In the committee’s judgment, the concept of incremental retail price equivalent cost is most appropriate for the NHTSA’s purposes because it best represents the full, long-run economic costs of increasing fuel economy. The NHTSA has used the RPE method in its rulemakings on fuel economy, for example in the final rule for model year 2011 light-duty vehicles (DOT/NHTSA, 2009, pp. 346-352). Incremental RPE estimates are intended to represent the average additional price that consumers would pay for a fuel economy technology implemented in a typical vehicle under average economic conditions and typical manufacturing practices. These estimates are intended to represent long-run, high-volume, industry-average production costs, incorporating rates of profit and overhead expenses including warranties, transport, and retailing. Although learning and technological progress never stop, RPEs are intended to represent costs after an initial period of rapid cost reduction that results from learning by doing.2 The committee uses the term substantially learned as opposed to fully learned to convey that cost reductions due to increasing volumes may continue to occur. RPEs are not intended to replicate the market price of a specific vehicle or a specific optional feature at a specific time. The market price of a particular vehicle at a particular time depends on many factors (e.g., market trends, marketing strategies, profit opportunities, business cycles, temporary shortages or surpluses) other than the cost of manufacturing and retailing a vehicle or any given component. It is not appropriate to base a long-term policy such as fuel economy standards on short-run conditions or special circumstances.

The RPE concept, unfortunately, is not easy to apply. It raises a number of difficult questions about appropriate premises and assumptions and reliable sources of data. It frequently relies on the application of markup factors, which could vary depending on the nature of the technology and the basis for the original cost estimate. When an RPE markup factor is used, the definition of the cost to which it applies is critical. Much of the disagreement over RPE multipliers can be traced to inconsistent definition of the cost to be marked up. The following are key premises of the committee’s application of the RPE method.

  • Incremental RPE. The relevant measure of cost is the change in RPE in comparison to an equivalent vehicle without the particular fuel economy technology. More often than not, a fuel economy technology replaces an existing technology. For example, a 6-speed automatic transmission replaces a 5-speed, a compression-ignition (CI) engine replaces a spark-ignition (SI) engine, or a set of low-rolling-resistance tires replaces a set with higher rolling resistance. What matters is the change in RPE rather than the total RPE of the new technology. This requires that an estimate of the RPE of the existing technology be subtracted from that of the new technology.

  • Equivalent vehicle size and performance. Estimating the cost of decreasing fuel consumption requires one to carefully specify a basis for comparison. The committee considers that to the extent possible, fuel consumption cost comparisons should be made at equivalent acceleration performance and equivalent vehicle size. Other vehicle attributes matter as well, such as reliability, noise, and vibration. Ideally, cost and fuel economy comparisons should be made on the basis of no compromise for the consumer. Often there are differences of opinion about what design and engineering changes may be required to ensure no compromise for the consumer. This, in turn, leads to differing bills of materials to be costed out, which leads to significant differences in incremental RPE estimates.

  • Learning by doing, scale economies, and competition. When new technologies are first introduced and only one or two suppliers exist, costs are typically higher than they will be in the long run due to lack of scale economies, as-yet-unrealized learning by doing, and limited competition. These transitional costs can be important to manufacturers’ bottom lines and should be considered. However, nearly all cost estimates are developed assuming long-run, high-volume, average economic conditions. Typical assumptions include (1) high volume, (2) substantially learned component costs, and (3) competition provided by at least three global suppliers available to each manufacturer (Martec Group, Inc., 2008a, slide 3). Under these assumptions, it is not appropriate to employ traditional learning curves to predict future reductions in cost as production experience increases. However, if cost estimates are for novel technology and do not reflect learning by doing, then the application of learning curves as well as the estimation of scale economies may be appropriate. The use of such methods introduces substantial uncertainty, however, since there are no proven methods for predicting the amount of cost reduction that a new technology will achieve.

  • Normal product cycles. As a general rule, premises include normal redesign and product turnover schedules. Accelerated rates of implementation can increase costs by decreasing amortization periods and by demanding more engineering and design resources than are available. Product cycles are discussed in Chapter 7.

  • Purchased components versus in-house manufacture. Costs can be estimated at different stages in the manufacturing process. Manufacturing cost estimates gen-


Learning by doing represents the increase in productivity and decrease in cost that occurs during a technology’s lifetime as a result of manufacturers’ gaining experience in producing the technology. The impacts of learning on costs can be represented as a volume-based learning where costs reductions occur with increasing production levels or as a time-based learning where cost reductions occur over time.

The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement