The Benefits and Costs of New Fuels and Engines for Light-Duty Vehicles in the United States

1. INTRODUCTION

Rising world oil prices, coupled with concerns about global climate change, are forcing a reconsideration of the fuels and engines that are used in the transport sector of the U.S. economy. How to propel cars and light trucks is of special interest because these “light-duty” vehicles account for approximately 60% of oil use in the U.S. transport sector and total oil demand for these vehicles is projected to rise by almost 20% between now and 2020.⁽¹⁾ A similar dilemma is facing leaders in other regions of the world, including the European Union, China, India, and Japan.

Using the gasoline-powered internal combustion engine as a baseline for comparison, we estimate the benefits and costs of alternative engines and fuels for new light-duty passenger vehicles in the United States. Based on explicit screening criteria, the scope of the alternatives is restricted to gasoline-electric hybrid technology, advanced diesel technology, and dedicated vehicles that run continuously on 85% ethanol and 15% gasoline (E85). Net benefit estimates are computed on a per vehicle basis. Separate estimates are provided for applications to passenger cars, sport utility vehicles, and large pickup trucks.

For analytic purposes, we assume in this article that market and policy trends cause each of the three alternative technologies to be used in at least 1 million U.S. vehicles. We focus not on the benefits and costs of the policies that might produce this outcome but on the marginal benefits and costs of a dedicated E85 vehicle, hybrid, or advanced diesel, after the transitional process has expired and an equilibrium is reached. This is a steady-state assumption that focuses on the long-run marginal benefits and costs of each technology under a condition of substantial economies of scale. The same steady-state assumption is made for all three technologies, although the economic cost and political will necessary to bring each technology to the steady-state condition may not be equal. We also omit any inefficiencies associated with policies (tax preferences or regulations) that may be necessary to ensure the penetration of the three technologies to 1 million units per year. For example, analysts argue that tighter federal fuel economy standards cause more inefficiency than higher gasoline taxes.^(2,3)

We examine each of the three alternatives using two perspectives: a “private” perspective that considers only those benefits and costs likely to be incurred by the vehicle owner, and a “societal” perspective that considers various externalities (e.g., climate change and energy security impacts) as well as private benefits and costs to vehicle owners. The societal perspective is favored in welfare economics and was applied in the pioneering studies of this issue.^(4,5)

The key outcome measure is the net present value (NPV), defined as the present value of benefits minus the present value of costs (2005$). We examine the robustness of the NPV estimates by exploring how these estimates change with plausible yet different assumptions about key input values. In order to provide further context for the NPV estimates, we also summarize recent market and policy trends in the United States that appear to be affecting the penetration of the three alternatives, including a variety of qualitative factors that we were not able to incorporate into the numeric benefit-cost analysis.

This article contributes to a small but growing body of benefit-cost literature on new propulsion systems for cars and light trucks, including previous work by Lave et al.,⁽⁵⁾ Lave and MacLean,⁽⁶⁾ the National Research Council,⁽⁷⁾ MacLean and Lave,⁽⁸⁾ and Lipman and Delucci.⁽⁹⁾ The National Research Council⁽⁷⁾ primarily examines improvements to the gasoline engine, with little analysis of hybrid and diesel technology and no analysis of E85. Studies that examine hybrid engines have reached conflicting conclusions, such as Lave and MacLean⁽⁶⁾ versus Lipman and Delucci.⁽⁹⁾ The most comprehensive assessment of technologies is provided by Lave et al.⁽⁵⁾ and refined by MacLean and Lave,⁽⁸⁾ but their primary assumptions about fuel prices ($1.50 per gallon), diesel emission control, and hybrid technology costs predate recent developments in the industry.

The strengths of the present study include application of a consistent analytic framework to four different technologies, a lifecycle perspective, updated inputs based on recent economic and technological developments, and extensive sensitivity analysis of results based on plausible changes to input values. The study also summarizes recent trends in the marketplace and public policy.

The U.S. market for new light-duty vehicles—now about 17 million units per year—is currently dominated by the gasoline-powered internal combustion engine. A wide range of fuel/engine combinations have been suggested as alternatives: advanced diesel technology, electric vehicles, fuel cells, liquid fuels derived from coal, ethanol, biodiesel, compressed natural gas, propane, gasoline-electric hybrid technology, diesel-electric hybrid technology, and plug-in hybrid technology.

In order to restrict the analysis to a manageable number of alternatives, we applied three screening criteria to each fuel/engine alternative: (1) Will the alternative help reduce U.S. oil consumption? (2) Will the alternative help reduce the emissions of greenhouse gases implicated in global climate change? (3) Could the alternative have significant market penetration (at least 1 million units of production per year) in the U.S. transport sector by the early part of the next decade (2010–2020)? We selected for analysis only those alternatives where a “yes” answer seemed appropriate for all three questions, based on our literature reviews and discussions with relevant specialists.

The three chosen alternatives are (1) full gasoline-electric hybrid technology, such as the system offered as standard equipment on the Toyota Prius or the system offered as an option by Ford Motor Company on the Escape SUV; (2) advanced diesel technology, such as the systems sold widely in Europe coupled with low-sulfur diesel fuel and recent advances in control technology that minimize tailpipe emissions of particulates and nitrogen dioxide (e.g., particulate traps and nitrogen dioxide catalysts coupled with low-sulfur diesel fuel); and (3) vehicles that can operate on a mixture of gasoline and up to 85% ethanol, such as the 5 million vehicles already on the road in the United States that were produced by General Motors Corporation, Ford Motor Company, and Chrysler Corporation. We acknowledge that some of the rejected alternatives also hold considerable promise (e.g., plug-in hybrids) and are worthy of future benefit-cost studies, even though their near-term market penetration may be slight.

Even for the three selected alternatives to the gasoline engine, meeting the market penetration criterion will not be easy or costless—even though 1 million vehicles per year is a small fraction of the 17 million U.S. vehicles sold each year. Producing and fueling 1 million diesel-powered light-duty vehicles per year would require transitional costs at engine suppliers (domestic and foreign), vehicle assembly lines, refineries, and refueling stations. For hybrid technology, a major expansion of the supplier network for batteries and other inputs would be required. Although producing 1 million E85 vehicles per year is quite plausible, fueling those vehicles would require major changes in infrastructure (e.g., pumps at refueling stations) and efforts (information and incentives) to persuade consumers to use the fuel. Our focus in this article is not on the transition process but on the marginal benefits and costs of the technologies once the transition is accomplished.

2. NET BENEFIT ESTIMATES: PRIVATE PERSPECTIVE

We begin with a simple analysis in which a hypothetical consumer faces a choice between a gasoline-powered internal combustion engine, a gasoline-electric hybrid vehicle, an advanced diesel-powered vehicle, or a vehicle that runs continuously on E85. The consumer's choice is analyzed separately assuming that a mid-sized passenger car, a mid-sized SUV, and a large pickup truck (or a large SUV) are under consideration.

For the nominal assumptions about technology costs, fuel efficiency, and performance in Table I, we rely primarily on estimates prepared by analysts in the federal government: Greene et al.,⁽¹⁰⁾ NHTSA,⁽¹¹⁾ and EPA⁽¹²⁾ for hybrids and diesels, and IMF,⁽¹³⁾ EPA,⁽¹⁴⁾ and EIA⁽¹⁾ for corn-based ethanol and E85. Our nominal assumptions about driving behavior (annual vehicle miles traveled (VMT), annual vehicle survival rates as a result of accidents/malfunctions, extra driving incurred as a result of increased fuel economy (the rebound effect), and percentage of urban travel are adapted from NHTSA,⁽¹⁵⁾ EPA,⁽¹⁴⁾ and Greening et al.⁽¹⁶⁾ In the sensitivity analyses, we considered a variety of estimates from different viewpoints.

The gasoline-electric hybrid engine is the most costly of the four systems but is also the most fuel-efficient and offers some performance benefit compared to the gasoline-powered engine. The diesel engine is more expensive than the gasoline engine but less expensive than the hybrid. The diesel is assumed to offer better performance (measured by torque) than the two other alternatives but is not quite as fuel-efficient as the hybrid. Note that the performance comparison excludes acceleration, which, though quantifiable, varies largely by make and model and less as a result of the particular engine technology or fuel. The E85 vehicle, which is also likely to have flex-fuel capability, has a small incremental vehicle cost compared to the gasoline engine. Yet the E85 vehicle generates the largest fuel expenses because ethanol, which is assumed to be corn-based, is more expensive to produce than gasoline after adjusting for differences in energy content.

The nominal average price of gasoline over the vehicle's life is assumed to be $2.50 per gallon (federal and state taxes included), based on oil price assumptions in 2007 IMF World Economic Outlook.⁽¹³⁾ Oil price projections from EIA's 2007 Annual Energy Outlook⁽¹⁾ and recent trends in oil prices drive the low and high ends of our fuel cost range. The nominal cost of E85 is assumed to be $2.42 per gallon based on IMF's cost estimates for corn-based ethanol,⁽¹³⁾ while the ends of the range are driven by EPA's recent Regulatory Impact Analysis on the Renewable Fuels Standard.⁽¹⁴⁾ We follow EIA's assumption in their analysis that the tax credit (a 51 cent per gallon tax credit for refiners that blend ethanol into their fuel) will persist through the period considered in this study. Our fuel price assumptions are summarized in Table II.

We assume that the consumer cares only about the cost of the engine technology, any performance gain or detriment, any mobility impacts, and the net fuel savings over the projected life of the vehicle (see Table I). Quantifying the benefit of increased mobility to the consumer proved difficult, and we have assumed that the additional miles traveled as a result of the “rebound effect” are worth at least the cost of traveling them. By including the nonfuel variable costs, such as increased vehicle maintenance, exposure to crash risks, and transit time, we arrived at an estimate of 1.3 times the fuel cost of the additional travel. Even this estimate is likely to understate the true benefit. Our assumptions about the monetary value of torque increases in diesel and hybrid vehicles are based on the econometric estimates of Crandall et al.⁽¹⁷⁾ for light trucks. We also assume that the consumer applies a 7% real discount rate to any future benefits and costs, and compares the alternatives according to their net present value of monetized savings, where savings can be positive (good) or negative (bad). The 7% rate is near the effective interest rate that consumers pay on loans for vehicle purchases, 6% for new cars and 9% for used cars.⁽¹⁸⁾ NHTSA⁽¹¹⁾ argues that the proper weighted average of the two loan rates is quite close to 7%.

The results from the private perspective are displayed in Table III. For all three vehicle types, the diesel has the largest positive savings ($460 for the car, $1,249 for the SUV, and $2,289 for the pickup truck). The hybrid has small positive savings for the passenger car ($198), larger savings for the SUV ($1,066), and somewhat smaller savings for the large pickup ($505). The E85 vehicle has consistently negative NPV (–$1,034 for cars, –$1,332 for SUVs, and −$1,632 for large pickups), primarily due to the higher fuel expenditures.

The impacts of the four components (fuel consumption, technology cost, performance, and mobility) on the private results are also reported in Table III. A key insight is that while the hybrid saves more fuel than the diesel, the overall advantage of the diesel arises from the assumptions about lower technology costs and improved performance. The negative “mobility” benefit for the E85 vehicle results from the increased cost of per mile travel (due to cost differences and the lower energy content of the fuel). So the E85 vehicle is estimated to travel fewer miles, thus saving fuel but reducing the mobility of the consumer.

In some regions of the country where E85 is more widely available (e.g., Minnesota), owning a flex-fuel vehicle might be especially advantageous in the event of long gas lines. Yet energy economists argue that long gas lines like those experienced in the United States in 1973–1974 are unlikely in the future due to the growing fungibility of the world oil market and the removal of price controls that contributed to long waiting times.⁽¹⁹⁾ We have not quantified any incremental benefit to the consumer of owning a vehicle that can operate on two fuels instead of one.

For purposes of this analysis, we assume that at least 1 million vehicles are being fueled continuously by E85, even if the relative costs of gasoline are lower than the energy-equivalent costs of E85. This usage obviously would not occur voluntarily if the relative price of E85 is higher, which is why our analysis should not be interpreted as projecting levels of any alternative fuel use. Expanded E85 usage would require other steps, such as larger subsidies or mandates for renewable fuels use. Countries such as Brazil and Sweden have already taken such steps to promote substantial use of E85, and policymakers in the corn-producing states of the United States and in the U.S. Congress are considering pro-E85 policies.⁽²⁰⁾

Our technology cost estimates do not necessarily include all of the transitional and one-time costs associated with each of the alternatives. If some refineries need to shift their mix of products toward diesel fuel instead of gasoline, there will be some one-time capital expenses at the refinery associated with the shift. If auto-assembly lines shift from making gasoline engines to hybrids or diesels, there will be some one-time costs associated with the conversion.⁽²¹⁾ And if E85 is used to power 1 million vehicles, there will be some one-time costs associated with upgrading distribution, storage facilities, and pumps at retail outlets—though some of these one-time costs (e.g., capital costs of ethanol plants) are already incorporated into the marginal cost estimates we are using.

In assessing the overall results from the private perspective, it is important to recognize that the net savings for diesels and hybrids are small compared to the overall purchase cost of a new light-duty vehicle, which can range from $20,000 to $50,000 per vehicle and even higher. Our inference is that, assuming fuel prices stabilize at $2.50 per gallon, market forces alone are unlikely to induce a powerful shift away from the gasoline-powered internal combustion engine, especially since steady fuel economy improvements in the gasoline engine will compete with advanced diesels and hybrids. Among those consumers with a very strong interest in fuel efficiency or performance, hybrid and diesel offerings will find some enthusiastic buyers. In contrast, E85 appears to have no appeal from a consumer perspective (given our nominal assumptions) because corn-based ethanol is more costly to produce than gasoline after adjusting for the energy content of the fuels.

3. NET BENEFIT ESTIMATES: SOCIETAL PERSPECTIVE

Our societal analysis introduces the following complications: (1) the impact of lowered U.S. oil consumption on U.S. energy security is quantified and monetized, (2) the lifecycle impacts of emissions of conventional tailpipe pollutants and greenhouse gas emissions are quantified and monetized, and (3) the indirect effect of each alternative on vehicle miles of travel through the “rebound effect,” including pollution, congestion, and other externalities, is quantified and monetized. Fuel prices are treated net of taxes and credits in the societal case because taxes and credits are considered to be transfers (i.e., they produce no net change in well-being for society as a whole). This accounts for the large difference in NPV fuel savings between the private and societal analyses. Since the societal analysis includes consideration of impacts on everyone, not just the vehicle owner or user, it is the most relevant perspective for public policy deliberations.

The key inputs used for the societal perspective are displayed in Table IV.

The assumptions about driving behavior are consistent with those made in the private analysis above. The climate analysis accounts for all greenhouse gases (e.g., carbon dioxide, methane, and nitrous oxide), weighted by their relative global warming potential and expressed in units of “CO₂ equivalents” that occur over the lifecycle of each automobile technology-fuel combination. Actual emission quantities were estimated using the GREET model developed by Argonne National Labs (see http://www.transportation.anl.gov/software/GREET/). Based on recent work by Nordhaus,⁽²²⁾ the shadow price of carbon dioxide shows the damages from climate change increasing over time. Nordhaus estimates the future social damages from climate change using a computable general equilibrium model known as DICE and finds that climate change damages increase over time. Accordingly, we assume an increasing social cost for greenhouse gas emissions.

The energy security externality represents an estimated cost of both U.S. ability to impact world oil price through decreased consumption and the U.S. economy's vulnerability to oil price volatility.⁽²³⁾ Though the magnitude of these effects is still debated in the literature, our nominal assumption is a revised estimate (adjusted for the increase in world oil price) based on a study by the National Research Council.⁽⁷⁾

Congestion enters the analysis because hybrids and diesels increase travel whereas E85 discourages travel. Standard damage estimates for congestion are drawn from regulatory analyses of federal fuel-economy standards.⁽¹¹⁾ Note that the congestion cost applies only to the rebound miles of travel, whereas the damages from pollution apply to all miles of travel.

The rebound effect has multiple ramifications in the societal analysis. For example, the rebound effect dampens the extent of fuel savings, pollution reductions, and greenhouse gas reductions attributable to the diesel engine and the hybrid engine. Interestingly, the external benefits of E85 are enhanced by the rebound effect because the higher energy-equivalent cost of E85 discourages travel, reduces greenhouse gases, and reduces congestion. But the mobility disbenefit hurts E85. Conventional air pollution values are estimated on a lifecycle basis for each technology-fuel combination using the GREET model. Multiple pollutants (see Table IV) are then aggregated using shadow prices for the different pollutants based on estimates from NHTSA.⁽¹¹⁾ The pollution impacts of the rebound effect (more travel causes more pollution) have a significant effect, as do the inherent differences in the cleanliness of the different fuels and engines.

We selected our nominal input values for the societal analysis based on several factors. The input values were used in recent, major analyses of federal government regulatory impacts, such as in Reference 11. Recent research from a recognized expert in the subject area provided an up-to-date estimate. We used this basis in choosing the nominal shadow price for climate change. We also selected values from older analyses and updated them for current conditions, which occurred for the energy security externality. We recognize that uncertainty exists for nearly all of the input values in the society analysis, and we establish broad, but plausible, ranges for each of the key values. The section below on sensitivity analyses shows how changes in the input assumptions affect the results.

