3
Cost Estimation

INTRODUCTION

As a general rule, reduced fuel consumption comes at a cost. The cost may be due to more expensive materials, increased manufacturing complexity, or a tradeoff with other vehicle attributes such as power or size. In addition to increased manufacturing costs, other costs of doing business are likely to be affected to a greater or lesser degree. These indirect costs include research and development (R&D), pensions and health care, warranties, advertising, maintaining a dealer network, and profits. The most appropriate measure of cost for the purpose of evaluating the costs and benefits of fuel economy regulations is the long-run increase in retail price paid by consumers under competitive market conditions.1 The retail price equivalent (RPE) cost of decreasing fuel consumption includes not only changes in manufacturing costs but also any induced changes in indirect costs and profit.

Most methods for estimating manufacturing costs begin by identifying specific changes in vehicle components or designs, and they then develop individual cost estimates for each affected item. Most changes result in cost increases, but some, such as the downsizing of a V6 engine to an I4, will reduce costs. Component cost estimates can come from a variety of sources, including interviews of original equipment manufacturers (OEMs) and suppliers, prices of optional equipment, and comparisons of models with and without the technology in question. Total costs are obtained by adding up the costs of changes in the individual components.

An alternative method, which has only just begun to be used for estimating fuel economy costs, is to tear down a component into the fundamental materials, labor, and capital required to make it, and then to estimate the cost of every nut and bolt and every step in the manufacturing process (Kolwich, 2009). A potential advantage of this method is that total costs can be directly related to the costs of materials, labor, and capital so that as their prices change, cost estimates can be revised. However, this method is difficult to apply to new technologies that have not yet been implemented in a mass-production vehicle, whose designs are not yet finalized and whose impact on changing related parts is not yet known.

Differences in cost estimates from different sources arise in a number of ways:

  • Assumptions about the costs of commodities, labor, and capital;

  • Judgments about the changes in other vehicle components required to implement a given technology;

  • Definitions of “manufacturing cost” and what items are included in it; and

  • Assessments of the impacts of technologies on indirect costs.

This chapter discusses the premises, concepts, and methods used in estimating the costs of fuel economy improvement, highlights areas where differences arise, and presents the committee’s judgments on the key issue of RPE markup factors.

Information on costs can be used with assumptions on payback periods, discount rates, price of fuel, and miles driven per year to provide an estimate of the cost-effectiveness of technologies. However, the statement of task given to the committee is to look at the costs and fuel consumption benefits of individual technologies. Performing cost-effectiveness analysis was not included within the committee’s task and was not done by the committee. The accurate calculation of benefits of improved fuel efficiency is a complex task that is being undertaken by the National Highway Traffic Safety Administration (NHTSA) and the U.S. Environmental Protection Agency (EPA) as part of their current joint regulatory efforts.

1

As explained below, this rests on the premise that the global automotive market can be reasonably characterized (in economic jargon) as either a perfectly competitive or a monopolistically competitive market. Under such market conditions, products are sold, in the long run, at their average cost of production, including a normal rate of return to capital but no excess profits. Increased costs of production will therefore be fully passed on to consumers. The total cost of resources plus the consumers’ surplus loss due to the price increase is, to a close approximation, equal to the increase in long-run retail price times the volume of sales.



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



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 24
3 Cost Estimation INTRODUCTION component into the fundamental materials, labor, and capital required to make it, and then to estimate the cost of every As a general rule, reduced fuel consumption comes at nut and bolt and every step in the manufacturing process a cost. The cost may be due to more expensive materials, (Kolwich, 2009). A potential advantage of this method is that increased manufacturing complexity, or a tradeoff with total costs can be directly related to the costs of materials, other vehicle attributes such as power or size. In addition to labor, and capital so that as their prices change, cost estimates increased manufacturing costs, other costs of doing business can be revised. However, this method is difficult to apply to are likely to be affected to a greater or lesser degree. These new technologies that have not yet been implemented in a indirect costs include research and development (R&D), pen- mass-production vehicle, whose designs are not yet finalized sions and health care, warranties, advertising, maintaining a and whose impact on changing related parts is not yet known. dealer network, and profits. The most appropriate measure of Differences in cost estimates from different sources arise cost for the purpose of evaluating the costs and benefits of fuel in a number of ways: economy regulations is the long-run increase in retail price paid by consumers under competitive market conditions.1 The • Assumptions about the costs of commodities, labor, retail price equivalent (RPE) cost of decreasing fuel consump- and capital; tion includes not only changes in manufacturing costs but also • Judgments about the changes in other vehicle compo- any induced changes in indirect costs and profit. nents required to implement a given technology; Most methods for estimating manufacturing costs begin • Definitions of “manufacturing cost” and what items are by identifying specific changes in vehicle components or included in it; and designs, and they then develop individual cost estimates for • Assessments of the impacts of technologies on indirect each affected item. Most changes result in cost increases, costs. but some, such as the downsizing of a V6 engine to an I4, will reduce costs. Component cost estimates can come from This chapter discusses the premises, concepts, and methods a variety of sources, including interviews of original equip- used in estimating the costs of fuel economy improvement, ment manufacturers (OEMs) and suppliers, prices of optional highlights areas where differences arise, and presents the com- equipment, and comparisons of models with and without the mittee’s judgments on the key issue of RPE markup factors. technology in question. Total costs are obtained by adding up Information on costs can be used with assumptions the costs of changes in the individual components. o n payback periods, discount rates, price of fuel, and An alternative method, which has only just begun to be miles driven per year to provide an estimate of the cost- used for estimating fuel economy costs, is to tear down a effectiveness of technologies. However, the statement of task given to the committee is to look at the costs and fuel consumption benefits of individual technologies. Perform- 1 As explained below, this rests on the premise that the global automo - ing cost-effectiveness analysis was not included within the tive market can be reasonably characterized (in economic jargon) as either committee’s task and was not done by the committee. The a perfectly competitive or a monopolistically competitive market. Under accurate calculation of benefits of improved fuel efficiency such market conditions, products are sold, in the long run, at their average cost of production, including a normal rate of return to capital but no excess is a complex task that is being undertaken by the National profits. Increased costs of production will therefore be fully passed on to Highway Traffic Safety Administration (NHTSA) and the consumers. The total cost of resources plus the consumers’ surplus loss due U.S. Environmental Protection Agency (EPA) as part of their to the price increase is, to a close approximation, equal to the increase in current joint regulatory efforts. long-run retail price times the volume of sales. 24

OCR for page 24
25 COST ESTIMATION PREMISES ignition (CI) engine replaces a spark-ignition (SI) engine, or a set of low-rolling-resistance tires replaces In the committee’s judgment, the concept of incre - a set with higher rolling resistance. What matters is the mental retail price equivalent cost is most appropriate for change in RPE rather than the total RPE of the new the NHTSA’s purposes because it best represents the full, technology. This requires that an estimate of the RPE long-run economic costs of increasing fuel economy. The of the existing technology be subtracted from that of NHTSA has used the RPE method in its rulemakings on the new technology. fuel economy, for example in the final rule for model year • Equivalent vehicle size and performance. Estimat- 2011 light-duty vehicles (DOT/NHTSA, 2009, pp. 346- ing the cost of decreasing fuel consumption requires 352). Incremental RPE estimates are intended to represent one to carefully specify a basis for comparison. The the average additional price that consumers would pay for a committee considers that to the extent possible, fuel fuel economy technology implemented in a typical vehicle consumption cost comparisons should be made at under average economic conditions and typical manufac- equivalent acceleration performance and equivalent turing practices. These estimates are intended to represent vehicle size. Other vehicle attributes matter as well, long-run, high-volume, industry-average production costs, such as reliability, noise, and vibration. Ideally, cost and incorporating rates of profit and overhead expenses including fuel economy comparisons should be made on the basis warranties, transport, and retailing. Although learning and of no compromise for the consumer. Often there are dif- technological progress never stop, RPEs are intended to rep- ferences of opinion about what design and engineering resent costs after an initial period of rapid cost reduction that changes may be required to ensure no compromise for results from learning by doing.2 The committee uses the term the consumer. This, in turn, leads to differing bills of substantially learned as opposed to fully learned to convey materials to be costed out, which leads to significant that cost reductions due to increasing volumes may continue differences in incremental RPE estimates. to occur. RPEs are not intended to replicate the market price • Learning by doing, scale economies, and competition. of a specific vehicle or a specific optional feature at a specific When new technologies are first introduced and only time. The market price of a particular vehicle at a particular one or two suppliers exist, costs are typically higher time depends on many factors (e.g., market trends, marketing than they will be in the long run due to lack of scale strategies, profit opportunities, business cycles, temporary economies, as-yet-unrealized learning by doing, and shortages or surpluses) other than the cost of manufacturing limited competition. These transitional costs can be and retailing a vehicle or any given component. It is not ap- important to manufacturers’ bottom lines and should propriate to base a long-term policy such as fuel economy be considered. However, nearly all cost estimates are standards on short-run conditions or special circumstances. developed assuming long-run, high-volume, average The RPE concept, unfortunately, is not easy to apply. economic conditions. Typical assumptions include It raises a number of difficult questions about appropriate (1) high volume, (2) substantially learned compo- premises and assumptions and reliable sources of data. It nent costs, and (3) competition provided by at least frequently relies on the application of markup factors, which three global suppliers available to each manufacturer could vary depending on the nature of the technology and the (Martec Group, Inc., 2008a, slide 3). Under these as- basis for the original cost estimate. When an RPE markup sumptions, it is not appropriate to employ traditional factor is used, the definition of the cost to which it applies is learning curves to predict future reductions in cost as critical. Much of the disagreement over RPE multipliers can production experience increases. However, if cost be traced to inconsistent definition of the cost to be marked estimates are for novel technology and do not reflect up. The following are key premises of the committee’s learning by doing, then the application of learning application of the RPE method. curves as well as the estimation of scale economies may be appropriate. The use of such methods intro- • Incremental RPE. The relevant measure of cost is the duces substantial uncertainty, however, since there are change in RPE in comparison to an equivalent vehicle no proven methods for predicting the amount of cost without the particular fuel economy technology. More reduction that a new technology will achieve. often than not, a fuel economy technology replaces • Normal product cycles. As a general rule, premises in- an existing technology. For example, a 6-speed auto- clude normal redesign and product turnover schedules. matic transmission replaces a 5-speed, a compression- Accelerated rates of implementation can increase costs by decreasing amortization periods and by demanding more engineering and design resources than are avail- 2 Learning by doing represents the increase in productivity and decrease in able. Product cycles are discussed in Chapter 7. cost that occurs during a technology’s lifetime as a result of manufacturers’ • gaining experience in producing the technology. The impacts of learning on Purchased components versus in-house manufacture. costs can be represented as a volume-based learning where costs reductions Costs can be estimated at different stages in the manu- occur with increasing production levels or as a time-based learning where facturing process. Manufacturing cost estimates gen- cost reductions occur over time.

OCR for page 24
26 ASSESSMENT OF FUEL ECONOMY TECHNOLOGIES FOR LIGHT-DUTY VEHICLES erally do not include warranty, profit, transportation, fixed or overhead costs scale with variable costs is a key area and retailing costs, and may not include overhead or of uncertainty. research and development. Other estimates are based Although many components are manufactured in-house on the prices that original equipment manufactur- by OEMs, it is useful to distinguish between component ers (OEMs) would pay a Tier 1 supplier for a fully and vehicle assembly costs, because many manufacturers manufactured component.3 These estimates include purchase 50 percent or more of a vehicle’s components from the supplier’s overhead, profit, and R&D costs, but not suppliers. Transaction prices and price estimates from Tier 1 costs incurred by the OEM. RPEs attempt to estimate and Tier 2 suppliers are a major source of information on the the fully marked-up cost to the ultimate vehicle pur- costs of fuel economy technologies. chaser. A key issue for cost estimates based on Tier 1 Variable manufacturing costs of components include ma- supplier costs is the appropriate markup to RPE. This terials, labor, and direct labor burden (Table 3.1). Variable will depend on the degree to which the part requires manufacturing costs are sometimes referred to as direct man- engineering and design changes to be integrated into ufacturing costs, although when this term is used it typically the vehicle, and other factors. includes the depreciation and amortization of manufacturing • Allocation of overhead costs. Specific changes in equipment. Fixed costs of component manufacturing include vehicle technology and design may affect some of tooling and facilities depreciation and amortization associ- an OEM’s costs of doing business and not others. A ated with capital investments, manufacturing overhead (e.g., reduction in engine friction, for example, might not R&D, engineering, warranty, etc.), and profit (or return to affect advertising budgets or transportation costs. To capital). Unfortunately, terminology frequently differs from date there is a very limited understanding of how to one study to another. Total manufacturing costs (variable determine which costs of doing business are affected plus fixed) are equivalent to the price that a Tier 1 supplier by each individual technology and how to develop would charge an OEM for a finished component, ready for technology-specific markups (e.g., Rogozhin et al., installation. 2009). In theory, this approach has the potential to OEM or assembly costs include the variable costs of yield the most accurate results. However, in practice, materials, labor, and direct labor burden for vehicle assem- unambiguous attribution of costs to specific vehicle components is difficult. For example, despite extensive reliability testing, it is not possible to predict with certainty what impact a technology or design change TABLE 3.1 Components of Vehicle Retail Price will have on warranty costs. Furthermore, there are Equivalent (Long-Run Average Cost) significant cost components that cannot logically be Component Manufacturing (Subassembly) allocated to any individual component. Among these Variable component manufacturing costs are the maintenance of a dealer network and advertis- Materials ing. Yet, these costs must be paid. The RPE method Labor assumes that such costs should be allocated in propor- Direct labor burden Fixed component manufacturing costs tion to the component’s cost and that overall overhead Tooling and facilities depreciation and amortization costs will increase in proportion to total vehicle cost. R&D This will not necessarily produce the most accurate Engineering estimate for each individual item but is consistent with Warranty the goal of estimating long-run average costs. Other overhead Profit Vehicle Assembly and Marketing COMPONENTS OF COST Variable costs Although different studies describe and group the com- Assembly materials ponents of the retail price equivalent (long-run average cost) Assembly labor Direct labor burden in different ways, there are four fundamental components: Fixed costs (1) the variable costs of manufacturing components, (2) fixed Tooling and facilities depreciation and amortization costs of manufacturing components, (3) variable costs of Warranty vehicle assembly, and (4) fixed costs of vehicle assembly R&D and sale. The distinction between variable and fixed costs is Engineering Warranty not a sharp one, because many “fixed” costs scale to some Other overhead extent with production volume. In fact, the degree to which Transportation Marketing and advertising Dealer costs and profit 3 Tier 1 suppliers contract directly with OEMs, whereas Tier 2 suppliers Original equipment manufacturer profit contract with Tier 1 suppliers.

OCR for page 24
27 COST ESTIMATION bly. Fixed costs include facilities and tooling depreciation experience future cost reductions relative to current estimates and amortization, warranty, R&D, engineering, advertising, through learning by doing. Technologies such as cylinder dealer expenses and profit, transportation, and OEM return deactivation, camless valve trains, gasoline direct injection on investment (profit). The sum total of all costs, divided by with lean burn, turbocharging with engine downsizing, and the Tier 1 supplier price (or equivalent), is called the RPE hybrid systems from stop-start to full hybrids and plug-in markup. hybrids were all assumed to have progress ratios of 0.8 (i.e., a The costs of inputs to the production process can vary doubling of cumulative production would reduce costs by 20 over time. Some key components, such as electrical sys- percent). Diesel emissions control systems were assumed to tems, emissions controls, and hybrid vehicle batteries, use have smaller progress ratios of 0.9 (EPA, 2008a, Table 4.2-3). relatively expensive metals whose prices can be volatile, If supplier cost estimates truly represent fully learned significantly impacting manufacturing costs. The prices of costs (at full scale economies), then there is no justification many of these metals increased dramatically prior to the for assuming future learning by doing. The cost estimates global recession beginning in 2008, but have since returned made by Martec for the Northeast States Center for a Clean to previous levels. Most publicly available estimates of Air Future (NESCCAF), for example, were intended to re- technology costs do not explicitly reflect uncertainties about flect cost reductions by learning that would occur over the future commodity prices. period 2009-2011. In its study for the Alliance of Automobile Manufacturers, Martec intended that its cost estimates reflect full scale economies and full learning: “Martec specified FACTORS AFFECTING COSTS OVER TIME AND an extremely high annual volume target [500,000 units per ACROSS MANUFACTURERS year] specifically to drive respondents to report mature, Cost estimates for fuel economy technologies are typi- forward costs expected in the future with the impact of learn- cally presented as a single point estimate or as a range. In ing fully reflected” (Martec Group, Inc., 2008b, p. 7). But fact, costs will vary over time and even across manufacturers Martec identifies two sources of learning: (1) improvement owing to technological progress, experience (learning by in manufacturing productivity, largely as a result of pro- doing), prices of commodities, labor and capital, and the duction volume; and (2) changes in system design. Martec nature of the vehicles manufactured. considered the latter to be technological innovations that would change the system architecture and thus the technol- ogy itself, requiring new cost estimates. Thus, the learning Economies of Scale considered by Martec in its estimates is based on the belief Scale economies describe the tendency for average manu- that the Tier 1 and Tier 2 suppliers would implicitly include facturing costs to decrease with increasing volume, as fixed learning effects of the first type in their high-volume cost costs are distributed over a greater number of units produced. estimates, and would exclude learning of the second type. The automobile industry is characterized by large economies In its 2011 corporate average fuel economy (CAFE) rule- of scale. Although sources differ, full scale economies are making, the NHTSA recognized two types of learning by generally considered to be reached at between 100,000 and doing: “volume-based” learning and “time-based” learning. 500,000 units per year. Martec Group, Inc. (2008a), for ex- Neither is based on cumulative production, as is much of the ample, asserts that production efficiencies are maximized at literature on learning by doing. DOT/NHTSA (2009, p. 185) 250,000 to 300,000 units. Honda cited a maximum efficiency judged that a first cycle of volume-based learning would of 300,000 units in its comments to the DOT/NHTSA (2009, occur at a volume of 300,000 units per year and that costs p. 185). would be reduced by 20 percent over low-volume estimates. A second learning threshold was set at 600,000 units per year, at which point a second cost reduction of 20 percent Technological Progress and Learning by Doing was taken. No further volume-based learning was assumed. Although cost estimates are generally premised on full The NHTSA applied this procedure to only three technolo- scale economies and fully learned technologies, both the gies in its 2011 rule: integrated starter generator, two-mode EPA and the NHTSA believe that not all Tier 1 supplier or hybrid, and plug-in hybrid. piece cost estimates represent fully learned technology costs. DOT/NHTSA (2009, p. 188) also applies time-based or In their view, learning curves should be applied for the more year-over-year learning by doing to widely available, high- novel technologies not in widespread use today.4 The EPA volume, mature technologies. Either time-based or volume- listed 16 advanced technologies that, in its judgment, would based learning, but not both, is applied to a particular technol- ogy. Time-based learning is applied at the rate of 3 percent per year in the second and all subsequent years of a technology’s 4 The EPA generally does not use typical continuous learning curves application. but instead stepwise learning as a function of time, rather than cumulative production. Usually, costs are assumed to decrease by 10 percent after the The use of learning curves poses a dilemma. On the one first year of production, and by another 10 percent after the second year, hand, there is no rigorous method for determining how much and then to remain constant.

OCR for page 24
28 ASSESSMENT OF FUEL ECONOMY TECHNOLOGIES FOR LIGHT-DUTY VEHICLES and how rapidly a specific technology’s costs can be reduced TABLE 3.2 Vehicle Classification by the National by learning by doing.5 On the other hand, the phenomenon Highway Traffic Safety Administration of learning by doing is widely and generally observed in the Passenger Cars manufacturing of new technologies (e.g., Wene, 2000). This Subcompact does not mean that no learning should be assumed. Rather, Subcompact performance learning curves should be applied cautiously and should Compact reflect average rates of learning based on empirical evidence Compact performance from the motor vehicle industry. Expert judgment should be Midsize Midsize performance used to determine the potential for learning, depending on Large the nature of the technology in question. Large performance Light Trucks Vehicle Type or Class Minivans Small SUV/pickup/van The costs of fuel economy technology also vary across Midsize SUV/pickup/van vehicle classes. To a large extent this is a function of vehicle Large SUV/pickup/van size and power. For example, an eight-cylinder engine has twice as many valves as a four-cylinder, and so the costs of valve train technologies will be higher. When technologies, such as turbocharging, increase the power output per unit (minivans) and footprint size (sport utility vehicles [SUVs], of displacement and thereby enable engine downsizing at pickups, and vans). constant performance, the starting cylinder count can affect Although classification can improve the accuracy of cost the options for downsizing. In general, an eight-cylinder estimates, there is no perfect classification system, and there engine can be replaced by a smaller six-cylinder engine of will always be some heterogeneity within a class. equivalent performance without additional costs for mitigat- ing vibration. Downsizing a four-cylinder to a three-cylinder METHODS OF ESTIMATING COSTS would require significant modifications to offset increased vibration, and this might even rule out reducing the cylinder As a generalization, there are two basic methods of cost count. Since most of the cost savings from downsizing accrue estimation. The first and most common is to obtain esti - from reducing the number of cylinders, technologies that mates of the selling prices of manufactured components. enable engine downsizing will be relatively more expensive The second is to tear down a technology into its most basic for four-cylinder engines. Since different vehicle classes materials and manufacturing processes and to construct a have different distributions of cylinder counts, the costs of bottom-up estimate by costing out materials, labor, and certain technologies should be class-dependent. As another capital costs for every step. Both methods ultimately rely example, the cost of a 1 percent weight reduction by material heavily on the expertise and the absence of bias on the cost substitution will depend on the initial mass of the vehicle. estimator’s part. National Research Council (2002) did not vary technolo- gy costs by vehicle class. The NHTSA’s Volpe model’s algo- Estimation Using Supplier Prices for Components, or rithm, however, operates at the level of make, model, engine, “Piece Costs” and transmission configuration. Some technology costs are scaled to the specific attributes of each vehicle. Other costs The supplier price method relies on comparing an es- are class-dependent. In its final rule for 2011, DOT/NHTSA timate of the price that a Tier 1 component manufacturer (2009, p. 165) specified eight passenger car classes and four would charge an OEM for a reference component to an esti- light truck classes (Table 3.2). Passenger cars were divided mate of the price that it would charge for an alternative that into size classes on the basis of their footprint. Each class delivered reduced fuel consumption. In the past, information was divided into a standard and high-performance class on on the prices that manufacturers pay to Tier 1 suppliers for the basis of class-specific cut-points determined using expert components has come from a variety of sources, including judgment. This reflects the NHTSA’s view that in addition to the following: size, performance is the key factor determining differences in technology applicability and cost. The classification of light • The NRC (2002) report on the CAFE standards; trucks was based on structural and design considerations • The NESCCAF (2004) study on reducing light-duty vehicle greenhouse gas emissions; • The California Air Resources Board study in support 5 Not only the progress ratio, but also the assumed initial cumulative of its greenhouse gas regulations; production (or threshold volume) strongly influences estimated future cost reductions. Numerous after-the-fact estimations of progress ratios are • The study by Energy and Environmental Analysis, Inc. available. However, in general, there is no scientific method for deciding (EEA, 2006) for Transport Canada; on these parameters ex ante.

OCR for page 24
29 COST ESTIMATION • Confidential data submitted by manufacturers to the and (2) fixed or burden costs. Estimating costs to the con- NHTSA in advance of rulemakings; and sumer (analogous to the retail price equivalent) requires • Confidential data shared by manufacturers in meetings additionally estimating the OEM’s amortized costs, as well with the NHTSA and the EPA in 2007. as other costs and profit. Dealers’ costs are added to the manufacturer’s cost plus profit to obtain the consumer’s Component cost estimates can be obtained from discus- cost (Figure 3.1). As the NHTSA report is careful to point sions with suppliers or OEMs, from published reports, or by out, estimating costs “is not an exact science” but rather one comparing the prices of vehicles with and without the com- strongly dependent on the expertise and judgment of the ponent in question (Duleep, 2008), bearing in mind that costs estimators at every step. and market prices may differ significantly. The NHTSA also The teardown method was applied by Kolwich (2009) to receives cost estimates in the form of confidential data sub- estimate the incremental manufacturing cost of a downsized mitted by manufacturers. Depending on how fuel economy 1.6-liter, four-cylinder, stoichiometric direct injection, turbo- technologies are defined, estimates for more than one com- charged engine versus a 2.4-liter, four-cylinder, naturally ponent may be involved. Given a supplier price estimate, a aspirated base engine. The study did not attempt to estimate markup factor is applied to estimate the RPE. A single markup the markup from manufacturing costs to RPE. Rather, the factor is often used for all components, but different markups cost estimated is equivalent to the price that a Tier 1 supplier may be used according to the nature of the component. The would charge an OEM for the fully manufactured engine. key issues are, therefore, the accuracy of the supplier price Unit costs are composed of direct manufacturing costs (mate- estimates and the accuracy of the markup factor(s). rial + labor + fixed manufacturing costs) + “markup costs” First at the request of NESCCAF (2004) and later at the (scrap + overhead + profit) + packaging costs (Figure 3.2). request of the Alliance of Automobile Manufacturers, Martec Manufacturing costs are estimated in a series of highly Group, Inc. (2008b) estimated the variable (or manufactur- detailed steps based on what is learned in disassembling the ing) costs of fuel economy technologies based on the bill of technology. Both the new and the base technologies must be materials (BOM) required. The term materials as used in torn down and costed in order to estimate the difference in the Martec studies refers to manufactured components sup- cost. First, the technology to be evaluated is identified and plied by Tier 1 and Tier 2 suppliers. The direct and indirect defined. Next, candidate vehicles for teardown are identified changes in vehicle components associated with a particular (this limits the analysis to technologies already in produc - technology were determined in discussions with engineering tion). A pre-teardown, high-level bill of materials (consist- consultants and OEM engineers. The Tier 1 and Tier 2 sup- ing of subsystems and components) is then created, subject pliers were the primary sources of information on the costs to amendment, as discoveries might be made during the of manufactured components required to implement the fuel teardown process. At that point, the actual teardown process economy increases (Martec Group, Inc., 2008b, p. 7). begins. During the teardown, all of the processes necessary for assembly are identified and recorded, and every compo- nent and the material of which it is made are identified. The Teardown or Bottom-Up Estimation data generated in the disassembly are then reviewed by a A change in the design and content of a vehicle induces team of experts. Following the review, the components are changes in the materials of which it is made, the quantity torn down and assembly processes are identified, as is each and types of labor required to construct it, and changes in the and every piece of each component. A worksheet is then capital equipment needed to manufacture it. Such estimates constructed for all parts, containing all cost elements. Parts not only are time-consuming but also require analysts with with high or unexpected cost results are double-checked, a thorough knowledge of and experience with automotive and then entered into a final spreadsheet in which they are manufacturing processes. totaled and formatted. Bottom-up cost estimation methods have been used by the Once manufacturing costs have been estimated, a markup NHTSA for assessing the impacts of safety regulations. For reflecting all other costs of doing business is typically applied example, in a study of air bag costs, an NHTSA contractor to estimate the long-run cost that consumers will have to used a teardown method to identify all components of 13 pay. Applying this markup was outside the scope of the FEV existing air bag systems. This study (Ludtke and Associates, (2009) study but was included in the Ludtke and Associates 2004) is described in Appendix F. The contractor analyzed (2004) study. Estimates of the consumer’s cost of curtain each part or assembly and identified each manufacturing air bag systems installed in five different vehicles from the process required for fabrication, from raw material to fin- Ludtke and Associates study are shown in Table 3.3. Although ished product. The analysis identified parts purchased from costs vary, it is clear that Ludtke and Associates used the same suppliers as well as parts made in-house. Process engineers markup factors for Tier 1 manufacturers’ markups over their and cost estimators then carried out a process and cost analy- direct costs (24 percent), OEM markups (36 percent), and sis for each part and assembly. Two costs were developed: dealer markups (11 percent). These markups result in multi- (1) variable costs associated with the actual manufacturing pliers for the consumer’s cost over the Tier 1 supplier’s cost of

OCR for page 24
30 ASSESSMENT OF FUEL ECONOMY TECHNOLOGIES FOR LIGHT-DUTY VEHICLES FIGURE 3.1 Determination of manufacturing and consumer cost. SOURCE: Ludtke and Associates (2004), p. B-10. Figure 3-1.eps low-resolution bitmap Net Component/Assembly Cost Impact To OEM = + + Total Manufacturing Mark -up Cost Packaging Cost Cost = = Raw Material In Process Scrap End Item Scrap Material + Purchased Parts + Sell, General & Administrative Direct Labor Costs + Indirect Labor Labor Maintenance, Repair, Other Profit + Fringe + Engineering, Design Manufacturing and Testing/R&D Overhead/Burden FIGURE 3.2 Unit cost model. SOURCE: FEV, Inc. (2009) (FEV.com), Figure 5. Figure 3-2 new editable

OCR for page 24
31 COST ESTIMATION TABLE 3.3 Estimated Consumer Cost (2003 dollars) for Installed Air Bag Systems and Markups Mercury Montereya Item VW Jetta Toyota Camry Cadillac CTS Jeep Grand Cherokee Material $30.04 $27.45 $48.46 $69.88 $54.43 Direct labor $11.11 $20.54 $16.54 $37.62 $17.68 Direct labor burden $22.59 $34.40 $24.61 $55.91 $23.93 Tier 1 markup $15.40 $19.89 $21.93 $39.66 $23.21 Manufacturer markup $28.49 $36.82 $40.15 $73.11 $42.93 Dealer markup $11.84 $15.30 $16.69 $30.38 $17.84 Consumer’s cost $119.47 $154.40 $168.38 $306.55 $180.02 Variable cost $63.74 $82.39 $89.61 $163.41 $96.04 Variable manufacturing cost $79.14 $102.28 $111.54 $203.07 $119.25 Markup Tier 1 cost 1.51 1.51 1.51 1.51 1.51 Markup variable manufacturing cost 1.87 1.87 1.88 1.88 1.87 Tier 1 markup 24.2% 24.1% 24.5% 24.3% 24.2% OEM markup 36.0% 36.0% 36.0% 36.0% 36.0% Dealer markup 11.0% 11.0% 11.0% 11.0% 11.0% NOTE: Original equipment manufacturer (OEM) manufacturing costs (2003$) per vehicle—head protection air bag systems (curtain-type system without a torso airbag already installed in vehicle). aCost estimates for the Mercury Monterey are substantially higher than those for the other vehicles. Ludtke and Associates (2004) do not offer an explana - tion for the design differences that account for the higher cost. SOURCE: Ludtke and Associates (2004). 1.51 (1.36 × 1.11 = 1.51), and for the consumer’s cost over the The FEV teardown study (FEV, 2009; Kolwich, 2009) al- direct variable costs of manufacturing (“Total Manufacturing lows total manufacturing costs to be broken down by engine Costs” minus “Manufacturing Overhead Burden” in the FEV subsystem as well as cost component. Figure 3.3 shows the [2009] study; see Figure 3.2 above) the component of 1.87 incremental manufacturing costs by cost component. The (1.24 × 1.36 × 1.11 = 1.87). The costs shown in Table 3.3 are largest single component of the $537.70 total is material in 2003 dollars and assume a manufacturing scale of 250,000 ($218.82), followed by manufacturing burden ($154.24), units per year for the air bags. labor ($72.58), corporate overhead ($33.96), profit ($33.12), While Ludtke and Associates (2004) use a markup factor engineering and R&D ($12.36), and scrap ($11.72). The of 1.24 for direct manufacturing costs, Kolwich (2009) uses total markup on manufacturing costs is just over 20 percent. markup factors ranging from 10.3 percent to 17.7 percent, Figure 3.4 shows the same total cost broken down by engine depending on the complexity of the component (Table 3.4). subsystem. By far the largest components are the induction Note that the Kolwich rates do not include manufacturing air charging system ($258.89) and the fuel induction system overhead whereas the Ludtke rates do, and thus the former ($107.32). Cost savings occur in counterbalance ($35.95) should be higher. and intake systems ($12.73). TABLE 3.4 Total Manufacturing Cost Markup Rates for Tier 1 and Tier 2/3 Suppliers End Item Scrap SG&A Profit ED&T Total Markup Markup Markup Markup Markup Primary Manufacturing Equipment Group (%) (%) (%) (%) (%) Tier 2/3—large size, high complexity 0.7 7.0 8.0 2.0 17.7 Tier 2/3—medium size, moderate complexity 0.5 6.5 6.0 1.0 14.0 Tier 2/3—small size, low complexity 0.3 6.0 4.0 0.0 10.3 Tier 1 complete system/subsystem supplier (system/subsystem integrator) 0.7 7.0 8.0 6.0 21.7 Tier 1 high-complexity-component supplier 0.7 7.0 8.0 4.0 19.7 Tier 1 moderate-complexity-component supplier 0.5 6.5 6.0 2.5 15.5 Tier 1 low-complexity-component supplier 0.3 6.0 4.0 1.0 11.3 SOURCE: Kolwich (2009), Table 2.

OCR for page 24
32 ASSESSMENT OF FUEL ECONOMY TECHNOLOGIES FOR LIGHT-DUTY VEHICLES FIGURE 3.3 Incremental cost of turbocharged, downsized, gasoline direct-injection I4 engine broken down by cost category. SOURCE: Kolwich (2009), Figure 19. Figure 3-3.eps bitmap FIGURE 3.4 Incremental cost of turbocharged, downsized, gasoline direct-injection I4 engine broken down by engine subsystem. SOURCE: Kolwich (2009), Figure 19. Figure 3-4.eps bitmap RETAIL PRICE EQUIVALENT MARKUP FACTORS which may or may not induce other changes in the cost of manufacturing. These integration costs can be substantial for Markup factors relating component costs to RPE add major components, such as engines, or when, as is more often significantly to the estimated costs of automotive technolo- the case than not, many changes are made simultaneously. gies and are the subject of continuing controversy. The cost There are also indirect costs for research and development, of making and selling light-duty vehicles is not limited to administrative overhead, warranties, and marketing and ad- the manufacture of components and their assembly. Even vertising. Vehicles must be transported to dealers who have for a single technological or design change, cost impacts are their own labor, material, and capital costs. All of these addi- generally not limited to the component that is changed. Engi- tional costs are represented by RPE markup factors. neering expertise must be supplied to design these changes,

OCR for page 24
33 COST ESTIMATION Existing RPE Markup Factors OEM was also supported by Bussmann (2008), who cited a 2003 study of the global automotive industry by McKinsey For the automobile industry, there is a reasonable consen- Global Institute that produced a markup factor of 2.08, and sus on the ratio of total costs of doing business to the cost his own analysis of Chrysler data for 2003-2004 that pro- of fully manufactured components (the price that a Tier 1 duced factors of 1.96 to 1.97. Information supplied by EEA, supplier would charge an OEM). This average RPE markup Inc., to the committee (Duleep, 2008) implies higher markup factor is approximately 1.5, according to the available evi- factors: 2.22 to 2.51 for the markup over variable costs and dence, reviewed in detail in Appendix F of this report. Part 1.65 to 1.73 for the markup over Tier 1 supplier costs. of the disagreement over the size of the RPE markup factor Average RPE factors can be inferred by costing out all arises from the difference between the variable costs versus the components of a vehicle, summing those costs to obtain the variable plus fixed costs of a manufactured component. an estimate of OEM Tier 1 costs or fully burdened in-house An appropriate RPE markup over the variable (or direct) manufacturing costs, and then dividing the sum into the costs of a component is approximately 2.0 (Bussmann and selling price of a vehicle. The committee contracted with Whinihan, 2009). Part of the disagreement arises over the IBIS Associates (2008) to conduct such an analysis for two difficulty of attributing indirect and other fixed costs to a high-selling model-year 2009 vehicles: the Honda Accord particular vehicle component. sedan and the Ford F-150 pickup truck. For the Honda, Every fuel economy technology does not affect fixed or the RPE multipliers were 1.39 to market transaction price indirect costs in the same way. Some costs may be affected by and 1.49 to manufacturer’s suggested retail price (MSRP). engineering and design changes to decrease fuel consump- The multiplier to dealer invoice cost is 1.35, implying that tion; others may not. This can have a very large impact on dealer costs, including profit, amount to about 4 percent of the appropriate RPE of a given fuel economy technology. manufacturing costs, not considering any dealer incentives Some studies use a single, average RPE markup factor (e.g., provided by OEMs. For the Ford F-150, the RPE multipliers NRC, 2002; Albu, 2008; DOT/NHTSA, 2009), while others were 1.52 for market price and 1.54 for MSRP. The markup attempt to tailor the markup to the nature of the technology factor for dealer invoice is 1.43, implying that dealer costs (Rogozhin et al., 2009; Duleep, 2008). The problem of how and profit amount to about 9 percent of total manufacturing best to attribute indirect and fixed costs to a specific change costs, not including any possible OEM incentives to dealers. in vehicle technology remains unresolved. Existing estimates of the RPE markup factor are similar The EPA Study on RPE Factors and Indirect Cost when interpreted consistently. Vyas et al. (2000) compared Multipliers their own markup factors to estimates developed by EEA, Inc., and Chrysler. Unfortunately, differences in the defini- Concerns with the Existing RPE Method tions of categories of costs preclude precise comparisons. Vyas et al. concluded that an appropriate markup factor Objections have been raised with respect to the use of a over the variable costs of manufacturing a motor vehicle was single RPE markup factor for components manufactured by 2.0. The Vyas et al. (2000) report also summarized the cost Tier 1 suppliers and sold to OEMs. The EPA has pointed out methodology used by EEA, Inc., in a study for the Office that not all technologies will affect indirect costs equally, and of Technology Assessment (OTA, 1995). Vyas et al. (2000) it has proposed to investigate technology-specific markups, concluded that the markup over variable manufacturing by attempting to identify only those indirect costs actually costs used in that study was 2.14, while the markup over affected by each technology (EPA, 2008b). In a similar vein, outsourced parts (e.g., purchased from a Tier 1 supplier) was the importance of “integration costs” has been cited as a fac- 1.56 (Table 3.5). tor that would justify different markup factors for different A markup factor of 1.5 was also used by the NHTSA technologies (Duleep, 2008).6 Because a vehicle is a system, (2009, p. 173) in its final fuel economy rule for 2011. A it is almost always the case that the design of one part affects somewhat lower RPE markup factor of 1.4 was used by the others. Manufacturers cannot simply buy a list of parts and NRC (2002) and Albu (2008), while the EPA has used a markup of 1.26 (EPA, 2008a). 6 Duleep (2008) recommends using different markup factors for differ- The use of a markup of approximately 2 over the direct ent kinds of components to account for differences in the cost of integrating manufacturing costs of parts manufactured in-house by an components into the overall vehicle design. For parts purchased from Tier 1 suppliers, Duleep recommends a range of markup factors from 1.45 to 1.7, depending chiefly on integration costs. As an example, Duleep presented to the committee an estimated markup factor of 1.72 for injector, pump, and rail TABLE 3.5 Comparison of Markup Factors costs for a stoichiometric GDI engine. This is at the high end of his markup Markup Factor for ANL Borroni-Bird EEA range, reflecting the greater integration costs for engine technologies. Duleep (2008) proposed using judgment to divide technologies into three In-house components 2.00 2.05 2.14 groups. He recommended a markup factor of 1.7 for technologies requiring Outsourced components 1.50 1.56 1.56 extensive integration engineering, 1.56 for those having average integration SOURCE: Vyas et al. (2000). costs, and 1.4 for those with little or no integration costs.

OCR for page 24
34 ASSESSMENT OF FUEL ECONOMY TECHNOLOGIES FOR LIGHT-DUTY VEHICLES bolt them together to produce a vehicle that meets custom- The EPA study (Rogozhin et al., 2009) estimated RPEs ers’ expectations and satisfies all regulatory requirements.7 for the largest manufacturers for the year 2007 using publicly Integrating a new engine or transmission to decrease fuel available data in manufacturers’ annual reports. Several as- consumption will have much greater ramifications for vehicle sumptions were required to infer components not reported, design and is likely to generate greater integration costs than or reported in different ways by different manufacturers. simpler components. The method is similar to that used by Bussman (2008) and In a presentation to the committee, the EPA raised con- produced similar results. One notable difference is that the cerns that markup factors on piece or supplier costs tended estimates shown in Table 3.6 attempt to exclude legacy health to overestimate the costs of most fuel economy technologies: care costs, estimated at 45 percent of total health care costs, “Our first preference is to make an explicit estimate of all which in turn were estimated to be 3 percent of fully bur- indirect costs rather than rely on general markup factors” dened manufacturing costs. This would lower the estimated (EPA, 2008b, slide 4). Nonetheless, in its assessment of the RPEs by 1 to 2 percent relative to estimates in other reports, costs of greenhouse gas mitigation technologies for light- all else being equal. The estimated RPE multipliers were duty vehicles, the EPA staff assumed a uniform markup of 50 remarkably consistent across manufacturers (Table 3.6) and percent over supplier costs (i.e., a markup factor of 1.5). Still, very comparable to the studies cited above. Estimated RPE the EPA maintains that such a markup is too large: “We be- multipliers ranged from 1.42 for Hyundai to 1.49 for Nissan, lieve that this indirect cost markup overstates the incremental with an industry average of 1.46. Adding 1 to 2 percent for indirect costs because it is based on studies that include cost health care costs would bring the average multiplier even elements—such as funding of pensions—which we believe closer to 1.5. are unlikely to change as a result of the introduction of new technology” (EPA, 2008a, p. 47). Estimating Technology-Specific Markup Factors and IC Following up on this assertion, the EPA commissioned Multipliers a study of RPE factors and indirect cost (IC) multipliers (Rogozhin et al., 2009). The IC multiplier attempts to im- The assertion that different technologies will induce dif- prove on the RPE by including only those specific elements ferent changes in indirect costs seems evident. The question of indirect costs that are likely to be affected by vehicle is how to identify and measure the differences. At the present modifications associated with environmental regulation. time a rigorous and robust method for estimating these dif- In particular, fixed depreciation costs, health care costs for ferential impacts does not exist (Bussmann and Whinihan, retired workers, and pensions may not be affected by many 2009). Therefore, it is not clear that the accuracy of fuel vehicle modifications caused by environmental regulations. consumption cost assessment would be increased by the use The EPA study (Rogozhin et al., 2009) also criticizes of technology-specific, as opposed to an industry-average, the RPE method on the grounds that an increase in the total markup factor. The EPA (Rogozhin et al., 2009), however, cost of producing a vehicle will not be fully reflected in the has taken the first steps in attempting to analyze this problem increased price of the vehicle due to elasticities of supply and in a way that could lead to a practical method of estimating demand. For this reason, the report argues that manufacturer technology-specific markup factors. profits should not be included in the RPE multiplier. The The EPA-sponsored study (Rogozhin et al., 2009) went committee disagrees with this assertion for two reasons. First, on to estimate IC multipliers as a function of the complexity as noted earlier, the global automotive industry approximates or scope of the innovation in an automaker’s products caused what economists term a monopolistically competitive market, by the adoption of the technology. A four-class typology of that is, a market in which there is product differentiation but a innovation was used: high degree of competition among many firms. In a monopo- listically competitive market, in the long run the full costs of • Incremental innovation describes technologies that production will be passed on to consumers. In the long run, require only minor changes to an existing product monopolistically competitive market supply is perfectly elas- and permit the continued use of an established design. tic at the long-run average cost of production (this includes Low-rolling-resistance tires were given as an example a normal rate of return on capital). Since cost estimates by of incremental innovation. convention assume long-run conditions (full scale economies • Modular innovation is that which does not change the and learning), long-run supply assumptions should be used to architecture of how components of a vehicle interact ensure consistency. The increase in RPE is a reasonable esti- but does change the core concept of the component re- mate of the change in welfare associated with the increased placed. No example was given for modular innovation. vehicle cost especially, as noted above, in the long run. • Architectural innovation was defined as innovation that requires changes in the way that vehicle components are linked together but does not change the core design 7 For some parts, the effort required for integration may be small. Tires are concepts. The dual-clutch transmission was offered often cited as an example. Still, even tires have implications for a vehicle’s as an example, in that it replaces the function of an suspension and braking systems.

OCR for page 24
35 COST ESTIMATION TABLE 3.6 Individual Manufacturer and Industry Average Retail Price Equivalent (RPE) Multipliers: 2007 Relative to Cost of Sales Industry Daimler RPE Multiplier Contributor Average Chrysler Ford GM Honda Hyundai Nissan Toyota VW Vehicle Manufacturing Cost of sales 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 Production Overhead Warranty 0.03 0.04 0.03 0.03 0.01 0.02 0.03 0.04 0.02 R&D product development 0.05 0.04 0.02 0.06 0.07 0.04 0.06 0.05 0.06 Depreciation and amortization 0.07 0.11 0.05 0.06 0.05 0.06 0.09 0.08 0.09 Maintenance, repair, operations cost 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 Total production overhead 0.18 0.22 0.13 0.17 0.16 0.15 0.21 0.19 0.20 Corporate Overhead General and administrative 0.07 0.05 0.12 0.07 0.11 0.08 0.03 0.06 0.03 Retirement <0.01 0.01 0.00 0.01 <0.01 <0.01 <0.01 <0.01 <0.01 Health 0.01 <0.01 <0.01 0.01 0.01 0.01 0.01 0.01 0.01 Total corporate overhead 0.08 0.06 0.13 0.08 0.14 0.09 0.04 0.07 0.04 Selling Transportation 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.10 Marketing 0.04 0.02 0.04 0.05 0.03 0.05 0.08 0.03 0.02 Dealers Dealer new vehicle net profit <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 Dealer new vehicle selling cost 0.06 0.06 0.06 0.06 0.06 0.06 0.06 0.06 0.06 Total selling and dealer contributors 0.14 0.12 0.14 0.14 0.13 0.15 0.18 0.12 0.17 Sum of Indirect Costs 0.40 0.40 0.39 0.40 0.44 0.39 0.43 0.38 0.41 Net income 0.06 0.07 0.05 0.05 0.04 0.03 0.06 0.09 0.02 Other costs (not included as contributors) 0.04 0.04 0.11 0.06 0.02 0.01 0.01 <0.01 0.05 RPE multiplier 1.46 1.47 1.45 1.45 1.47 1.42 1.49 1.48 1.43 SOURCE: Rogozhin et al. (2009), Table 3-3. existing transmission but does require redesign and TABLE 3.7 Weighted Industry Average RPE Components reintegration with other components. Omitting Return on Capital • Differential innovation involves significant changes Cost Contributor Light Car Industry Average in the core concepts of vehicle components, as well as Production Overhead their integration. Hybrid vehicle technology was cited Warranty 0.03 as an example because it changes the functions of such R&D (product development) 0.05 key components as the engine, brakes, and battery. Depreciation and amortization 0.07 Maintenance, repair, operations cost 0.03 An industry average was computed for each component Total production overhead 0.18 Corporate Overhead of the RPE, omitting profit, or net income. As stated above, General and administrative 0.07 the committee considers this omission to be in error. The Retirement 0.00 resulting components are shown in Table 3.7. Next, based Health care 0.01 on the judgment of an expert panel, short- and long-term Total corporate overhead 0.08 effects on the RPE components were estimated for the four Selling categories of technology innovation (Rogozhin et al., 2009). Transportation 0.04 A value of zero for the effect of a technology innovation Marketing 0.04 Dealers on an RPE component implies that the application of that Dealer new vehicle selling cost 0.06 technology has no impact on the cost of that particular RPE Total selling and dealer costs 0.14 component. There will be no increase in expenditure on that Sum of Indirect Costs 0.40 RPE component as a result of the adoption of the technol- SOURCE: Rogozhin et al. (2009), Table 4-1. ogy. A value of 1 implies that the cost of the component will increase directly with the increased cost of the component. Values greater than 1 imply a greater-than-proportional in- crease. Each RPE component is multiplied by its respective short- or long-term effect, and the results are summed and

OCR for page 24
36 ASSESSMENT OF FUEL ECONOMY TECHNOLOGIES FOR LIGHT-DUTY VEHICLES Finding 3.2: RPE factors certainly do vary depending on TABLE 3.8 Short- and Long-Term Indirect Cost Multipliers the complexity of the task of integrating a component into a vehicle system, the extent of the required changes to other Low Medium High Industry components, the novelty of the technology, and other factors. Complexity Complexity Complexity Average RPE However, until empirical data derived by means of rigorous Short term 1.05 1.20 1.45 1.46 estimation methods are available, the committee prefers to Long term 1.02 1.05 1.26 1.46 use average markup factors. SOURCE: Rogozhin et al. (2009), Table 4-5. Finding 3.3: Available cost estimates are based on a variety of sources: component cost estimates obtained from sup- added to 1.0 to produce the IC multipliers. The multipliers pliers, discussions with experts at OEMs and suppliers, range from 1.05 to 1.45 in the short run and 1.02 to 1.26 in comparisons of actual transaction prices when publicly the long run (Table 3.8). This implies that none of the fuel available, and comparisons of the prices of similar vehicles economy technologies considered, no matter how complex, with and without a particular technology. There is a need could cause an increase in indirect costs as large as the in- for cost estimates based on a teardown of all the elements dustry average indirect costs, especially in the long run. This of a technology and a detailed costing of material costs, ac- result would imply that the more that regulatory requirements counting for labor time and capital costs for all fabrication increase the cost of automobile manufacturing, the lower the and assembly processes. Such studies are more costly than overall industry RPE would be. the current approaches listed above and are not feasible for advanced technologies whose designs are not yet finalized and/or whose system integration impacts are not yet fully FINDINGS understood. Nonetheless, estimates based on the more rig- Large differences in technology cost estimates can result orous method of teardown analysis are needed to increase from differing assumptions. Carefully specifying premises confidence in the accuracy of the costs of reducing fuel and assumptions can greatly reduce these differences. These consumption. include the following: Technology cost estimates are provided in the follow- • Whether the total cost of a technology or its incre- ing chapters for each fuel economy technology discussed. mental cost over the technology that it will replace is Except as indicated, the cost estimates represent the price estimated; that an OEM would pay a supplier for a finished component. • Whether long-run costs at large-scale production are Thus, on average, the RPE multiplier of 1.5 would apply. assumed or short-run, low-volume costs are estimated; • Whether learning by doing is included or not; REFERENCES • Whether the cost estimate represents only direct in- Albu, S. 2008. ARB perspective on vehicle technology costs for reducing house manufacturing costs or the cost of the purchase greenhouse gas emissions. Presentation by Assistant Chief, Mobile of a component from a Tier 1 supplier; Source Division, California Air Resources Board, to the Committee on • Whether the RPE multiplier is based on industry aver- Technologies for Improving Light-Duty Vehicle Fuel Economy, Janu- age markups or is specific to the nature of the technol- ary 24, Detroit, Mich. ogy; and Bussmann, W.V. 2008. Study of industry-average markup factors used to estimate retail price equivalents (RPE). Presentation to the Committee • What other changes in vehicle design, required to on Assessment of Technologies for Improving Light-Duty Vehicle Fuel maintain vehicle quality (e.g., emissions, towing, Economy, January 24, Detroit, Mich. gradability, launch acceleration, noise, vibration, Bussmann, W.V., and M.J. Whinihan. 2009. The estimation of impacts on harshness, manufacturability), have been included in retail prices of regulations: A critique of “Automobile Industry Retail the cost estimate. Price Equivalent and Indirect Cost Multipliers.” Prepared for Alliance of Automobile Manufacturers, May 6, Southfield, Mich. DOT/NHTSA (Department of Transportation/National Highway Traffic Finding 3.1: For fully manufactured components purchased Safety Administration). 2009. Average fuel economy standards, pas - from a Tier 1 supplier, a reasonable average RPE markup senger cars and light trucks, model year 2001: Final rule. 49 CFR Parts factor is 1.5. For in-house direct (variable) manufacturing 523, 531, 533, 534, 536, and 537, Docket No. NHTSA-2009-0062, RIN costs, including only labor, materials, energy, and equipment 2127-AK29. March 23. Washington D.C. Duleep, K.G. 2008. Analysis of technology cost and retail price. Presenta - amortization, a reasonable average RPE markup factor is tion to Committee on Assessment of Technologies for Improving Light- 2.0. In applying such markup factors, it is essential that the Duty Vehicle Fuel Economy, January 25, Detroit, Mich. cost basis be appropriately defined and that the incremental EEA (Energy and Environmental Analysis, Inc.). 2006. Technologies to cost of fuel economy technology is the basis for the markup. Reduce Greenhouse Gas Emissions from Light-Duty Vehicles. Report The factors given above are averages; markups for specific to Transport Canada, Ottawa, Ontario, March. Arlington, Va. technologies in specific circumstances will vary.

OCR for page 24
37 COST ESTIMATION EPA (U.S. Environmental Protection Agency). 2008a. EPA Staff Techni- Martec Group, Inc. 2008b. Variable Costs of Fuel Economy Technologies. cal Report: Cost and Effectiveness Estimates of Technologies Used Report prepared for the Alliance of Automobile Manufacturers. June 1. to Reduce Light-Duty Vehicle Carbon Dioxide Emissions. EPA420- Washington, D.C. R-08-008. Washington, D.C. NESCCAF (Northeast States Center for a Clean Air Future). 2004. Reducing EPA. 2008b. The use of markup factors to estimate indirect costs. Presen - Greenhouse Gas Emissions from Light-Duty Motor Vehicles. Available tation from National Vehicle and Fuel Emissions Laboratory, Office at http://www.nesccaf.org/documents/rpt040923ghglightduty.pdf. Ac- of Transportation and Air �uality, to the Committee on Assessment cessed July 6, 2009. of Technologies for Improving Light-Duty Vehicle Fuel Economy, NRC (National Research Council). 2002. Effectiveness and Impact of Cor- National Research Council, March 31, Washington, D.C. porate Average Fuel Economy (CAFE) Standards. National Academy FEV, Inc. 2009. Light-duty technology cost analysis overview. Presentation Press, Washington, D.C. to the Committee on the Assessment of Technologies for Improving OTA (Office of Technology Assessment). 1995. Advanced Automotive Light-Duty Vehicle Fuel Economy, March 16, Washington, D.C. Technology: Visions of a Super-Efficient Family Car. OTA-ETI-638. IBIS Associates, Inc. 2008. Data Collection and Analysis: Vehicle Systems Washington, D.C. Costs. Report to the Committee on Assessment of Technologies for Im- Rogozhin, A., M. Gallaher, and W. McManus. 2009. Automobile Industry proving Light-Duty Vehicle Fuel Economy, National Research Council, Retail Price Equivalent and Indirect Cost Multipliers. Report by RTI December, 2008, Waltham, Mass. International to Office of Transportation and Air �uality. U.S. Envi- Kolwich, G. 2009. Light-Duty Technology Cost Analysis Pilot Study. ronmental Protection Agency. RTI Project Number 0211577.002.004. Report FEV 07-069-103F. Prepared for the U.S. Environmental Protec- February. Research Triangle Park, North Carolina. tion Agency, Ann Arbor, Michigan, September 3. FEV, Inc., Auburn Vyas, A., D. Santini, and R. Cuenca. 2000. Comparison of Indirect Cost Hills, Mich. Multipliers for Vehicle Manufacturing. Center for Transportation Ludtke and Associates. 2004. Perform Cost and Weight Analysis, Head Research, Argonne National Laboratory. April. Argonne, Ill. Protection Air Bag Systems, FMVSS 201. DOT HS 809 842. December. Wene, C. 2000. Experience Curves for Energy Technology Policy. Interna- U.S. Department of Transportation, National Highway Traffic Safety tional Energy Agency, Paris. Administration, Washington, D.C. Martec Group, Inc. 2008a. Variable costs of fuel economy technologies. Presentation to the Committee on Assessment of Technologies for Im - proving Light-Duty Vehicle Fuel Economy, National Research Council, January 24, Detroit, Mich.