, including household income, fuel prices, vehicle fuel efficiency, road supply, the number of vehicles in use, vehicle prices, and other factors.\501\ The results of NHTSA's analysis are consistent with the findings from other recent research: The average long-run rebound effect ranged from 16 percent to 30 percent over the period from 1950 through 2007, while estimates of the rebound effect in 2007 range from 8 percent to 14 percent. Projected values of the rebound effect for the period from 2010 through 2030, which the agency developed using forecasts of personal income, fuel prices, and fuel efficiency from AEO 2009's Reference Case, range from 4 percent to 16 percent, depending on the specific model used to generate them. --------------------------------------------------------------------------- \499\ Some studies estimate that the long-run rebound effect is significantly larger than the immediate response to increased fuel efficiency. Although their estimates of the adjustment period required for the rebound effect to reach its long-run magnitude vary, this long-run effect is most appropriate for evaluating the fuel savings and emissions reductions resulting from stricter standards that would apply to future model years. \500\ For details of the agency's analysis, see Chapter VIII of the PRIA and Chapter 4 of the draft joint TSD accompanying this proposed rule. \501\ The agency used several different model specifications and estimation procedures to control for the effect of fuel prices on fuel efficiency in order to obtain accurate estimates of the rebound effect. --------------------------------------------------------------------------- In light of these results, the agency's judgment is that the apparent decline over time in the magnitude of the rebound effect justifies using a value for future analysis that is lower than historical estimates, which average 15-25 percent. Because the lifetimes of vehicles affected by the alternative CAFE standards considered in this rulemaking will extend from 2012 until nearly 2050, a value that is significantly lower than historical estimates appears to be appropriate. Thus NHTSA has elected to use a 10 percent rebound effect in its analysis of fuel savings and other benefits from higher CAFE standards for this NPRM. NHTSA also invites comment on other alternatives for estimating the rebound effect. As one illustration, variation in the price per gallon of gasoline directly affects the per-mile cost of driving, and drivers may respond just as they would to a change in the cost of driving resulting from a change in fuel economy, by varying the number of miles they drive. Because vehicles' fuel economy is fixed in the short run, variation in the number of miles driven in response to changes in fuel prices will be reflected in changes in gasoline consumption. Under the assumption that drivers respond similarly to changes in the cost of driving whether they are caused by variation in fuel prices or fuel economy, the short-run price elasticity of gasoline--which measures the sensitivity of gasoline consumption to changes in its price per gallon--may provide some indication about the magnitude of the rebound effect itself. NHTSA invites comment on the extent to which the short- run elasticity of demand for gasoline with respect to its price can provide useful information about the size of the rebound effect. Specifically, we seek comment on whether it would be [[Page 49672]] appropriate to use the price elasticity of demand for gasoline, or other alternative approaches, to guide the choice of a value for the rebound effect. Additionally, NHTSA recognizes that as the world price of oil falls in response to lower U.S. demand for oil, there is the potential for an increase in oil use and, in turn, greenhouse gas emissions outside the U.S. This so called international oil ``take back'' effect is difficult to estimate. Given that oil consumption patterns vary across countries, there will be different demand responses to a change in the world price of crude oil. In addition, many countries around the world subsidize their oil consumption. It is not clear how oil consumption would change due to changes in the market price of oil given the current pattern of demand and subsidies. Further, many countries, especially in the developed countries/regions (i.e., the European Union), already have or anticipate implementing policies to limit GHG emissions. Further out in the future, it is anticipated that developing countries would take actions to reduce their GHG emissions as well. Any increases in petroleum consumption and GHG emissions in other nations that occurs in response to a decline in world petroleum prices would be attributed to those nations, and recorded in their respective GHG emissions inventories. Thus, including the same increase in emissions as part of the impact of adopting CAFE standards in the U.S. would risk double- counting of global emissions totals. NHTSA seeks comment on how to estimate the international ``take back'' effect and its impact on fuel consumption and GHG emissions. See the Energy Security section of the TSD, 4.2.8, for more discussion of the impact of the proposed vehicle rule on oil markets. h. Benefits From Increased Vehicle Use The increase in vehicle use from the rebound effect provides additional benefits to their owners, who may make more frequent trips or travel farther to reach more desirable destinations. This additional travel provides benefits to drivers and their passengers by improving their access to social and economic opportunities away from home. As evidenced by their decisions to make more frequent or longer trips when improved fuel economy reduces their costs for driving, the benefits from this additional travel exceed the costs drivers and passengers incur in making more frequent or longer trips. The agency's analysis estimates the economic benefits from increased rebound-effect driving as the sum of fuel costs drivers incur plus the consumer surplus they receive from the additional accessibility it provides.\502\ Because the increase in travel depends on the extent of improvement in fuel economy, the value of benefits it provides differs among model years and alternative CAFE standards. Under even those alternatives that would impose the highest standards, however, the magnitude of these benefits represents a small fraction of total benefits. --------------------------------------------------------------------------- \502\ The consumer surplus provided by added travel is estimated as one-half of the product of the decline in fuel cost per mile and the resulting increase in the annual number of miles driven. --------------------------------------------------------------------------- i. The Value of Increased Driving Range Improving vehicles' fuel economy may also increase their driving range before they require refueling. By reducing the frequency with which drivers typically refuel, and by extending the upper limit of the range they can travel before requiring refueling, improving fuel economy thus provides some additional benefits to their owners.