Technologies’ costs and fuel economy impacts are both important for setting fuel economy and greenhouse gas (GHG) emissions standards. The direct manufacturing costs of component technologies or technology packages are the most important pieces of information, but the degree to which the technology and design changes will affect the indirect costs of firms also matters. The rate of technological and design changes required by the standards can also affect costs by making capital investments prematurely obsolete or requiring greater than normal engineering effort. Technology and design changes can also have secondary impacts on consumers’ satisfaction and corporate profits.
Uncertainty about future costs is inescapable because of the uncertain rate and direction of future technological progress, as well as uncertainties about the future prices of materials, energy, labor, and capital. Although technological change is certain, its direction, magnitude, and impacts on cost are difficult to predict. For most components, manufacturing costs tend to decrease with increased production volumes and with the accumulation of experience. However, there are no exact methods for predicting future rates of learning by doing or technological progress. Assuming no technological progress or cost reduction via learning will likely overestimate the costs of compliance. On the other hand, overly optimistic assumptions will result in underestimation of costs. In this chapter the committee discusses methodological issues with estimating the costs of fuel economy technologies and manufacturing issues with deploying these technologies.
Rigorously estimating costs requires carefully defining terms. The most important concepts are direct manufacturing costs, indirect costs, and total costs.1
The elementary cost components of manufacturing are materials, energy, labor, and capital. Cost estimation begins with an understanding of the quantities of each component required to produce a certain number of units per year: Unit cost2 is calculated by multiplying the quantities by their prices, summing to yield total cost, and dividing by the volume of the production. In general, total cost is not a linear function of production volume. There are fixed costs that are required to produce any units and variable costs that do increase more linearly with production volume, but even this is an oversimplification. In general, total costs increase less than linearly with production volume due to economies of scale. Thus, estimates of unit costs require specification of a production volume. Fixed costs are typically amortized over assumed levels of production for a certain number of years. Thus, unit cost estimates also depend on production volumes and the usable lifetimes of fixed investments. In addition, prices of all of the components tend to vary over time and location, to different degrees. Prices may vary with the degree of competition in the supply chain. Price variability is not the only source of uncertainty. The quantities of labor, capital, and energy required may also vary over time and with improvements in manufacturing processes.
Direct manufacturing costs (DMC) are defined alternatively as the price an original equipment manufacturer (OEM) would pay a supplier for a fully manufactured part ready for assembly in a vehicle, or the OEM’s total cost of internally manufacturing the same part. Thus, if the part is made by the OEM, the DMC is typically considered to be
1 As discussed in Chapter 1, the committee defines total costs for a technology as direct manufacturing costs plus indirect costs. Total cost does not represent the cost of ownership to the consumer, which is discussed in Chapter 9 as the private cost of ownership.
2 Unit cost is the cost of a specified unit of a product or service.
the materials, labor, and energy required by the OEM to manufacture the part, similar to the definitions provided by Helfand and Sherwood (2013) and Blincoe (2013). If the part is made by a supplier, the DMC includes the research and development and indirect costs of the supplier. This would not include integration and other indirect costs borne by the OEM, however. It includes all the suppliers’ fixed costs necessary to manufacture the part, as well as a normal rate of return for the supplier on capital. DMCs are the most important element of cost estimation because they typically make up the majority of total costs and because indirect costs tend to vary with DMC.
Indirect costs (IC) include expenditures not directly required for manufacturing a component technology but necessary for the operation of an automobile manufacturing firm. Direct manufacturing costs plus indirect costs equal what economists call the long-run average cost of manufacturing, or, simply, total costs. In an ideal competitive market in equilibrium, price would equal long-run average cost. In the less-than-perfect dynamic markets of the real world, the two frequently differ. Indirect cost estimates may also have a large impact on estimated total costs. For a typical OEM, indirect costs (including a typical rate of profit) average about 50 percent of direct manufacturing costs (RTI/UMTRI 2009).
There are many ways to describe indirect costs. In the following section the committee discusses the two principal approaches to representing indirect costs for a technology: (1) the retail price equivalent (RPE) markup and (2) the indirect cost multiplier (ICM). Though the methods for estimating these multipliers are different, both the RPE and ICM represent the costs for producing a technology that are not included in the estimate of the direct manufacturing costs. Table 7.1 shows a breakdown of the components included in indirect costs based on information available for publicly traded firms (RTI/UMTRI 2009, Table 3-3).
|OEM vehicle production overhead||Warranty|
|Research and development|
|Capital depreciation and amortization|
|Maintenance, repair, and operations|
|Employee benefits (e.g. health insurance, retirement plans)|
|Sales including dealer costs||Transportation|
|Dealer selling costs (inventory, advertising, labor, etc.)|
Estimating Direct Manufacturing Costs
There are a variety of sources of information about current manufacturing costs. Teardown studies performed by competent third parties are the most reliable sources of cost estimates but are also the most expensive. Other sources are useful but may have biases or inadequacies (e.g., basing direct manufacturing costs on market prices that also reflect dynamic effects of supply and demand) and so must be interpreted with caution. Cost quotes for fully manufactured components may be provided by OEMs or suppliers. The different sources, as well as their strengths and weaknesses, were described in the Phase 1 report (NRC 2011).
The 2025 rulemaking makes use of a small number of detailed teardown cost analyses. The NRC Phase 1 report recognized the need for cost estimates based on teardown studies and concluded that increasing the use of this approach would increase confidence in the accuracy of the costs (NRC 2011). In 2009, the Environmental Protection Agency (EPA) contracted with FEV, Inc., to perform a cost analysis of converting a conventional naturally aspirated, fuel-injected I4 engine to a stoichiometric, direct-injected, turbocharged and downsized I4 engine (FEV 2009). This appears to be the first instance of the Agencies’ use of cost estimation based on tearing down vehicles, itemizing each part, estimating the purchase or direct manufacturing cost of each part and subsystem, and constructing detailed cost estimation models for the complete system change. This level of detail, together with documentation of the costs of every item and process, makes possible a more complete understanding of the reasons for differences in cost among technologies. Since the Phase 1 report, additional teardown cost studies have been carried out for mild (FEV 2011a) and full hybrid (FEV 2012), valve-train technologies (FEV 2011a), advanced transmissions (FEV 2011b), and mass reduction (FEV 2012). Key assumptions include high-volume manufacturing (450,000 units/year) in North America. The volume assumptions made by the Agencies in order to scale costs are further discussed in the following section on economies of scale.
Estimating Indirect Costs
Rogozhin et al. (2010) calculated indirect costs by category for eight major automobile manufacturers using publicly available annual reports from the OEMs for 2007. Total indirect costs as a ratio to direct manufacturing costs were similar for all manufacturers (Figure 7.1): The minimum ratio was 0.45, the maximum was 0.49, and the average was 0.46. The indirect cost calculations excluded a category entitled “Other costs (not included as contributors),” shown in Figure 7.1 as “Other.” Had the “Other” costs been included, it would have raised the average to 0.50. The distribution of indirect costs by category is more variable than the total, however. Costs classified as “Corporate overhead” vary from 0.04 to 0.14,
FIGURE 7.1 Indirect costs as a percent of direct manufacturing costs by OEM, 2007.
SOURCE: Rogozhin et al. (2010).
and “Production overhead” ranges from 0.13 to 0.22. The rate of manufacturer profit (excluding dealer profit) in 2007 also varied in a relatively narrow range across OEMs from 3 percent to 9 percent.
The small variations in total indirect costs among firms may reflect differences in accounting conventions as much as real cost differences. There is no strict definition of what must be included in direct manufacturing costs. If the standard is the price an OEM would pay a supplier for a fully manufactured part ready for assembly, then direct manufacturing costs should include amortization of capital required for the subassembly, maintenance of facilities used, and profit on the operations. It is not clear that internal definitions of direct manufacturing costs always include all of these components.
The industry average ratio of indirect costs to direct manufacturing costs appears to fluctuate within a range of +/− 0.1 over time. An investigation of the ratio of total costs to direct manufacturing costs by the National Highway Traffic Safety Administration (NHTSA) for 1972-1997 found that the ratios fluctuated between 1.4 and 1.6 but without any apparent trend (Figure 7.2). In this regard, the ratio of total to direct costs represents the average markup above direct costs for all technologies produced in a given year. Figure 7.2 indicates that to cover costs, provide a return to investors, and remain competitive in the marketplace, OEMs have typically set prices that average 1.5 times direct costs (Blincoe 2013). The consistency of the 1.5, or 50 percent, markup from manufacturing costs to retail price is noteworthy and suggests that it is reasonable to assume that the relative share of indirect costs
FIGURE 7.2 Total costs as a ratio to direct manufacturing costs (RPE), 1972-1997 and 2007.
SOURCE: EPA/NHTSA (2012).
may not change much in the future. For this reason, the Final Rule increased the estimated 46 percent markup to 50 percent for use in estimating the indirect cost multipliers used in calculating the total costs of technologies used to meet the standards (EPA/NHTSA 2012). The similarity of the indirect cost shares across manufacturers and over time is also consistent with the view that competition among manufacturers is robust in that a manufacturer with substantially above-average indirect costs would be unsuccessful in the market.
RPE and ICM Methods
Retail Price Equivalent
The RPE makes no distinctions among technologies with respect to their impacts on indirect costs. It assigns an average indirect cost percentage to all technologies, thereby avoiding the question of attributing changes in indirect costs to specific technologies or design changes. This approach maintains the typical markup rate because fuel economy technologies add to the price of the vehicle. Based on the available data, a reasonable RPE multiplier would be 1.5.
The RPE method was used in previous rulemakings by NHTSA (DOT/NHTSA 2009, 173) and previous NRC fuel economy studies. NRC (2002) used a value of 1.4, while the Phase 1 study and NHTSA (DOT/NHTSA 2009) determined that a value of 1.5 was appropriate (NRC 2011, 36). A study comparing estimates of RPE multipliers concluded that a value of 1.56 was appropriate for outsourced parts purchased from suppliers (Vyas et al. 2000). Duleep (2008) recommended RPEs ranging from 1.65 to 1.73 depending on the complexity of the part; In 2003, McKinsey (as quoted in Bussmann 2008) produced a 1.7 RPE multiplier; Bussmann (2008) calculated an RPE of 2.0 for data obtained from Chrysler for 2003-2004. The Phase 1 report committee commissioned a study by IBIS Associates (2008) that costed out all components of a Honda Accord sedan and a Ford F-150 pickup truck. For the Honda, the RPE multiplier based on average transaction price was 1.39, while the markup to manufacturer’s suggested retail price (MSRP) was 1.49. For the F-150 the corresponding markups were 1.52 for transaction price and 1.54 for MSRP (NRC 2011, 33).
Indirect Cost Multiplier
The alternative is to estimate the impact of each technology on each component of indirect cost. In theory, this approach seems clearly superior to assuming identical impacts for all technologies regardless of their nature. However, attribution can be ambiguous, especially for future costs. Whether or not a specific technology will add to warranty or advertising costs is highly uncertain, for example. Does selling cars with a higher value added require additional corporate staff? Does it add to dealers’ inventory costs? Unfortunately, there is a general lack of empirical data on which to base such attributions.
The indirect cost multiplier (ICM) method is an application of activity based costing (ABC) methods to fuel economy technologies. ABC accounting attempts to assign costs to products based on the activities they require. In the case of ICMs, however, the costs are not assigned to the final product (an automobile) but rather to specific components of the final product. The difficulty is that the accounting data for automobile manufacturers that is publicly available in annual reports is organized according to standard financial accounting principles. Financial accounting is intended to give creditors, investors, and the government a fair and accurate representation of the firm’s transactions, revenues, profits, and losses. Costs are generally classified by function rather than by attributing activities to specific products or components of products. Therefore, additional analysis is required to assign costs to components.
The ICM method attempts to estimate the specific impacts of technologies and technology packages on indirect costs. While the ICM method is logically appealing, rigorous implementation is very difficult because it requires extensive knowledge of a firm’s operations and involves uncertainty about components such as warranty costs. RTI International and UMTRI (2009) and Rogozhin et al. (2010) provide descriptions and examples of how EPA estimates ICMs. As described by Helfand and Sherwood (2013), EPA relied on its engineers and scientists with expertise in automotive product development and production to develop the ICMs used in the Corporate Average Fuel Economy (CAFE)/GHG standards. The agency used two different experiments, a consensus approach (Rogozhin et al. 2010) and a Delphi-based approach (Helfand and Sherwood 2009) to develop the ICM.
The ICM estimation process applied by the Agency begins by associating each technology with one of four degrees of innovation and complexity (Rogozhin et al. 2010). The different levels of innovation and complexity are judged to affect research and development, corporate and dealer labor, and warranty components of the ICM differently. The levels of innovation used by the Agency were these:
- Incremental: compatible with existing core components of the automobile;
- Modular: changes core components but not their interaction;
- Architectural: changes interactions of components but not their fundamental function; and
- Differential: establishes new functions for components and changes their interaction.
The technologies were further classified by complexity. The panel of Agency experts was given four examples:
- Low complexity: passive aerodynamic drag reductions;
- Medium complexity: turbocharging with downsizing;
- High 1 complexity: hybrid electric vehicle; and
- High 2 complexity: plug-in hybrid electric vehicle.
In the final rule, the Agencies point out the limitations of both the RPE and ICM methods. Because the accounting methods of manufacturers differ and costs are generally not classified as direct or indirect, the estimation of RPEs requires judgment to allocate costs between the two categories. As described above and in the related references, the ICM method requires grouping technologies into different categories based on levels of innovation and complexity, which requires judgment and assumes that all technologies within a group have identical impacts on indirect costs. Expert judgment is also relied upon to estimate the impact values for each technology and cost component. Finally, ICM estimates have not been validated by directly measuring the indirect cost impacts of specific technologies.
The current rulemaking attempts to estimate transitional as well as long-run average indirect cost impacts. Higher short-run costs are represented by higher ICMs. The higher short-run ICM impact factors are intended to represent additional engineering effort that is initially required to integrate a new technology component into the overall vehicle system. Engine changes, for example, will require adjustments to transmissions, computer control algorithms, and other elements of the vehicle. Once a new technology has been integrated into a first-generation vehicle, the indirect cost multipliers assume the long-run values. The short run is defined as one production cycle lasting 4-5 years, and >5 years is the long run (Rogozhin et al. 2010). Short-run ICMs range from 4 percent to 18 percent higher than the long-run ICMs, with greater differences as the degree of complexity/innovation increases (Table 7.2).
Undoubtedly, different technologies and different design changes to vehicles affect indirect costs differently. It is therefore appropriate for the Agencies to work toward a methodology that assigns different indirect cost multipliers to different fuel economy and emissions technologies. As described above, the empirical basis for such multipliers is still lacking, and, since their application depends on expert judgment, it is not possible to determine whether the Agencies’ ICMs are accurate or not. In a presentation to the committee, the EPA presented evidence that on average the ICM method resulted in a ratio of total costs to direct manufacturing costs of approximately 1.5 (EPA 2014). The committee notes the seeming incongruity in the result where, on average, EPA estimates that the ratio of total costs to direct costs is about 1.5, but almost all of the individual ICMs are less than 1.5. The committee encourages the Agencies to continue research on ICMs with the goal of developing a sound empirical basis for their estimation. One possible method for developing such an empirical basis is through case studies to break down the costs associated with the specific steps required to integrate a new technology and trace its impacts on indirect costs. The committee notes that the specific values for the ICMs are critical since they may affect the overall estimates of cost and benefits for the overall standards and the cost effectiveness of the individual technologies.
|Complexity/Innovation||Near Term||Long Term|
SOURCE: EPA/NHTSA (2012, Table 3-1).
Economies of Scale
Scale economies are an important determinant of cost in the automobile industry. The NRC 2011 report asserted that scale economies would generally be reached at between 100,000 and 500,000 units per year, citing evidence from Martec (2008) and Honda (DOT/NHTSA 2009, 185). Husan (1997) cites estimates of scale economies for different production processes that range from 120,000 to 240,000 for powertrain manufacturing to 2,000,000 for foundry/forging and pressing. Even within a category, estimates of maximum scale economies vary widely, for example, from 120,000 to 1,000,000 for powertrain manufacturing. The exception is final assembly, where the range of estimates for volumes at which maximum scale economies are achieved is narrower: 100,000 to 300,000.
Economies of scale are often summarized by a scale elasticity that represents the relative reduction in cost with a 1 percent increase in scale as the optimum is approached. Data from Husan (1997) suggest a scale elasticity of approximately −0.1 for final assembly (Figure 7.3). The evidence also suggests that both optimal production volumes and scale elasticities will vary by manufacturing process. It should be noted that the Agencies do not separate cost reductions from increasing scale and cost reductions from learning by doing.
The Agencies use economies of scale to develop cost estimates of direct manufacturing costs. The incremental DMC of a technology is developed as discussed previously, assuming a North American production of 450,000 units. This is justified based on an FEV turbocharging and downsizing pilot study (FEV 2009) which used MY 2007 WardsAuto data to estimate U.S. domestic light-duty engine production volumes of 350,000 to 480,000 for moderate- to high- volume applications. The assumption of 450,000 units therefore is not production of a given technology by each manufacturer but by the entire United States or North American market (it is not clear which). The scale assumption of 450,000 units is used in such areas as the teardown studies supporting the rule and as an input to the BatPac model for battery costing, for example.
FIGURE 7.3 Estimates of scale economies in automobile manufacturing.
SOURCE: Data from Husan (1997).
As discussed further below, assuming such high volumes across the board assumes that all technologies will be implemented at large scales, and that economies of scale operate across multiple manufacturers, not simply within a single manufacturer. This volume assumption is problematic for technologies at low volume, like hybrid technologies, and for technologies with significant proprietary issues preventing suppliers from producing components at scale, such as for transmissions. The Agencies recognized that if their assumption of high volume is wrong, then actual costs will be higher than those predicted. Economies of scale are also connected to learning as defined by the Agencies and described below.
Learning by Doing
Numerous retrospective studies have documented how the price of a novel technology declines with cumulative production (Wene 2000). For many technologies price, p, has been shown to decrease as a function of the logarithm of cumulative production, X, known as a learning curve.
pX = p0X–a
The progress ratio (equal to 2–a) measures the relative cost for each doubling of production. For example, if a = 0.074, the progress ratio is 0.95, which implies a 5 percent reduction in manufacturing costs for every doubling of cumulative production. In theory, p0 is the price of the first unit of sales. However, in practice it is very difficult to know the price of the very first unit of any commodity, and the calibration of the learning curve can be very sensitive to the assumed initial conditions.
While numerous studies have estimated learning curves using historical data, there is no rigorous method for predicting learning curves for novel technologies in the absence of empirical data. Further complicating matters, observed price reductions are typically due to a combination of scale economies, exogenous technological change, and learning by doing in manufacturing processes. It is useful to separate the three since scale economies are a function of the current volume of production, rather than cumulative production, and exogenous technological change represents general advances in science and technology over time. Mathematically, it is straightforward to model separate effects for the three components, yet few empirical analyses have done so. If the overall combined effect is attributed to learning by doing, the result will be an overestimate of the potential for cost reduction via cumulative production (Nordhaus 2009).
The NRC Phase 1 report noted that cost estimates were usually based on three key assumptions (NRC 2011, 25):
- High-volume production (100,000 to 500,000 units);
- Learned production costs; and
- Competition in the supply chain from at least three global suppliers.
Under these assumptions, the Phase 1 committee concluded that it was not appropriate to use traditional learning curves to predict future cost reductions for technologies already in mass production. On the other hand, that committee also concluded that it would be appropriate to apply learning curves to cost estimates for truly novel technologies that did not reflect learning by doing. The Technical Support Document (TSD) for the 2017-2025 standards explains that the Agencies employ a nontraditional learning curve method using time- instead of volume-based learning and distinguishing between novel and established technologies (EPA/NHTSA 2012).
The Agencies’ rulemaking distinguishes two types of learning by doing, “steep” and “flat,” as well as the case of no learning (EPA/NHTSA 2012). Only newer technologies are subject to steep learning. The list of newer technologies includes only six items, subject to steep learning during the following periods:
- Air conditioner alternative refrigerant, MY 2016-2020;
- P2 hybrid vehicle battery pack components, MY 2012-2016;
- Electric/plug-in vehicle battery pack components, MY 2012-2025;
- Electric/plug-in vehicle battery charger components, MY 2012-2025;
- Stop-start, MY 2012-2015; and
- Lower rolling resistance tires, level 2, MY 2017-2021.
After their periods of steep learning, each of the above technologies converts to flat learning until 2025. Twenty-one technologies are subject to only flat learning, which begins in 2012 and continues through 2025. Only five technologies are considered to have no opportunities for future cost reduction via learning:
- Engine modification to accommodate low friction lubes;
- Engine friction reduction – levels 1 and 2;
- Lower rolling resistance tires – level 1;
- Low drag brakes; and
- Electric/plug-in vehicle battery charger installation labor.
Chapter 8 shows the various learning curves used by the Agencies and adopted by the committee in most cases to estimate future costs. The Agencies use a base year within their model of learning by doing. This base year is when the technology is considered “mature” and is when the Agencies assume volumes of 450,000, discussed previously. This base year is also used as the year off which the indirect cost is calculated for each technology. As discussed in Chapter 8, the concept of negative learning is used for low-volume technologies.
Steep learning follows a step function that produces a 20 percent reduction in cost after the first 2 years of implementation and another 20 percent reduction after another 2 years of production. Once two fast learning cycles are complete, new technologies follow the flat learning cycle. Flat learning begins with 5 years of cost reduction at 3 percent per year, followed by 5 years at 2 percent per year, and finally 5 years at 1 percent per year. For a technology subject to steep learning, this results in a more than 50 percent (52.8 percent) cost reduction after 19 years. Flat learning results in more than a 25 percent (26.2 percent) reduction in cost after 15 years. It is common practice in the automotive industry for OEMs to negotiate contracts with suppliers that stipulate annual cost reductions in the range of 1 percent to 3 percent, depending on the technology. Since such contracts are generally considered proprietary, it is difficult to document and measure this phenomenon. The Agencies apparently assign learning functions to technologies based on their expert judgment.
The Agencies’ use of the learning-by-doing concept is unconventional in that it is strictly a function of time rather than cumulative production. This formulation avoids the complication of endogeneity, the simultaneous determination of the cost of a technology and its adoption by manufacturers. It also sidesteps the problems of choosing a learning rate and an initial level of cumulative production. Within the TSD, the Agencies provide a detailed discussion of how learning is applied (EPA/NHTSA 2012). For example, as described above, when steep learning is in effect, then a 20 percent decrease in cost is assumed for every doubling in production volume. This is implemented by time rather than volume, however, so it implicitly assumes that a doubling of volume occurs every 2 years for technologies subject to steep learning. They also describe the difficulties of implementing volume-based learning within the models used by NHTSA (the Volpe model) and EPA (the OMEGA model), which would require the models to endogenously estimate a production volume and then apply a volume-based learning curve for any specific time period through an iterative feedback loop (EPA/NHTSA 2012). The committee appreciates this difficulty. On the other hand, this approach to learning allows a technology to accomplish very significant cost reductions even if its production volumes remain very low. Further, while the Agencies’ description of how they apply learning and how they are not separating learning into volume-based and time-based learning is clear, the figure displaying their conceptual model of learning (Figure 3-1 in the TSD) titles the curve as the volume-based learning curve with an x-axis labeled “cumulative production.” This conveys to the reader that learning is estimated from production volume. It would be helpful to rectify this situation in order to eliminate ambiguity about what was actually done.