Despite the numerous complications that have been added to the societal perspective, the results are qualitatively similar to the results found for the private perspective. Measured by NPV, the diesel is the most promising alternative. The NPV for the hybrid is slightly negative for cars but positive for SUVs and pickups. The NPV of E85 is again uniformly negative. A careful look at Table V reveals an underlying pattern: the most important inputs appear to be fuel savings, technology cost, performance, and mobility, the same inputs that drive the private perspective. In other words, the externalities are relatively small compared to the private impacts.⁽²⁴⁾

4. SENSITIVITY ANALYSES

We examined the robustness of our analytic results with respect to plausible changes in the assumed values of inputs. We begin with a sensitivity analysis of the results from the private perspective and then explore the societal perspective. Numerous sensitivity analyses were performed, though not all are reported here. We focus on those that affect the rank ordering of the technologies or are likely to be of significant interest to readers.

4.1. Robustness of Private Results

In the private case, the most interesting sensitivity analyses concern the projected fuel price and the incremental costs of competing technologies. As we look over the maximum 25-year lifetime of a light-duty vehicle, the anticipated price of fuel looms as a large unknown. The low end of our fuel price range may seem very low by the experience of recent years in the United States, but it remains higher than what U.S. motorists saw at the pump through most of the 1980s and 1990s. The high-price scenario may seem low given recent experience, but is much larger than EIA's upper-end, long-run forecast⁽¹⁾ and larger than what U.S. motorists experienced after Hurricane Katrina and the unexpected shutdown of Gulf Coast refineries. The low- and high-price scenarios are intended to stretch the bounds of plausibility, thereby serving a useful bounding role. Thus, our sensitivity analysis covers a large range of potential fuel prices.

The results of the fuel-price analysis are reported in Table VI. At the low gasoline price hybrids have negative NPV for all three vehicle types, but the NPV of the diesel remains positive for both SUVs and pickups. At the high gasoline price ($3.50 per gallon), both the hybrid and diesel have positive NPV for all three vehicle types, but the NPV of the hybrid is higher than the NPV of diesel in car and SUV applications. All of these analyses assume that gasoline and diesel fuel prices move together, with diesel priced slightly higher than gasoline. Since the world oil price is the primary driver of both gasoline and diesel fuel prices, we do not expect the two fuel prices to diverge in the long run.

Table VI. NPV of Hybrid, Diesel, and E85: Alternative Oil and Ethanol Prices (Private Perspective)

Oil Price Case	Hybrid NPV	Diesel NPV	E85 NPV E85 Price Case
			Low	Nominal	High
Low	−$692	−$172	$39	−$2,929	−$4,720
Nominal	$198	$460	$2,155	−$1,034	−$3,008
High	$1,449	$1,390	$5,439	$1,870	−$290
Sport Utility Vehicle
Low	−$48	$462	$15	−$3,697	−$5,930
Nominal	$1,066	$1,249	$2,667	−$1,332	−$3,800
High	$2,633	$2,417	$6,791	$2,304	−$405
Large Pickup Truck
Low	−$633	$1,293	$69	−$4,638	−$7,471
Nominal	$505	$2,289	$2,237	−$1,632	−$4,755
High	$2,116	$3,761	$5,732	$3,002	−$432

The NPV of E85 is quite sensitive to the estimated cost of producing E85 from corn. At the low cost estimate ($1.70 per gallon), E85 has a positive NPV for all three vehicle types, higher than the NPV of the diesel or hybrid. At the high-cost estimate ($3.04 per gallon), unfavorable results for E85 become more strongly unfavorable.

The low-cost ethanol scenario is based on EPA,⁽¹⁴⁾ but some analysts question whether the cost of producing corn-based ethanol can really decline by 50% over the next decade. Progress is being made at the farm, where more corn is being grown with the same inputs, and at the ethanol plant, where more ethanol is being produced from a bushel of corn.^(25,26) Creative ways are also being explored to generate commercial value from the byproducts of corn-based ethanol production.⁽²⁷⁾ Despite these encouraging trends, continued growth in the price of corn could make the low-cost ethanol scenario completely implausible. A different interpretation of the low-cost scenario would be a breakthrough in cellulosic ethanol production or a removal of the tariff on imported ethanol from low-cost producers such as Brazil. A detailed analysis of these noncorn possibilities is beyond the scope of this article.