\503\ NHTSA re-examined this issue for purposes of this rulemaking, and found no information in comments or elsewhere that would cause the agency to revise its previous approach. Since no direct estimates of the value of extended vehicle range are available, NHTSA calculates directly the reduction in the annual number of required refueling cycles that results from improved fuel economy, and applies DOT-recommended values of travel time savings to convert the resulting time savings to their economic value.\504\ --------------------------------------------------------------------------- \503\ If manufacturers respond to improved fuel economy by reducing the size of fuel tanks to maintain a constant driving range, the resulting cost savings will presumably be reflected in lower vehicle sales prices. \504\ See Department of Transportation, Guidance Memorandum, ``The Value of Saving Travel Time: Departmental Guidance for Conducting Economic Evaluations,'' Apr. 9, 1997. http:// ostpxweb.dot.gov/policy/Data/VOT97guid.pdf (last accessed August 9, 2009); update available at http://ostpxweb.dot.gov/policy/Data/ VOTrevision1_2-11-03.pdf (last accessed August 9, 2009). --------------------------------------------------------------------------- As an illustration, a typical small light truck model has an average fuel tank size of approximately 20 gallons. Assuming that drivers typically refuel when their tanks are 55 percent full (i.e., 11 gallons in reserve), increasing this model's actual on-road fuel economy from 24 to 25 mpg would extend its driving range from 216 miles (= 9 gallons x 24 mpg) to 225 miles (= 9 gallons x 25 mpg). Assuming that it is driven 12,000 miles/year, this reduces the number of times it needs to be refueled each year from 55.6 (= 12,000 miles per year/ 216 miles per refueling) to 53.3 (= 12,000 miles per year/225 miles per refueling), or by 2.3 refuelings per year. Weighted by the nationwide mix of urban and rural driving, personal and business travel in urban and rural areas, and average vehicle occupancy for driving trips, the DOT-recommended values of travel time per vehicle-hour is $24.64 (in 2007 dollars).\505\ Assuming that locating a station and filling up requires five minutes, the annual value of time saved as a result of less frequent refueling amounts to $4.72 (calculated as 5/60 x 2.3 x $24.64). This calculation is repeated for each future year that model year 2012-2016 cars and light trucks would remain in service. Like fuel savings and other benefits, the value of this benefit declines over a model year's lifetime, because a smaller number of vehicles originally produced during that model year remain in service each year, and those remaining in service are driven fewer miles. --------------------------------------------------------------------------- \505\ The hourly wage rate during 2008 is estimated to average $25.50 when expressed in 2007 dollars. Personal travel in urban areas (which represents 94 percent of urban travel) is valued at 50 percent of the hourly wage rate, while business travel (the remaining 6 percent of urban travel) is valued at 100 percent of the hourly wage rate. For intercity travel, personal travel (87 percent of total intercity travel) is valued at 70 percent of the wage rate, while business travel (13 percent) is valued at 100 percent of the wage rate. The resulting values of travel time are $12.67 for urban travel and $17.66 for intercity travel, and must be multiplied by vehicle occupancy (1.6) to obtain the estimated values of time per vehicle hour in urban and rural driving. Finally, about 66% of driving occurs in urban areas, while the remaining 34% takes place in rural areas, and these percentages are used to calculate a weighted average of the value of time in all driving. --------------------------------------------------------------------------- NHTSA recognizes that many assumptions made in its estimate for the value of increased driving range are subject to uncertainty. Please see Chapter 4 of the TSD and Chapter 8 of NHTSA's PRIA for more information about the uncertainty regarding these assumptions. j. Added Costs From Congestion, Crashes and Noise Increased vehicle use associated with the rebound effect also contributes to increased traffic congestion, motor vehicle accidents, and highway noise. NHTSA relies on estimates of per-mile congestion, accident, and noise costs caused by increased use of automobiles and light trucks developed by the Federal Highway Administration to estimate these increased costs.\506\ NHTSA employed these estimates previously in its analysis accompanying the MY 2011 final rule, and continues [[Page 49673]] to find them appropriate for this NPRM after reviewing the procedures used by FHWA to develop them and considering other available estimates of these values. The agency multiplies FHWA's estimates of per-mile costs by the annual increases in automobile and light truck use from the rebound effect to yield the estimated increases in congestion, accident, and noise externality costs during each future year. --------------------------------------------------------------------------- \506\ These estimates were developed by FHWA for use in its 1997 Federal Highway Cost Allocation Study; see http://www.fhwa.dot.gov/ policy/hcas/final/index.htm (last accessed August 9, 2009). --------------------------------------------------------------------------- k. Petroleum Consumption and Import Externalities U.S. consumption and imports of petroleum products also impose costs on the domestic economy that are not reflected in the market price for crude petroleum, or in the prices paid by consumers of petroleum products such as gasoline. These costs include (1) higher prices for petroleum products resulting from the effect of U.S. oil import demand on the world oil price; (2) the risk of disruptions to the U.S. economy caused by sudden reductions in the supply of imported oil to the U.S.; and (3) expenses for maintaining a U.S. military presence to secure imported oil supplies from unstable regions, and for maintaining the strategic petroleum reserve (SPR) to cushion against resulting price increases.\507\ --------------------------------------------------------------------------- \507\ See, e.g., Bohi, Douglas R. and W. David Montgomery (1982). Oil Prices, Energy Security, and Import Policy, Washington, DC: Resources for the Future, Johns Hopkins University Press; Bohi, D. R., and M. A. Toman (1993). ``Energy and Security: Externalities and Policies,'' Energy Policy 21:1093-1109; and Toman, M. A. (1993). ``The Economics of Energy Security: Theory, Evidence, Policy,'' in A. V. Kneese and J. L. Sweeney, eds. (1993). Handbook of Natural Resource and Energy Economics, Vol. III. Amsterdam: North-Holland, pp. 1167-1218. --------------------------------------------------------------------------- Higher U.S. imports of crude oil or refined petroleum products increase the magnitude of these external economic costs, thus increasing the true economic cost of supplying transportation fuels above their market prices. Conversely, lowering U.S. imports of crude petroleum or refined fuels by reducing domestic fuel consumption can reduce these external costs, and any reduction in their total value that results from improved fuel economy represents an economic benefit of more stringent CAFE standards, in addition to the value of saving fuel itself. NHTSA has carefully reviewed its assumptions regarding the appropriate value of these benefits for this proposed rule. In analyzing benefits from its recent actions to increase light truck CAFE standards for model years 2005-07 and 2008-11, NHTSA relied on a 1997 study by Oak Ridge National Laboratory (ORNL) to estimate the value of reduced economic externalities from petroleum consumption and imports.\508\ More recently, ORNL updated its estimates of the value of these externalities, using the analytic framework developed in its original 1997 study in conjunction with recent estimates of the variables and parameters that determine their value.\509\ The updated ORNL study was subjected to a detailed peer review by experts selected by EPA, and its estimates of the value of oil import externalities were subsequently revised to reflect their comments and recommendations.\510\ --------------------------------------------------------------------------- \508\ Leiby, Paul N., Donald W. Jones, T. Randall Curlee, and Russell Lee, Oil Imports: An Assessment of Benefits and Costs, ORNL- 6851, Oak Ridge National Laboratory, November 1, 1997. Available at http://pzl1.ed.ornl.gov/ORNL6851.pdf (last accessed August 9, 2009). \509\ Leiby, Paul N. ``Estimating the Energy Security Benefits of Reduced U.S. Oil Imports,'' Oak Ridge National Laboratory, ORNL/ TM-2007/028, Revised July 23, 2007. Available at http:// pzl1.ed.ornl.gov/energysecurity.html (click on link below ``Oil Imports Costs and Benefits'') (last accessed August 9, 2009). \510\ Peer Review Report Summary: Estimating the Energy Security Benefits of Reduced U.S. Oil Imports, ICF, Inc., September 2007. --------------------------------------------------------------------------- At the request of EPA, ORNL further revised its 2008 estimates of external costs from U.S. oil imports to reflect recent changes in the outlook for world petroleum prices and continuing changes in the structure and characteristics of global petroleum supply and demand. These most recent revisions increase ORNL's estimates of the ``monopsony premium'' associated with U.S. oil imports, which measures the reduced value of payments from U.S. oil purchasers to foreign oil suppliers beyond the savings from reduced purchases of petroleum itself that results when lower U.S. import demand reduces the world price of petroleum.\511\ Consistency with NHTSA's use of estimates of the global benefits from reducing emissions of CO2 and other greenhouse gases in this analysis, however, requires the use of a global perspective for assessing their net value. From this perspective, reducing these payments simply results in a transfer of resources from foreign oil suppliers to U.S. purchasers (or more properly, in a savings in the value of resources previously transferred from U.S. purchasers to foreign producers), and provides no real savings in resources to the global economy. Thus NHTSA's analysis of the benefits from adopting higher CAFE standards for MY 2012-2016 cars and light trucks excludes the reduced value of monopsony payments by U.S. oil consumers that might result from lower fuel consumption by these vehicles. --------------------------------------------------------------------------- \511\ The reduction in payments from U.S. oil purchasers to domestic petroleum producers is not included as a benefit, since it represents a transfer that occurs entirely within the U.S. economy. --------------------------------------------------------------------------- The literature on the energy security for the last two decades has routinely combined the monopsony and the macroeconomic disruption components when calculating the total value of the energy security premium. However, in the context of using a global value for the Social Cost of Carbon (SCC) the question arises: How should the energy security premium be used when some benefits from the proposed rule, such as the benefits of reducing greenhouse gas emissions, are calculated at a global level? Monopsony benefits represent avoided payments by the U.S. to oil producers in foreign countries that result from a decrease in the world oil price as the U.S. decreases its consumption of imported oil. Although there is clearly a benefit to the U.S. when considered from the domestic perspective, the decrease in price due to decreased demand in the U.S. also represents a loss of income to oil-producing countries. Given the redistributive nature of this effect, do the negative effects on other countries ``net out'' the positive impacts to the U.S.? If this is the case, then, the monopsony portion of the energy security premium should be excluded from the net benefits calculation for the rule. As the preceding discussion has indicated, the agencies omitted the reduction in monopsony payments that occurs when U.S. petroleum consumption and imports are reduced from their estimates of economic benefits for the proposed rules. Since the reduction in monopsony payments by U.S. oil consumers is exactly offset by a decline in income to suppliers of imported oil, this omission ensures consistency of the agencies' analysis with the inclusion of global benefits from reducing emissions of greenhouse gas emissions. The agencies seek comment on whether, from other perspectives, it would be reasonable to include both the global value of reducing GHG emissions and the reduction in monopsony payments by U.S. consumers of petroleum products in their estimates of total economic benefits from reducing U.S. fuel consumption. ORNL's most recently revised estimates of the increase in the expected costs associated with potential disruptions in U.S. petroleum imports imply that each gallon of imported fuel or petroleum saved reduces the expected costs of oil supply disruptions [[Page 49674]] to the U.S. economy by $0.16 per gallon (in 2007$). The reduction in expected disruption costs represents a real savings in resources, and thus contributes economic benefits in addition to the savings in fuel production costs that result from increasing fuel economy. NHTSA employs this value in its evaluation of the economic benefits from adopting higher CAFE standards for MY 2012-2016 cars and light trucks. NHTSA's analysis does not include savings in budgetary outlays to support U.S. military activities among the benefits of higher fuel economy and the resulting fuel savings.\512\ NHTSA's analysis of benefits from alternative CAFE standards for MY 2012-2016 also excludes any cost savings from maintaining a smaller SPR from its estimates of the external benefits of reducing gasoline consumption and petroleum imports. This view concurs with that of the recent ORNL study of economic costs from U.S. oil imports, which concludes that savings in government outlays for these purposes are unlikely to result from reductions in consumption of petroleum products and oil imports on the scale of those resulting from higher CAFE standards. --------------------------------------------------------------------------- \512\ However, the agency conducted a sensitivity analysis of the potential effect of assuming that some reduction military spending would result from fuel savings and reduced petroleum imports in order to investigate its impacts on the standards and fuel savings. --------------------------------------------------------------------------- Based on a detailed analysis of differences in fuel consumption, petroleum imports, and imports of refined petroleum products among the Reference Case, High Economic Growth, and Low Economic Growth Scenarios presented in AEO 2009, NHTSA estimates that approximately 50 percent of the reduction in fuel consumption resulting from adopting higher CAFE standards is likely to be reflected in reduced U.S. imports of refined fuel, while the remaining 50 percent would be reduce domestic fuel refining.