The Agencies decided that learning should affect only direct manufacturing costs and not indirect costs, except for warranty costs. Their reasoning is that learning affects only direct manufacturing costs, except that warranty costs involve replacement of parts whose costs should also decrease with learning. It is difficult to evaluate this assertion due to the lack of empirical data.
If the rate of fuel economy improvement or GHG reduction required by the standards necessitates replacing capital investments before their normal depreciated lifetime, it may be appropriate to attribute the remaining amortized cost of the capital equipment, the “stranded capital,” to the replacement technology.3 The NRC Phase 1 report also noted that accelerated rates of redesign and technology adoption could demand more engineering resources than are available, potentially driving up labor costs.
3 Even this attribution is somewhat arbitrary in that there are generally many technologies available with which to improve fuel economy. The attribution of stranded capital cost to a specific “replacement” technology may make it a less economical choice than alternatives.
An estimate of the potential cost impact of stranded capital was made by FEV, Inc., (2011c) for the EPA. The study is not based on historical data nor is it an analysis of how the standards are likely to cause capital investments to be stranded in the future. Rather it is a “what if” assessment of the potential impact of stranded capital costs under a variety of different assumptions. The study estimates “the potential saddling of cost onto a new technology configuration as a result of the production equipment and/or tooling for the baseline configuration being abandoned before the planned fully depreciated life” (FEV, Inc. 2011a, 1-1). Six case studies were carried out: two conventional engines replaced by downsized, turbocharged, direct-injection engines; three upgraded transmissions; and one conventional V6 powertrain replaced by a power-split hybrid system. The additional cost of truncating the useful life of the productive capital after 3, 5, and 8 years was estimated. The results indicate relatively modest yet nontrivial cost impacts relative to the total cost of the technologies in question (Table 7.3). EPA and NHTSA (2012) provide descriptions of how the regulatory models apply costs representing the stranded capital cost of the replaced technology to the cost for producing the newly applied technology.
The FEV analysis does not address whether premature retirement is likely for specific technologies and how prevalent it might be. Although more analysis on this topic would improve the estimate of costs of the standards, the automobile industry in general is moving toward more rapid model updates due to consumer and competitive pressures (Oliver Wyman 2013; IHS Automotive 2013; Bloomberg 2014; Finlay 2014).
There are alternative approaches to using the engineering costs of prospective technologies for estimating the costs of the standards. Empirical economic analyses have used models of manufacturer and consumer behavior to infer the costs of CAFE standards. Jacobsen (2013) estimated the costs to those automakers that appeared to be substantially constrained by the standards during the late 1990s, when gasoline prices were low. He estimates the added costs of the last miles per gallon improvement to meet the standard will vary a good deal across the affected manufacturers, with direct costs to the domestic manufacturers ranging from $52 to $438 per car, and from $157 to $264 per truck. Using a different approach, Anderson and Sallee (2011) take the decision by companies to produce additional flexible fuel vehicles (FFVs) in lieu of reducing emissions in other ways as indicative of the cost of the fuel economy regulations at the margin. Their analysis covers the period from 1996 to 2006, and they find that the implied direct costs of the standards at the margin are quite low, between about $9 and $27 per vehicle. These studies examine a period when standards were relatively flat, which may not be as relevant to the future regulations, but the approach to estimating costs of regulatory compliance ex post is valuable. To date, there has been no careful study comparing ex-post economic studies to the ex-ante forecasted engineering costs for the same period.
There are also economic models that attempt to forecast the future effects, costs, and benefits of variations in the fuel economy regulations. These include Liu et al. (2014), Skerlos and Whitefoot (2012), and Jacobsen (2013). These studies use underlying engineering cost estimates but incorporate economic decisions in response to the regulations and thus the ability to capture more complete measures of the costs of the regulations. Such models and approaches can serve as important additional input in the regulatory analysis of the rules.
Technology Introduction in Vehicle Manufacturing
The timing for introducing new vehicle features and technologies is a significant strategic and competitive issue for automotive manufacturers. More rapid technology introduction will keep products current and more likely to be recognized as new and state of the art, resulting in higher consumer demand. Accelerating new product development provides
|Replaced Technology||New Technology||Potential Stranded Capital Cost per Vehicle (Productive Capital Stranded After X Years)|
|3 Years||5 Years||8 Years|
|Conventional V6||DSTGDI I4||$56||$40||$16|
|Conventional V8||DSTGDI V6||$60||$43||$17|
|Six-speed AT||6-speed DCT||$55||$39||$16|
|Six-speed AT||8-speed AT||$48||$34||$14|
|Six-speed DCT||8-speed DCT||$28||$20||$8|
|Conventional V6||Power-split HEV||$111||$79||$32|
SOURCE: FEV (2011c).
a market advantage to the manufacturer and, as discussed earlier, is a general industry trend in response to consumer and competitive pressures. In the context of fuel economy, new products are likely to have more recently developed technologies that reduce fuel consumption. Adopting new technologies, however, must be considered in the broader context of vehicle design and manufacturing. More substantial technological advancements, such as lightweighting or downsizing powertrains, may require major reengineering efforts and their introduction would be, therefore, timed with a new vehicle or new powertrain program. Other technologies that may require less integration engineering (such as low rolling resistance tires, low global warming potential (GWP), A/C systems, or other technologies) can be implemented during minor upgrades, which occur more frequently throughout the overall model life cycle.
Resource constraints that limit rapid implementation of new technology are costs, development and validation lead time, engineering resource availability, and financial risk. These same resource constraints also apply to suppliers who may be cooperating in new product development with the manufacturer. Technologies requiring high capital expenditures (e.g., tooling for powertrain or body components) tend to have longer product life cycles to help lower amortization rates. Much newer technologies, such as forming parts out of aluminum instead of steel, may require greater engineering lead time and resources than simply changing the shape of a part made of the same material. Additionally, the greater the change in the technology the greater the challenge of achieving both production and product validation, which can result in higher risk. Stop/start, safety issues with batteries, and utilization of die cast structural components instead of formed panels are examples of technological changes that potentially pose such risks.
Lightweighting Body––Near Term
As discussed in Chapter 6, while the majority of the car body today is made from steel, the body is also made from a mixture of materials that includes many grades of steel, grades of aluminum, a variety of composites, magnesium, and other materials. The industry, in general, is trending toward a broader distribution of materials that will bring about a significant transformation in the manufacturing process. The evolution of steel implementation in vehicles has been significant, from the frequent use of mild-strength steel in the early 1980s to today’s vehicles having over 50 percent of their body made from high-strength steel (HSS). It is not unusual to have over 15 grades of steel in a single vehicle body. Today, the implementation of steel in vehicles utilizes the following design tools:
- Engineering modeling software to simulate the forming of parts and assembly (welding) of the body. Modeling expertise and algorithms have to be modified as new material grades get introduced.
- Extensive cold-stamping infrastructure with stamping plants that in some cases have 30 or more press lines. Today’s state-of-the-art press lines (e.g., a servo press)
FIGURE 7.4 Example vehicle incorporating a combination of steel and aluminum types.
SOURCE: Audi MediaServices (n.d.).
As an example, a body structure is shown in Figure 7.4, composed of conventional and ultra-high-strength steel panels, both traditionally stamped and hot formed, as well as aluminum panels, extruded and rolled profiles, and die cast components.
can cost upwards of $70 million, whereas a tandem or transfer press line might cost $20 to $40 million apiece. A single high-volume vehicle may require 8 to 10 press lines to support it with stamped steel parts. Automation between the presses typically uses magnets, designed for steel components, to pick up parts.
- Tooling industry that produces hundreds of dies to fabricate parts for a single vehicle body. The automotive tool and die industry is steel-centric, and the die cost for an all-new vehicle body will cost $200 million to $400 million. There are also assembly tools (to weld) and checking tools (to measure parts) that are designed specifically for steel. Assembly makes occasional use of adhesive and other joining technologies but is dominated by spot welding, where the typical car has 4,000 or more spot welds in it. Most spot welding is performed by robots, and a new body shop will have 300 to 1,000 robots in it. The number of new joining technologies has been increasing in the body shop––for example, adhesive gluing, laser welding, and fastening (e.g., rivets). With adjustments, this general infrastructure can be used with new materials such as high-strength steel or even aluminum. However while the traditional steel body shop is dominated by spot welding, a body shop with other materials will likely be dominated by other joining technologies.
- The body-in-white is sent to the paint shop, where it is sealed, cleaned, and painted. Different materials paint differently and many paint shops rely on electrostatic paint, thus requiring a conductive metal surface. Paint shops last 20 or 30 years or longer and can cost over $500 million. They are subject to strict environmental laws, and upgrades are expensive and time consuming. Their design often influences the type of vehicle that can be made at the assembly plant.
As the materials being used in a vehicle’s body change, there can be significant impact on the design tools described above. Today, the industry has less experience designing, tooling, and joining aluminum components of a vehicle than it has with those of steel. A task as menial as moving a finished component around the factory floor requires a different engineering approach since the magnets used to move steel pieces do not function on aluminum pieces. However, some of these design and engineering challenges may be mitigated by OEMs moving towards an all-aluminum vehicle design. This can be seen today, as design, tooling, and stamping processes for aluminum are migrated from steel with a transitory learning curve.
While OEMs are already looking toward aluminum as the next step from steel, vehicle design further into the future may yield vehicle bodies that use multiple materials to achieve the fuel consumption and safety requirements of future standards. However, the industry’s limited experience base with and supply infrastructure for composites for the body presents a bigger challenge than the those associated with aluminum. As the number of materials in a body increases, joining and painting them becomes more difficult. Joining with adhesive has made significant advancements over the past 10 years, and it now provides a better joint than spot welding, although at higher cost. The paint shop can then be tuned to this new material, even if the paint shop was initially designed for steel. Other materials such as magnesium and composites are used only in selective locations while the body is still dominated by steel and aluminum.
Lightweighting Body––Long Term
The next migration to more mixed materials will occur as new assembly plants and paint shops get upgraded in the next 10 to 20 years. It is important to note in this context that the changeover to new materials almost never is a 1:1 substitution of a part or component. It entails, instead, a concept change for the functional design as well as for the manufacturing processes and systems used, as all the recent major changeovers reflect. While the opportunity for reducing mass with a mixed material vehicle is much better than with a monolithic design, mixing materials into the car poses new challenges not confronted by a steel-intensive vehicle.
The material sectors (steel, aluminum, magnesium, composites) are largely autonomous, with little to no collaboration between them. They are, individually, fierce competitors attempting to promote the use of their material in the car, often at the expense of another material. Consequently, there has been little to no cooperation between the material sectors. Challenges to mixing materials into the car, though significant, are not insurmountable; they just have not received the attention that individual materials have received. The principal challenges include the following:
- Coefficient of thermal expansion. Different materials expand and contract differently, and this can distort the body, especially when it is exposed to a heated paint station, for example.
- Joining different materials. Different materials require different joining methods. The methods of primary interest include spot welding, laser welding, friction stir welding, weld-bond adhesive, riveting, and fasteners. Each of these joining methods has an extensive research base, but except for spot welding, most experience is outside the auto industry. Organizations like EWI (Columbus, Ohio) have been researching these methods for years.
- Predictive modeling. Two aspects of CAE modeling include static and dynamic analysis. The static analysis principally looks at individual component forming and strength properties. Dynamic modeling attempts to evaluate a structure during a simulated crash. Both modeling methods become increasingly challenging with different materials and different joining methods.
- Supply chain. The maturity level of each supply chain is different. Steel is well established with many commodity materials. Aluminum is positioned to expand, but price volatility is a concern. Composites have a unique challenge: Many suppliers and extensive branding by company for their products makes material standardization, specification, and testing difficult.
The current U.S. knowledge base for automotive lightweighting materials, material properties, designing, forming, and joining is distributed throughout the vehicle manufacturers and Tier 1 and Tier 2 suppliers. Ford has pioneered the move to mass-produced aluminum structures with annual volumes of about 650,000 units for the 2015 MY aluminum body F-150, which is about 10 times the volume of niche-market luxury vehicles made of aluminum (SAE 2014). Ford worked with suppliers, including Alcoa and Novelis, to fine-tune the compositional specifications of the aluminum alloys for the F-150. In 1993, Ford developed the experimental aluminum-intensive Mercury Sable and introduced the first aluminum-intensive Jaguar XJ in 2003.
Changes to Powertrain (Adding Technologies, Downsizing, and Electrification)
Traditionally, development cycles in powertrains have been disengaged from vehicle model cycles. In part, this practice was due to the strategy of continuously making smaller incremental upgrades to the powertrains. In addition, manufacturers also prefer to limit the risk of combining the launch of a new vehicle with a new powertrain at the same time. Periods to completely retool for machines and assembly lines for engines and transmissions have generally been in the 10+ years range.
While powertrain design improvements have mainly been driven by performance improvements, a recent focus on fuel economy improvements, GHG emission reduction, and vehicle lightweighting has prompted radical changes in system and component designs as well as more dynamic and quicker-paced design implementation. This trend has had a significant impact on the manufacturing process of powertrain components, both in the initial steps of casting or forging and in the highly complex and highly automated machining and assembly processes.
In order to understand and assess these changes, a more differentiated view of the components of the powertrain makes sense. Essentially the powertrain comprises the engine, turbochargers, transmission, one or more drive shafts, differentials, and the related axle drives. Engines are traditionally manufactured in-house by the OEMs and very rarely shared by multiple OEMs in the form of collaborative projects. This indicates a lesser desire for standardization of engines among the different OEMs.
For example, downsizing the engine does not change the casting process of engine blocks and cylinder heads itself, but does necessitate a changeover of casting molds, cores, risers, and other components. Advanced engine concepts usually lead to more intricate and complex shapes and surfaces of the castings, however. These concepts require a more controlled flow of air and/or fuel, which leads to more complex and intricate casting geometries on cylinder heads. Other challenges in casting technologies can include mechanical core stability, riser geometries, and cycle times.
Within each OEM, considerable efforts are being made to standardize engine features to accommodate increased manufacturing flexibility. Designs having three, four, or six in-line cylinders with identical cylinder specifications can be machined and assembled on the same lines; this is the case for V6 and V8 engines as well. This standardization of engine features can even extend to the changeover capability between gasoline and diesel engines and will be facilitated by the foreseeable introduction of aluminum blocks for diesel engines. Flexible automation has made the youngest generation of production lines more capable of introducing even substantial product innovations at a quicker pace than before without a substantial cost penalty.
Attached to the engine, but fully separate from a manufacturing point of view, turbochargers have become important elements of powertrains with the use of sometimes up to three units on a single engine. Casting aside, manufacturing challenges arise when balancing the machined parts and the precision assembly of ball bearing systems for high rpm and low friction performance, especially with such advanced concepts as twin scroll or variable geometry systems. Turbochargers, or other supercharging systems, are generally manufactured by suppliers aiming to benefit from economies of scale and then delivered to the engine assembly sites.
Transmissions are often manufactured by suppliers, but many OEMs maintain in-house production capability for various volumes. For vehicle designs ranging from more basic and inexpensive to the more luxurious and sophisticated, transmissions with six or more speeds have become the standard. As discussed in Chapter 5, transmissions with eight, nine, or ten speeds will become increasingly prevalent, and continuously variable transmissions (CVTs) and dual-clutch transmissions will continue to play important roles. From a manufacturing viewpoint, all of these designs share a much higher package density than their predecessors. Consequently, machining and assembly tolerances have tightened significantly, leading to more process monitoring and control efforts on the production side and on overall integration.
The manufacture of the remaining components of the powertrain, drive shafts, differentials, and axles, which are mostly produced by suppliers instead of the OEMs, is impacted by fuel consumption and lightweighting targets. Issues here are predominantly caused by the need to reduce friction, which results in tighter machining, balancing, assembly tolerances, and bearing specifications, which all have a potential to increase manufacturing cost.
The Product Development Process
The timing for implementing change on the automobile has trended to different product life cycles for different subsystems on the car. Table 7.4 approximates the product development process (PDP) for different changes.
While automakers wish to keep the vehicle “fresh” for consumers, cost, lead time, resource availability, and risk limit the rate of change. Although accelerating the PDP will challenge these constraints, the availability of new technologies (e.g., safety technologies, electronics, and software) may warrant faster introduction than initially planned, and this typically increases cost. Traditionally, the rate of introducing new technology was embedded in an OEM’s overall strategic planning over several model life cycles and incorporated into the PDP and the manufacturing process. However, automakers have indicated that the steadily increasing mandates to improve fuel economy will necessitate more frequent product updates than their PDPs are designed to accommodate. Technologies that are ready for deployment cannot wait until the next upgrade, which might be as long as a 4-8 year cycle for upgrading the powertrains or models shown in Table 7.4, or the automaker risks not meeting the fuel economy target and consumer expectations. Therefore, a shorter deployment cycle is needed and a similar speeding up of the PDP is also necessary for the complete vehicle assembly. This also reduces the length of time available for engineering and predeployment testing, which means that automakers have relied more on accelerated laboratory testing and environmental test chambers and less on testing in the field. Introducing technology faster than planned will lead to issues related to stranded capital, which is discussed earlier in this chapter, and higher product development costs.
Standardized product design and process design principles are a hallmark of mass production to ensure competitive cost and quality performance. Major automakers have standardized procedures for both designing and producing vehicle subsystems. As with all standards, there has to be a balance between constraining innovation and achieving cost and quality objectives. For example, changing from steel to aluminum for a body part will probably require a change in standard practice as to how the part is designed (since aluminum cannot always bend into the same shape as steel), tooled, fabricated, and assembled. Deviations from standard practice result in higher development costs and risk.
Another industry trend inhibiting the rate of technology deployment is the globalization of platforms, discussed below. The reasons for designing different vehicles on a single platform are to reduce cost by keeping the overall volume of the part/component at high volumes, to decrease the need for engineering development resources, to improve quality, and to reduce risk. However, one consequence of this level of global standardization is that the cost to modify the platform now impacts the global vehicle design, not just those vehicles sold in a particular region. Another concern is that when a problem occurs, it occurs to a very large, potentially worldwide, population of vehicles, resulting in higher warranty or recall costs. Global platforms and standardized product and process development by the automakers will add complexity and cost to introducing new technology, and these impacts will weigh against the scale economies of having higher volumes on a smaller number of platforms. Another potential cost savings is that fewer platforms means that fewer redesigns will be needed to add technology across the fleet.
Table 7.5 summarizes the timing, cost, and integration into manufacturing of a variety of fuel economy technologies as estimated by this committee.
|Type of Change||Typical Frequency||Description (Engineering and Tooling)||Investment (Approximate Scale) (billion $)|
|New vehicle platform (clean sheet)||7 - 10 years||Total engineering for chassis and body and trim. Sporadically in conjunction with new powertrain development.||1.0 - 2.5|
|Major vehicle upgrade (on established platform)||6 - 8 years||Most of chassis may be carried over. Major body changes (with some carryover).||0.5 – 0.75|
|Minor vehicle upgrade (re-skin)||2 - 4 years||Minimal engineering with mostly cosmetic changes such as trim. Changes may be implemented that affect aerodynamics, rolling resistance or vehicle accessories,||0.25|
|New powertrain/transmission||10 - 15 years||All new engine or transmission design. Little to no carryover from previous generation.||0.75 - 1.5|
|Upgrade powertrain and transmission||4 - 8 years||Technology advancement such as changing 6-speed AT to 6-speed DCT or converting V6 to an I4 with turbo resulting in modifications of the assembly process and tooling||0.2 to 0.4|
SOURCE: FEV (2011c, Table 4-1).
|Technology||Description||Integration with Manufacturing||Time||Cost|
|Downsize and turbocharge||Engine downsize and addition of boosting||New engine re-design – requires new product development. (Turbo charger available from outsource)||Engine PDP – 2 to 3 years for engine development, 4 to 5 years including emissions certification||Expensive re-tooling requirements|
|Stop-start||Modifications to ICE||Minor modification of existing manufacturing line||Not significant, but best planned with new model launch||Not difficult to integrate into existing facility|
|Mild hybrid||Modifications to ICE, new batteries and power electronics||Minor modification of existing manufacturing line||Not significant, but best planned with new model launch||Not difficult to integrate into existing facility- relatively high development costs if volumes remain low|
|P2 hybrid||New chassis with existing or common body architecture, new batteries and power electronics||Unique chassis, torque converter removed and electric motor/generator installed in its place without changing the engine or transmission||All new engine and powertrain assembly||New and unique hybrid powertrain line operating outside of standard PDP - relatively high development costs if volumes remain low|
|PS hybrid||New chassis and powertrain with existing or common body architecture, including batteries and power electronics||Unique chassis and powertrain system to be integrated with traditional body assembly||All new engine and powertrain assembly||New and unique ICE and hybrid powertrain line operating outside of standard PDP - relatively high development costs if volumes remain low|
|PHEV 40 mi electric range||New chassis and powertrain with existing or common body architecture, including batteries and power electronics||Unique chassis and powertrain system to be integrated with traditional body assembly||All new engine and powertrain assembly||New and unique electrified powertrain line, perhaps with new ICEs, operating outside of standard PDP - relatively high development costs if volumes remain low|
|EV 75 mi range||New chassis and powertrain with existing or common body architecture, including batteries and power electronics||Unique chassis and powertrain system to be integrated with traditional body assembly||All new engine and powertrain assembly||New and unique electrified powertrain line operating outside of standard PDPP - relatively high development costs if volumes remain low|
|Body – Lightweighting|
|Steel to High-Strength Steel||HSS substitution for individual parts||HSS up to ~1000 MPa has minor issues and minimal increase in cost. Over ~1000 MPa requires change in forming process that will increase cost 50% or more.||Not a significant time impact in most cases. Supply chain availability concern.||Generally, cost premium increases as the steel strength increases. Significant cost increase for steel over 1000MPa due to hot forming. A hot formed part may cost two or more times a cold stamped part.|
|Steel to Aluminum (Closures)||Hood, deck lid, doors (aluminum hood and deck lids are common today. Aluminum doors and roof panels will be increasing||The challenges for converting to aluminum closure panels are more in design than manufacturing. Not a major manufacturing concern, but some changes with handling, dust/dirt and joining complexity (e.g., fasteners and adhesives) will add cost.||Minimal||Aluminum closures will cost about 25% more than steel equivalents. The cost premium is over $3.00 per pound saved.|
|Technology||Description||Integration with Manufacturing||Time||Cost|
|Steel to Aluminum (Body)||Aluminum body-in-white (traditional for lower volume and premium vehicles)||Significant changes required. The technology is known but the execution at scale production will be a challenge in stamping and body assembly.||Increased launch efforts as the industry learns to do this at volume. There are no significant timing hurdles for this except for the supply chain. Aluminum supply requires at least 30 months lead-time for a high volume vehicle.||Cost increase for converting from a steel to aluminum body-in-white (and closures) is in the order of $1.50 to $2.00 per pound weight reduction (high volume) – over $500 per vehicle.|
|Steel or Aluminum to Composites||Semi-structural components can be made with composites. A high-volume composite body-in-white is not considered viable due to cost and complexity until sometime past 2025/2030.||Complex. Integration hurdles in assembly (using adhesive) and new painting processes will be required that take a long time and large expense to convert.||Not expected until past 2030 for high volume. Will see composite panels for premium vehicles, usually with a metal subframe.||Expensive – several thousand dollars. New supply chain needed and existing infrastructure (presses, welders and tool making) require overhaul.|
|Aerodynamics||Passive and active technologies. Passive technologies can be readily implemented. Active ones generally require more time. Impact on styling will constrain options.||Minimal barriers||Can generally be implemented with a minor facelift.||Not significant|
|Low Rolling Resistance Tires||Integration with vehicle, road noise, etc. required.||Not an issue||Not an issue||Not significant|
|Electric Power Steering||Integration with vehicle||Not an issue||Not an issue||Not significant|
|Improved Accessories||Integration with vehicle||Not an issue||Not an issue||Not significant|
|Smart vehicle technology||Sensors and computers and communication devices require design into structure. System design for redundancy.||Electronics primarily from external supply chain with minimal constraints. Integration into vehicle requires some effort.||Not a significant issue (assuming electronics are available with necessary performance)||Not significant|
Growing Impact of Global Platforms on Vehicle Design Optimization
The automotive industry is experiencing significant growth in the use of global platforms as a way to reduce cost and increase engineering efficiency. It’s estimated that 30 percent of the vehicles produced in 2013 will be made on global platforms (Sedgwick 2014), and this number is continuing to increase. Recently, Ford announced that it will reduce its number of platforms from 15 down to 9 by 2015, and that those 9 platforms will account for 99 percent of the vehicles they manufacture. General Motors announced even more aggressive plans to reduce its number of platforms from 26 to 4, and Volkswagen has indicated that it plans to produce 40 different vehicle models globally on a single vehicle platform. Other companies such as Volvo, Nissan-Renault, BMW, and Toyota are in the process of executing their own versions of global platforms. Each manufacturer is developing its own methods and focusing on different areas of the vehicle to standardize. While potential benefits may be realized by increasing economies of scale and reducing cost and time to develop new model variations, there are also potential drawbacks such as the reduction of design flexibility and suboptimal vehicle design. Vehicle platforms are designed to accommodate certain component modifications but generally not new major technology advances. Despite these possible limitations, it is estimated that the top 10 global platforms will account for over 200 vehicle models by 2017.
There are a number of methods by which vehicle manufacturers are developing global platforms. For example, the modular transverse matrix (MQB), developed by Volkswagen, established a uniform mounting position for all engines and a
standardized front carriage structure, which allows it to produce models with different wheelbases and track widths on the same assembly line. Likewise, Nissan-Renault has developed its common module family (CMF), an architecture based on the assembly of compatible modules for the engine bay, cockpit, front underbody, rear underbody, and electrical/electronic architecture. Although these new methods do not necessarily fit the definition of a typical platform, they share the common goal of increasing commonality and standardization across vehicle models (increasing economies of scale and standardizing supply chains). The resulting reduction in unique engineering content and components across different models reduces cost while maintaining product choices for the consumer. In some cases, this can also result in over-engineered parts (designed for the greatest application load).
Global platforms are engineered to anticipate the introduction of various future modifications to the platform as technology advances or other changes are desired. The platform design might limit the ability to implement some changes but expedite the implementation of others that fit within the standard design, reducing development costs. However, since global platforms produce vehicles for different countries, they are designed to accommodate the most stringent requirements for powertrain performance, emission controls, and safety.
Vehicle manufacturers see global platforms as a way to maximize efficiency and reduce cost over a wide range of vehicle models. Nissan-Renault estimated that it will reduce its engineering cost by 30 to 40 percent and part cost by 20 to 30 percent by moving to its CMF system. Volkswagen has also estimated that its MQB could cut production cost by as much as 20 percent. The primary objective of the global platform is to reduce costs through economies of scale. Some regulatory technologies may benefit from platforms, whereas some may not because of differences in different countries. A vehicle platform is essentially the basic building block of components and systems from which a vehicle can be built. Increasing the number of vehicles shared on a single platform––which accounts for nearly half of the product development cost––can significantly reduce engineering cost. Similarly, purchasing and tooling cost can be reduced through economies of scale of component sharing and single sourcing of equipment.
There are several risks and potential limitations that vehicle manufacturers must manage when developing global platforms. With common systems and components shared across many vehicles, design flaws can significantly increase the exposure of a manufacturer’s vehicle fleet to recalls. Manufacturers will have to commit significant upfront investment, which may limit the flexibility to modify components and manufacturing processes over time. This could lead to stranded capital if the OEM is unable to amortize the initial investment due to the increased frequency of new design implementations. At the same time, platforms must have enough flexibility to differentiate the product from model to model for consumers to feel as if each product is different. This differentiation may be achieved by standardizing only systems that do not significantly affect styling. An important challenge is that components and structures may be overengineered for some vehicles such that they meet the requirements of all vehicles shared under the same platform. It is probable that shared components will be specified based on the most demanding and more expensive vehicles within the platform. Under such a condition some vehicles may incur increased cost and suboptimal design to meet the specifications of the platform. Each manufacturer must find a balance between the desire to increase economies of scale and the risk of overengineering its vehicles.
Finding 7.1 The committee conceptually agrees with the Agencies’ method of using an indirect cost multiplier instead of a retail price equivalent to estimate the costs of each technology since ICM takes into account design challenges and the activities required to implement each technology. In the absence of empirical data, however, the committee was unable to determine the accuracy of the Agencies’ ICMs. Due to this lack of empirical information, the committee generally assessed only the direct manufacturing costs for each technology. Historically, many studies have concurred on an average markup factor of 1.5.
Recommendation 7.1 The Agencies should continue research on indirect cost multipliers with the goal of developing a sound empirical basis for their estimation. One possible method for developing such an empirical basis is through case studies to break down the costs associated with the specific steps required to integrate a new technology and trace the impacts of a new technology on indirect costs. The committee provides an example earlier in the report where the committee used its knowledge to attempt to construct an ICM that demonstrates some of the insights that can be gathered from an empirical approach.
Finding 7.2 The Agencies’ use of the learning-by-doing concept is unconventional in that it is strictly a function of time rather than cumulative production. It allows a technology to accomplish significant cost reductions even if its production volumes remain very low. And in some of its presentations on how they approach learning, the Agencies convey the notion that cumulative production volume is used in the estimates. However, the committee appreciates the difficulties of implementing volume-based learning in the compliance models used by the Agencies.
Recommendation 7.2 The Agencies should make clear the terminology associated with learning and should assess whether and how volume-based learning might be better incorporated into their cost estimates, especially for low-
volume technologies. The Agencies should also continue to conduct and review empirical evidence for the cost reductions that occur in the automobile industry with volume, especially for large-volume technologies that will be relied on to meet the CAFE/GHG standards. The committee also recommends that, once the Agencies have decided on an implementation scenario, they should regard their production volumes as fixed and look for inconsistencies in their scenario with respect to cost reductions from learning (i.e., they have assigned a large cost reduction from learning for technologies with very low market penetrations).
Finding 7.3 The committee disagrees with the methodology of assigning direct manufacturing costs based on a 450,000 unit production volume since some technologies, especially those related to electric and hybrid vehicles, may take many years, if ever, to reach this market penetration. Additionally, since the 450,000 unit production volume applies to the entire United States, the smaller production volumes of each manufacturer would not bring the expected cost reductions.
Recommendation 7.3 For technologies such as electric vehicles, which may not reach 450,000 unit production volumes in North America in the time frame of the standards, the Agencies should use an appropriate, lower production volume to project direct manufacturing costs.
Finding 7.4 The product development process of auto manufacturers is accelerating for several reasons, one of which is to implement new technologies faster. Manufacturers traditionally bundle technologies and implement them in predetermined cycles. New regulations now encourage more rapid deployment as soon as a technology is ready to avoid falling behind on satisfying the steadily increasing regulations. More rapid deployment, although better for meeting regulations and responding to consumer demands, will increase stranded capital and incur higher product deployment costs.
Finding 7.5 The growth of global platforms used by automakers supports scale economies with shared components and engineering content. However, shared content may inhibit the use of lightweighting materials and fuel-economy technologies if they are not readily available in local markets, or because these technologies are not appropriate for the local market. Global platforms thus can be considered a constraint, especially in the short term, whereby supply chains are not fully developed, as well as an opportunity, especially in the long term, whereby scale economics can provide cost reductions. Since attributes that are unique to a local market (such as emission/fuel economy regulations or crashworthiness) may call for unique technologies, there is a risk that these may not be compatible with the global platform. In some cases, where regionally unique attributes are embedded in the global platform, a less-than-optimal solution (such as a subsystem engineered for the greatest load case) is likely to exist in the global platform that prevents an “optimal” design in every region.
Audi. 2014. Audi TT Coupe. Audi MediaServices. https://www.audi-mediaservices.com/publish/ms/content/de/public/grafiken/2014/03/20/TTC140091.html.
Blincoe, L. 2013. NHTSA’s Application of Indirect Costs. Presentation to the National Research Council Committee on Assessment of Technologies for Improving Light-Duty Vehicle Fuel Economy, Phase 2, March 27.
Bloomberg. 2014. VW plans Apple-like model updates. Automotive News, July 28. http://www.autonews.com/article/20140728/OEM03/140729897/vw-plans-apple-like-model-updates#.
Bussmann, W.V. 2008. Study of industry-average markup factors used to estimate retail price equivalents (RPE). Presentation to the National Research Council Committee on Assessment of Technologies for Improving Light-Duty Vehicle Fuel Economy, January 24.
DOT (Department of Transportation)/NHTSA (National Highway Traffic Safety Administration). 2009. Average fuel economy standards, passenger cars and light trucks, model year 2001: Final rule. 49 CFR Parts 523, 531, 533, 534, 536, and 537, Docket No. NHTSA-2009-0062, RIN 2127-AK29. Washington D.C., March 23.
Duleep, K.G. 2008. Analysis of technology cost and retail price. Presentation to the National Research Council Committee on Assessment of Technologies for Improving Light-Duty Vehicle Fuel Economy, January 25.
EPA (Environmental Protection Agency). 2012. Regulatory Impact Analysis: Final Rulemaking for 2017-2025 Light-Duty Vehicle Greenhouse Gas Emission Standards and Corporate Average Fuel Economy Standards. PA-420-R-12-016.
EPA. 2014. MTE Technologies and Costs: 2022-2025 GHG Emissions Standards. Presentation to the National Research Council Committee on Assessment of Technologies for Improving Fuel Economy of Light Duty Vehicles, Phase 2, July 31.
EPA (Environmental Protection Agency)/NHTSA (National Highway Traffic Safety Administration). 2012. Joint Technical Support Document: Final Rulemaking for 2017-2025 Light-Duty Vehicle Greenhouse Gas Emission Standards and Corporate Average Fuel Economy Standards, EPA-420-R-12-901, U.S. Environmental Protection Agency, Ann Arbor, MI.
FEV, Inc. 2009. Light-Duty Technology Cost Analysis Pilot Study. EPA-420-R-09-020, U.S. Environmental Protection Agency, Ann Arbor, MI.
FEV, Inc. 2011a. Light-Duty Vehicle Technology Cost Analysis, Mild Hybrid and Valvetrain Technology. EPA-420-R-11-023, U.S. Environmental Protection Agency, Ann Arbor, MI.
FEV, Inc. 2011b. Light-Duty Vehicle Technology Cost Analysis, Advanced 8-Speed Transmissions. EP-C-07-069, U.S. Environmental Protection Agency, Ann Arbor, MI.
FEV, Inc. 2011c. Potential Stranded Capital Analysis on EPA Light-Duty Technology Cost Analysis. EPA-420-R-11-019, U.S. Environmental Protection Agency, Ann Arbor, MI, November.
FEV, Inc. 2012. Light-Duty Vehicle Mass Reduction and Cost Analysis – Midsize Crossover Utility Vehicle. EPA-420-R-12-026, U.S. Environmental Protection Agency, Ann Arbor, MI, August.
Finlay, S. 2014. ’15 Toyota Camry Reflects Shorter Product Cycle. WardsAuto, October 1. http://wardsauto.com/vehicles-technology/15-toyota-camry-reflects-shorter-product-cycle.
Helfand, G., and T. Sherwood. 2013. Automobile industry retail price equivalent and indirect cost multipliers studies. Presentation to the National Research Council Committee on Assessment of Technologies for Improving Light-Duty Vehicle Fuel Economy, Phase 2, March 27.
Husan, R. 1997. The continuing importance of economies of scale in the automotive industry. European Business Review 97(1): 38-42.
IBIS Associates, Inc. 2008. Data Collection and Analysis: Vehicle Systems Costs. Report to the National Research Council Committee on Assessment of Technologies for Improving Light-Duty Vehicle Fuel Economy, December.
IHS Automotive. 2013. Ford Earns Top Marks in Polk Automotive Loyalty Awards; Volkswagen Named Most Improved. Polk, January 15. https://www.polk.com/company/news/ford_earns_top_marks_in_polk_automotive_loyalty_awards_volkswagen_named_mos.
Martec Group, Inc. 2008. Variable costs of fuel economy technologies. Presentation to the National Research Council Committee on Assessment of Technologies for Improving Light-duty Vehicle Fuel Economy, January 24.
NHTSA. 2012. Final Regulatory Impact Analysis: Corporate Average Fuel Economy for MY 2017-MY2025 Passenger Cars and Light Trucks. Office of Regulatory Analysis and Evaluation, National Center for Statistics and Analysis.
Nordhaus, W.D. 2009. The Perils of the Learning Model for Modeling Endogenous Technologies Change. Working Paper 14638, National Bureau of Economic Research, Cambridge, MA, January.
NRC (National Research Council). 2002. Effectiveness and Impact of Corporate Average Fuel Economy Standards. Washington, D.C.: The National Academies Press.
NRC. 2011. Assessment of Fuel Economy Technologies for Light-Duty Vehicles. Washington, D.C.: The National Academies Press.
Oliver Wyman. 2013. Automotive Value-Chain Structure Changes Massively. Automotive Manager: Trends, Opportunities and Solutions along the Entire Value Chain. Oliver Wyman, January. http://www.oliverwyman.com/content/dam/oliver-wyman/global/en/files/archive/2013/Oliver_Wyman_Automotive_Manager_I_2013.pdf.
Rogozhin, A., M. Gallaher, G. Helfand, and W. McManus. 2010. Using indirect cost multipliers to estimate the total cost of adding new technology in the automotive industry. International Journal of Production Economics 124: 360-368.
RTI International/UMTRI (University of Michigan Transportation Research Institute). 2009. Automobile Industry Retail Price Equivalent and Indirect Cost Multipliers. EPA-420-R-09-003, U.S. Environmental Protection Agency, Ann Arbor, MI.
SAE. 2014. Automotive Engineering, November 4.
Sedgwick, D. 2014. Carmakers bet on big global platforms to cut costs. Autonews, August 4. http://www.autonews.com/article/20140804/OEM10/308049988/carmakers-bet-on-big-global-platforms-to-cut-costs.
Vyas, A., D. Santini, and R. Cuenca. 2000. Comparison of Indirect Cost Multipliers for Vehicle Manufacturing. Center for Transportation Research, Argonne National Laboratory, Argonne, IL.
Wene, C.O. 2000. Experience Curves for Energy Technology Policy. International Energy Agency, OECD, Paris.
This page intentionally left blank.