\513\ Of this latter figure, 90 percent is anticipated to reduce U.S. imports of crude petroleum for use as a refinery feedstock, while the remaining 10 percent is expected to reduce U.S. domestic production of crude petroleum.\514\ Thus on balance, each 100 gallons of fuel saved as a consequence of higher CAFE standards is anticipated to reduce total U.S. imports of crude petroleum or refined fuel by 95 gallons.\515\ --------------------------------------------------------------------------- \513\ Differences between forecast annual U.S. imports of crude petroleum and refined products among these three scenarios range from 24-89 percent of differences in projected annual gasoline and diesel fuel consumption in the U.S. These differences average 49 percent over the forecast period spanned by AEO 2009. \514\ Differences between forecast annual U.S. imports of crude petroleum among these three scenarios range from 67-97 percent of differences in total U.S. refining of crude petroleum, and average 85 percent over the forecast period spanned by AEO 2009. \515\ This figure is calculated as 50 gallons + 50 gallons * 90% = 50 gallons + 45 gallons = 95 gallons. --------------------------------------------------------------------------- l. Air Pollutant Emissions i. Impacts on Criteria Air Pollutant Emissions Criteria air pollutants emitted by vehicles and during fuel production include carbon monoxide (CO), hydrocarbon compounds (usually referred to as ``volatile organic compounds,'' or VOC), nitrogen oxides (NOX), fine particulate matter (PM2.5), and sulfur oxides (SOX). While reductions in domestic fuel refining and distribution that result from lower fuel consumption will reduce U.S. emissions of these pollutants, additional vehicle use associated with the rebound effect from higher fuel economy will increase their emissions. Thus the net effect of stricter CAFE standards on emissions of each criteria pollutant depends on the relative magnitudes of its reduced emissions in fuel refining and distribution, and increases in its emissions from vehicle use. Because the relationship between emissions in fuel refining and vehicle use is different for each criteria pollutant, the net effect of fuel savings from the proposed standards on total emissions of each pollutant is likely to differ. We note that any benefits in terms of criteria air pollutant reductions resulting from this rule would not be direct benefits. With the exception of SO2, NHTSA calculated annual emissions of each criteria pollutant resulting from vehicle use by multiplying its estimates of car and light truck use during each year over their expected lifetimes by per-mile emission rates appropriate to each vehicle type, fuel, model year, and age. These emission rates were developed by U.S. EPA using its Motor Vehicle Emission Simulator (Draft MOVES 2009).\516\ Emission rates for SO2 were calculated by NHTSA using average fuel sulfur content estimates supplied by EPA, together with the assumption that the entire sulfur content of fuel is emitted in the form of SO2.\517\ Total SO2 emissions under each alternative CAFE standard were calculated by applying the resulting emission rates directly to estimated annual gasoline and diesel fuel use by cars and light trucks. --------------------------------------------------------------------------- \516\ The MOVES model assumes that the per-mile rates at which these pollutants are emitted are determined by EPA regulations and the effectiveness of catalytic after-treatment of engine exhaust emissions, and are thus unaffected by changes in car and light truck fuel economy. \517\ These are 30 and 15 parts per million (ppm, measured on a mass basis) for gasoline and diesel respectively, which produces emission rates of 0.17 grams of SO2 per gallon of gasoline and 0.10 grams per gallon of diesel. --------------------------------------------------------------------------- As with other impacts, the changes in emissions of criteria air pollutants resulting from alternative increases in CAFE standards for MY 2012-2016 cars and light trucks were calculated from the differences between emissions under each alternative that would increase CAFE standards, and emissions under the baseline alternative. NHTSA estimated the reductions in criteria pollutant emissions from producing and distributing fuel that would occur under alternative CAFE standards using emission rates obtained by EPA from Argonne National Laboratories' Greenhouse Gases and Regulated Emissions in Transportation (GREET) model.\518\ The GREET model provides separate estimates of air pollutant emissions that occur in different phases of fuel production and distribution, including crude oil extraction, transportation, and storage, fuel refining, and fuel distribution and storage.\519\ EPA modified the GREET model to change certain assumptions about emissions during crude petroleum extraction and transportation, as well as to update its emission rates to reflect adopted and pending EPA emission standards. NHTSA converted these emission rates from the mass per fuel energy content basis on which GREET reports them to mass per gallon of fuel supplied using estimates of fuel energy content supplied by GREET. --------------------------------------------------------------------------- \518\ Argonne National Laboratories, The Greenhouse Gas and Regulated Emissions from Transportation (GREET) Model, Version 1.8, June 2007, available at http://www.transportation.anl.gov/modeling_ simulation/GREET/index.html (last accessed August 9, 2009). \519\ Emissions that occur during vehicle refueling at retail gasoline stations (primarily evaporative emissions of volatile organic compounds, or VOCs) are already accounted for in the ``tailpipe'' emission factors used to estimate the emissions generated by increased light truck use. GREET estimates emissions in each phase of gasoline production and distribution in mass per unit of gasoline energy content; these factors are then converted to mass per gallon of gasoline using the average energy content of gasoline. --------------------------------------------------------------------------- The resulting emission rates were applied to the agency's estimates of fuel consumption under each alternative CAFE standard to develop estimates of total emissions of each criteria pollutant during fuel production and distribution. The assumptions about the effects of changes in fuel consumption on domestic and imported sources of fuel supply discussed above were then employed to calculate the effects of [[Page 49675]] reductions in fuel use from alternative CAFE standards on changes in imports of refined fuel and domestic refining. NHTSA's analysis assumes that reductions in imports of refined fuel would reduce criteria pollutant emissions during fuel storage and distribution only. Reductions in domestic fuel refining using imported crude oil as a feedstock are assumed to reduce emissions during fuel refining, storage, and distribution, because each of these activities would be reduced. Reduced domestic fuel refining using domestically-produced crude oil is assumed to reduce emissions during all four phases of fuel production and distribution.\520\ --------------------------------------------------------------------------- \520\ In effect, this assumes that the distances crude oil travels to U.S. refineries are approximately the same regardless of whether it travels from domestic oilfields or import terminals, and that the distances that gasoline travels from refineries to retail stations are approximately the same as those from import terminals to gasoline stations. --------------------------------------------------------------------------- Finally, NHTSA calculated the net changes in domestic emissions of each criteria pollutant by summing the increases in emissions projected to result from increased vehicle use, and the reductions anticipated to result from lower domestic fuel refining and distribution.\521\ As indicated previously, the effect of adopting higher CAFE standards on total emissions of each criteria pollutant depends on the relative magnitudes of the resulting reduction in emissions from fuel refining and distribution, and the increase in emissions from additional vehicle use. Although these net changes vary significantly among individual criteria pollutants, the agency projects that on balance, adopting higher CAFE standards would reduce emissions of all criteria air pollutants except carbon monoxide (CO). --------------------------------------------------------------------------- \521\ All emissions from increased vehicle use are assumed to occur within the U.S., since CAFE standards would apply only to vehicles produced for sale in the U.S. --------------------------------------------------------------------------- The net changes in domestic emissions of fine particulates (PM2.5) and its chemical precursors (such as NOX, SOX, and VOCs) are converted to economic values using estimates of the reductions in health damage costs per ton of emissions of each pollutant that is avoided, which were developed and recently revised by EPA. These savings represent the estimated reductions in the value of damages to human health resulting from lower atmospheric concentrations and population exposure to air pollution that occur when emissions of each pollutant that contributes to atmospheric PM2.5 concentrations are reduced. The value of reductions in the risk of premature death due to exposure to fine particulate pollution (PM2.5) account for a majority of EPA's estimated values of reducing criteria pollutant emissions, although the value of avoiding other health impacts is also included in these estimates. These values do not include a number of unquantified benefits, such as reduction in the welfare and environmental impacts of PM2.5 pollution, or reductions in health and welfare impacts related to other criteria pollutants (ozone, NO2, and SO2) and air toxics. EPA estimates different PM-related per-ton values for reducing emissions from vehicle use than for reductions in emissions of that occur during fuel production and distribution.\522\ NHTSA applies these separate values to its estimates of changes in emissions from vehicle use and fuel production and distribution to determine the net change in total economic damages from emissions of these pollutants. --------------------------------------------------------------------------- \522\ These reflect differences in the typical geographic distributions of emissions of each pollutant, their contributions to ambient PM2.5 concentrations, pollution levels (predominantly those of PM2.5), and resulting changes in population exposure. --------------------------------------------------------------------------- EPA projects that the per-ton values for reducing emissions of criteria pollutants from both mobile sources (including motor vehicles) and stationary sources such as fuel refineries and storage facilities will increase over time. These projected increases reflect rising income levels, which are assumed to increase affected individuals' willingness to pay for reduced exposure to health threats from air pollution, as well as future population growth, which increases population exposure to future levels of air pollution. ii. Reductions in CO2 Emissions Emissions of carbon dioxide and other greenhouse gases (GHGs) occur throughout the process of producing and distributing transportation fuels, as well as from fuel combustion itself. By reducing the volume of fuel consumed by passenger cars and light trucks, higher CAFE standards will reduce GHG emissions generated by fuel use, as well as throughout the fuel supply cycle. Lowering these emissions is likely to slow the projected pace and reduce the ultimate extent of future changes in the global climate, thus reducing future economic damages that changes in the global climate are expected to cause. By reducing the probability that climate changes with potentially catastrophic economic or environmental impacts will occur, lowering GHG emissions may also result in economic benefits that exceed the resulting reduction in the expected future economic costs caused by gradual changes in the earth's climatic systems. Quantifying and monetizing benefits from reducing GHG emissions is thus an important step in estimating the total economic benefits likely to result from establishing higher CAFE standards. The agency estimated emissions of CO2 from passenger car and light truck use by multiplying the number of gallons of each type of fuel (gasoline and diesel) they are projected to consume under alternative CAFE standards by the quantity or mass of CO2 emissions released per gallon of fuel consumed. This calculation assumes that the entire carbon content of each fuel is converted to CO2 emissions during the combustion process. Carbon dioxide emissions account for nearly 95 percent of total GHG emissions that result from fuel combustion during vehicle use. iii. Economic Value of Reductions in CO2 Emissions NHTSA has taken the economic benefits of reducing CO2 emission into account in this rulemaking, both in developing proposed CAFE standards and in assessing the economic benefits of each alternative that was considered. Since direct estimates of the economic benefits from reducing GHG emissions are generally not reported in published literature on the impacts of climate change, these benefits are typically assumed to be the ``mirror image'' of the estimated incremental costs resulting from an increase in those emissions. That is, the benefits from reducing emissions are usually measured by the savings in estimated economic damages that an equivalent increase in emissions would otherwise have caused. The ``social cost of carbon'' (SCC) is intended to be a monetary measure of the incremental damage resulting from carbon dioxide (CO2) emissions, including (but not limited to) net agricultural productivity loss, human health effects, property damages from sea level rise, and changes in ecosystem services. Any effort to quantify and to monetize the consequences associated with climate change will raise serious questions of science, economics, and ethics. But with full regard for the limits of both quantification and monetization, the SCC can be used to provide an estimate of the social benefits of reductions in GHG emissions. For at least four reasons, any particular figure will be contestable. First, scientific and economic knowledge about the impacts of climate change continues to grow. With new and better information about relevant questions, including the cost, burdens, and possibility of adaptation, current [[Page 49676]] estimates will inevitably change over time. Second, some of the likely and potential damages from climate change--for example, the loss of endangered species--are generally not included in current SCC estimates. These omissions may turn out to be significant; in the sense that they may mean that the best current estimates are too low. As noted by the IPCC Fourth Assessment Report, ``It is very likely that globally aggregated figures underestimate the damage costs because they cannot include many non-quantifiable impacts.'' Third, it is unlikely that the damage estimates account for the directed technological change that will lead to innovations that reduce the costs of responding to climate change--for example, it is likely that scientists will develop crops that are better able to withstand high temperatures. In this respect, the current estimates may overstate the likely damages. Fourth, controversial ethical judgments, including those involving the treatment of future generations, play a role in judgments about the SCC (see in particular the discussion of the discount rate, below). To date, SCC estimates presented in recent regulatory documents have varied within and among agencies, including DOT, DOE, and EPA. For example, a regulation proposed by DOT in 2008 assumed a value of $7 per ton CO2 \523\ (2006$) for 2011 emission reductions (with a range of $0-14 for sensitivity analysis). A regulation finalized by DOE used a range of $0-$20 (2007$). Both of these ranges were designed to reflect the value of damages to the United States resulting from carbon emissions, or the ``domestic'' SCC. In the final MY 2011 CAFE EIS, DOT used both a domestic SCC value of $2/tCO2 and a global SCC value of $33/tCO2 (with sensitivity analysis at $80/ tCO2), increasing at 2.4 percent per year thereafter. The final MY 2011 CAFE rule also presented a range from $2 to $80/ tCO2. EPA's Advance Notice of Proposed Rulemaking for Greenhouse Gases discussed the benefits of reducing GHG emissions and identified what it described as ``very preliminary'' SCC estimates ``subject to revision'' that spanned three orders of magnitude. EPA's global mean values were $68 and $40/tCO2 for discount rates of 2 percent and 3 percent respectively (in 2006 real dollars for 2007 emissions).\524\ --------------------------------------------------------------------------- \523\ For the purposes of this discussion, we present all values of the SCC as the cost per ton of CO2 emissions. Some discussions of the SCC in the literature use an alternative presentation of a dollar per ton of Carbon. The standard adjustment factor is 3.67, which means, for example, that a SCC of $10 per ton of CO2 would be equivalent to a cost of $36.70 for a ton of carbon emitted. \524\ 73 FR 44416 (July 30, 2008). EPA, ``Advance Notice of Proposed Rulemaking for Greenhouse Gases Under the Clean Air Act, Technical Support Document on Benefits of Reducing GHG Emissions,'' June 2008. www.regulations.gov. Search for ID ``EPA-HQ-OAR-2008- 0318-0078.'' --------------------------------------------------------------------------- The current Administration has worked to develop a transparent methodology for selecting a set of interim SCC estimates to use in regulatory analyses until a more comprehensive characterization of the distribution of SCC is developed. This discussion proposes a set of values for the interim social cost of carbon. It should be emphasized that the analysis here is preliminary. Today's proposed joint rulemaking presents SCC estimates that reflect the Administration's current understanding of the relevant literature. These interim estimates are being used for the short-term while an interagency group develops a more comprehensive characterization of the distribution of SCC values for future economic and regulatory analyses. The interim values should not be viewed as a statement about the results of the longer-term process. The Administration will be evaluating and seeking comment in the preamble to today's proposed rule on all of the scientific, economic, and ethical issues before establishing final estimates for use in future rulemakings. The outcomes of the Administration's process to develop interim values are judgments in favor of (a) global rather than domestic values, (b) an annual growth rate of 3%, and (c) interim global SCC estimates for 2007 (in 2006 dollars) of $55, $33, $19, $10, and $5 per ton of CO2. Notably, we have centered our current attention on a SCC of $19. The proposed figures are based on the following judgments. 1. Global and domestic measures. Because of the distinctive nature of the climate change problem, we present both a global SCC and a fraction of that value that represents impacts that may occur within the borders of the U.S. alone, or a ``domestic'' SCC, but center our current attention on the global measure. This approach represents a departure from past practices, which relied, for the most part, on domestic measures. As a matter of law, both global and domestic values are permissible; the relevant statutory provisions are ambiguous and allow selection of either measure.\525\ --------------------------------------------------------------------------- \525\ It is true that Federal statutes are presumed not to have extraterritorial effect, in part to ensure that the laws of the United States respect the interests of foreign sovereigns. But use of a global measure for the SCC does not give extraterritorial effect to Federal law and hence does not intrude on such interests. --------------------------------------------------------------------------- It is true that under OMB guidance, analysis from the domestic perspective is required, while analysis from the international perspective is optional. The domestic decisions of one nation are not typically based on a judgment about the effects of those decisions on other nations. But the climate change problem is highly unusual in the sense that it involves (a) a global public good in which (b) the emissions of one nation may inflict significant damages on other nations and (c) the United States is actively engaged in promoting an international agreement to reduce worldwide emissions. In these circumstances, we believe the global measure is preferred. Use of a global measure reflects the reality of the problem and is expected to contribute to the continuing efforts of the United States to ensure that emissions reductions occur in many nations. Domestic SCC values are also presented. The development of a domestic SCC is greatly complicated by the relatively few region- or country-specific estimates of the SCC in the literature. One potential source of estimates comes from a recent unpublished EPA modeling effort using the FUND model. The resulting estimates suggest that the ratio of domestic to global benefits varies with key parameter assumptions. With a 3 percent discount rate, for example, the U.S. benefit is about 6 percent of the global benefit for the ``central'' (mean) FUND results, while, for the corresponding ``high'' estimates associated with a higher climate sensitivity and lower global economic growth, the U.S. benefit is less than 4 percent of the global benefit. With a 2 percent discount rate, the U.S. share is about 2-5 percent of the global estimate. Based on this available evidence, an interim domestic SCC value equal to 6 percent of the global damages is proposed. This figure is in the middle of the range of available estimates from the literature. It is recognized that the 6 percent figure is approximate and highly speculative and alternative approaches will be explored before establishing final values for future rulemakings. 2. Filtering existing analyses. There are numerous SCC estimates in the existing literature, and it is reasonable to make use of those estimates in order to produce a figure for current use. A starting point is provided by the meta-analysis in Richard Tol, 2008.\526\ With [[Page 49677]] that starting point, the Administration proposes to ``filter'' existing SCC estimates by using those that (1) are derived from peer-reviewed studies; (2) do not weight the monetized damages to one country more than those in other countries; (3) use a ``business as usual'' climate scenario; and (4) are based on the most recent published version of each of the three major integrated assessment models (IAMs): FUND, PAGE, and DICE. --------------------------------------------------------------------------- \526\ Richard Tol, The Social Cost of Carbon: Trends, Outliers, and Catastrophes, Economics: The Open-Access, Open-Assessment E- Journal, Vol. 2, 2008-25. http://www.economics-ejournal.org/ economics/journalarticles/2008-25(2008). --------------------------------------------------------------------------- Proposal (1) is based on the view that those studies that have been subject to peer review are more likely to be reliable than those that have not been. Proposal (2) is based on a principle of neutrality and simplicity; it does not treat the citizens of one nation (or different citizens within the U.S.) differently on the basis of speculative or controversial considerations. Further, it is consistent with the potential compensation tests of Kaldor (1939) and Hicks (1940), which use unweighted sums of willingness to pay. Finally, this is the approach used in rulemakings across a variety of settings and consequently keeps U.S. government policy consistent across contexts. Proposal (3) stems from the judgment that as a general rule, the proper way to assess a policy decision is by comparing the implementation of the policy against a counterfactual state where the policy is not implemented. In addition, our expectation is that most policies to be evaluated using these interim SCC estimates will constitute small enough changes to the larger economy to safely assume that the marginal benefits of emissions reductions will not change between the baseline and policy scenarios. A departure from this approach would be to consider a more dynamic setting in which other countries might implement policies to reduce GHG emissions at an unknown future date and the U.S. could choose to implement such a policy now or at a future date. Proposal (4) is based on four complementary judgments. First, the FUND, PAGE, and DICE models now stand as the most comprehensive and reliable efforts to measure the economic damages from climate change. Second, the latest versions of the three IAMs are likely to reflect the most recent evidence and learning, and hence they are presumed to be superior to those that preceded them. Third, any effort to choose among them, or to reject one in favor of the others, would be difficult to defend at the present time. In the absence of a clear reason to choose among them, it is reasonable to base the SCC on all of them. Fourth, in light of the uncertainties associated with the SCC, the additional information offered by different models is important. 3. Use a model-weighted average of the estimates at each discount rate. At this time, a scientifically valid reason to prefer any of the three major IAMs (FUND, PAGE, and DICE) has not been identified. Accordingly, to address the concern that certain models not be given unequal weight relative to the other models, the estimates are based on an equal weighting of the means of the estimates from each of the models. Among estimates that remain after applying the filter, we begin by taking the average of all estimates within a model. The estimated SCC is then calculated as the average of the three model-specific averages. This approach is used to ensure that models with a greater number of published results do not exert unequal weight on the interim SCC estimates. 4. Apply a 3 percent annual growth rate to the chosen SCC values. SCC is assumed to increase over time, because future emissions are expected to produce larger incremental damages as physical and economic systems become more stressed as the magnitude of climate change increases. Indeed, an implied growth rate in the SCC can be produced by most of the models that estimate economic damages caused by increased GHG emissions in future years. But neither the rate itself nor the information necessary to derive its implied value is commonly reported. In light of the limited amount of debate thus far about the appropriate growth rate of the SCC, applying a rate of 3 percent per year seems appropriate at this stage. This value is consistent with the range recommended by IPCC (2007) and close to the latest published estimate (Hope 2008). (1) Discount Rates For estimation of the benefits associated with the mitigation of climate change, one of the most complex issues involves the appropriate discount rate. OMB's current guidance offers a detailed discussion of the relevant issues and calls for discount rates of 3 percent and 7 percent. It also permits a sensitivity analysis with low rates (1-3 percent) for intergenerational problems: ``If your rule will have important intergenerational benefits or costs you might consider a further sensitivity analysis using a lower but positive discount rate in addition to calculating net benefits using discount rates of 3 and 7 percent.'' \527\ --------------------------------------------------------------------------- \527\ See OMB Circular A-4, pp. 35-36, citing Portney and Weyant, eds. (1999), Discounting and Intergenerational Equity, Resources for the Future, Washington, DC. --------------------------------------------------------------------------- The choice of a discount rate, especially over long periods of time, raises highly contested and exceedingly difficult questions of science, economics, philosophy, and law. See, e.g., William Nordhaus, The Challenge of Global Warming (2008); Nicholas Stern, The Economics of Climate Change (2007); Discounting and Intergenerational Equity (Paul Portney and John Weyant eds. 1999). It is not clear that future generations would be willing to trade environmental quality for consumption at the same rate as the current generations. Under imaginable assumptions, decisions based on cost-benefit analysis with high discount rates might harm future generations--at least if investments are not made for the benefit of those generations. See Robert Lind, Analysis for Intergenerational Discounting, id. at 173, 176-177. It is also possible that the use of low discount rates for particular projects might itself harm future generations, by ensuring that resources are not used in a way that would greatly benefit them. In the context of climate change, questions of intergenerational equity are especially important. Reasonable arguments support the use of a 3 percent discount rate. First, that rate is among the two figures suggested by OMB guidance, and hence it fits with existing national policy. Second, it is standard to base the discount rate on the compensation that people receive for delaying consumption, and the 3 percent is close to the risk-free rate of return, proxied by the return on long term inflation-adjusted U.S. Treasury Bonds, as of this writing. Although these rates are currently closer to 2.5 percent, the use of 3 percent provides an adjustment for the liquidity premium that is reflected in these bonds' returns. At the same time, others would argue that a 5 percent discount rate can be supported. The argument relies on several assumptions. First, that rate can also be justified by reference to the level of compensation for delaying consumption, because it fits with market behavior with respect to individuals' willingness to trade-off consumption across periods as measured by the estimated post-tax average real returns to risky private investments (e.g., the S&P 500). In the climate setting, the 5 percent discount rate may be preferable to the riskless rate because it is based on risky investments and the return to projects to mitigate climate change is also risky. In contrast, the 3 percent riskless rate may be a more appropriate discount rate for [[Page 49678]] projects where the return is known with a high degree of confidence (e.g., highway guardrails). In principal, the correct discount rate would reflect the variance in payoff from climate mitigation policy and the correlation between the payoffs of the policy and the broader economy.\528\ --------------------------------------------------------------------------- \528\ Specifically, if the benefits of the policy are highly correlated with the returns from broader economy, then the market rate should be used to discount the benefits. If the benefits are uncorrelated with the broader economy the long term government bond rate should be applied. Furthermore, if the benefits are negatively correlated with the broader economy a rate less than that on long term government bonds should be used (Lind, 1982 pp. 89-90). --------------------------------------------------------------------------- Second, 5 percent, and not 3 percent, is roughly consistent with estimates implied by reasonable inputs to the theoretically derived Ramsey equation, which specifies the optimal time path for consumption. That equation specifies the optimal discount rate as the sum of two components. The first term (the product of the elasticity of the marginal utility of consumption and the growth rate of consumption) reflects the fact that consumption in the future is likely to be higher than consumption today, so diminishing marginal utility implies that the same monetary damage will cause a smaller reduction of utility in the future. Standard estimates of this term from the economics literature are in the range of 3 percent-5 percent. The second component reflects the possibility that a lower weight should be placed on utility in the future, to account for social impatience or extinction risk, which is specified by a pure rate of time preference (PRTP). A common estimate of the PRTP is 2 percent, though some observers believe that a principle of intergenerational equity suggests that the PRTP should be close to zero. It follows that discount rate of 5 percent is near the middle of the range of values that are able to be derived from the Ramsey equation. It is recognized that the arguments above--for use of market behavior and the Ramsey equation--face objections in the context of climate change, and of course there are alternative approaches. In light of climate change, it is possible that consumption in the future will not be higher than consumption today, and if so, the Ramsey equation will suggest a lower figure. However, the historical evidence is consistent with rising consumption over time. Some critics note that using observed interest rates for inter- generational decisions imposes current preferences on future generations, which some economists say may not be appropriate. For generational equity, they argue that the discount rate should be below market rates to correct for market distortions and inefficiencies in inter-generational transfers of wealth (which are presumed to compensate future generations for damage), and to treat generations equitably based on ethical principles (see Broome 2008).\529\ --------------------------------------------------------------------------- \529\ See Arrow, K.J., W.R. Cline, K-G Maler, M. Munasinghe, R. Squiteri, J.E. Stiglitz, 1996. ``Intertemporal equity, discounting and economic efficiency,'' in Climate Change 1995: Economic and Social Dimensions of Climate Change, Contribution of Working Group III to the Second Assessment Report of the Intergovernmental Panel on Climate Change. See also Weitzman, M.L., 1999. In Portney, P.R. and Weyant J.P. (eds.), Discounting and Intergenerational Equity, Resources for the Future, Washington, DC. --------------------------------------------------------------------------- Additionally, some analyses attempt to deal with uncertainty with respect to interest rates over time. We explore below how this might be done.\530\ --------------------------------------------------------------------------- \530\ Richard Newell and William Pizer, Discounting the distant future: how much do uncertain rates increase valuations? J. Environ. Econ. Manage. 46 (2003) 52-71. --------------------------------------------------------------------------- (2) Proposed Interim Estimates The application of the methodology outlined above yields interim estimates of the SCC that are reported in Table IV.C.3-2. These estimates are reported separately using 3 percent and 5 percent discount rates. The cells are empty in rows 10 and 11, because these studies did not report estimates of the SCC at a 3 percent discount rate. The model-weighted means are reported in the final or summary row; they are $33 per tCO2 at a 3 percent discount rate and $5 per tCO2 with a 5 percent discount rate. --------------------------------------------------------------------------- \531\ Most of the estimates in Table 1 rely on climate scenarios developed by the Intergovernmental Panel on Climate Change (IPCC). The IPCC published a new set of scenarios in 2000 for use in the Third Assessment Report (Special Report on Emissions Scenarios-- SRES). The SRES scenarios define four narrative storylines: A1, A2, B1 and B2, describing the relationships between the forces driving greenhouse gas and aerosol emissions and their evolution during the 21st century for large world regions and globally. Each storyline represents different demographic, social, economic, technological, and environmental developments that diverge in increasingly irreversible ways. The storylines are summarized in Nakicenovic et al., 2000 (see also http://sedac.ciesin.columbia.edu/ddc/sres/