The best fuel-price scenario for E85 is a combination of high world oil prices and low costs of producing ethanol. Under those conditions, the NPV of E85 is not only positive, it is larger than the NPV of both diesel and hybrid engines. But the probability of these two unlikely conditions occurring simultaneously for a decade or more is exceedingly small.

We have assumed that the price of diesel fuel will track gasoline as it has done in the past (and is forecasted to continue by EIA). However, as recent experience at the pump suggests, this need not always be the case. When the prices of diesel and gasoline are very similar, the fuel savings benefits of the diesel engine can be significant. But if the two fuel prices diverge we get a different story. Table VII illustrates the consequences of this price divergence by showing the price at which the NPV of fuel savings for the diesel technology are zero. Since the fuel savings are based on the relative cost of gasoline, we performed this calculation for each of the three price scenarios under consideration. In the nominal price case, diesel fuel would have to cost about 55 cents more than gasoline per gallon (history suggests such a discrepancy is unlikely over the long term). It is worth noting that even when the NPV of fuel savings is zero, there is still a mobility benefit for the diesel since the consumer is still able to travel more miles for the same cost. This price discrepancy can impact the results in another way as well: as the discrepancy between the two fuel prices increases, the rank ordering of technologies can change as the hybrid vehicles gain ground on the diesels (Table VII).

Table VII. Effect of Diesel Price Increases Over Gasoline (Private Perspective)

Oil Price Case	Price Difference Where NPV Fuel Savings = 0 (cents/gal)	Diesel NPV at Those Prices	Hybrid NPV
Low	39	−$1,131	−$692
Nominal	55	−$964	$198
High	77	−$682	$1,449
Sport Utility Vehicle
Low	39	−$727	−$48
Nominal	55	−$514	$1,066
High	76	−$154	$2,633
Large Pickup Truck
Low	39	−$226	−$633
Nominal	54	$35	$505
High	76	$477	$2,116

The costs of advanced diesel engines and the hybrid engine are also not known with certainty. Even the modern diesel engine, which is sold widely in Europe, must be augmented with new emission-control equipment (e.g., a particulate trap and NOx catalyst coupled with ultra low-sulfur diesel fuel) in order to be sold in the United States (especially in California and four other states). Our nominal cost assumption for the additional emissions control ($500 per vehicle plus a 3% fuel-economy penalty) may underestimate overall costs of advanced diesel emission controls.⁽²¹⁾

On the other hand, the costs of both diesel and hybrid technology could decline significantly over time. Diesel engine costs could decline as advanced emission controls are refined and optimized with engine performance. Hybrid costs have already declined significantly and could decline further with breakthroughs in battery technology.⁽²⁸⁾ The future relative costs of the two technologies become a key issue, one that is perplexing many decisionmakers in the automotive industry.

In Fig. 1, we display a sensitivity analysis for the passenger car using pairs of different technology cost estimates for diesel and hybrid vehicles. The technology cost sensitivities could also represent the inherent uncertainty in the cost of emissions controls (for the diesel) and battery costs (for the hybrid). The relationship between relative technology costs and NPV is revealed. The vertical axis represents the additional cost for the diesel, while the horizontal axis represents the additional cost for the hybrid. The diagonal line across the middle of the graph shows all pairs of cost estimates for which the NPV of the two technologies are equal (holding all other input assumptions at their nominal values). For pairs above the diagonal, hybrids are superior to diesels; for pairs below the diagonal line, diesels are preferred to hybrids. The dashed horizontal line represents the breakeven technology cost for the diesel. If the cost of the diesel falls above this line, the diesel NPV will be negative; below the line, it is positive. The dashed vertical line, in turn, defines the breakeven points for the hybrid. The dashed line from each axis to the diagonal line shows where the NPV is equal to zero ($3,698 for hybrids and $2,760 for diesels in the case of passenger cars). In a separate calculation, we determined that the costs of the hybrid must fall below $3,238 in order for the NPV of the hybrid to exceed the NPV of the diesel (setting all other input values at our nominal assumptions). For SUVs and large pickups, those switch point values for the hybrid costs are $4,017 and $3,256, respectively.

Though not detailed here, we consider different analytic treatments of the discount rate (3%, 5%, 7%, 9%, and 11%) used to express future costs and benefits in present value. As expected, lower discount rates help the diesel and hybrid engines (since their benefits are spread out over the long life of the vehicle) while larger rates help E85 when compared to hybrids and diesels. However, even the high discount rate did not change the rank ordering of the technologies for any vehicle class.

4.2. Robustness of Societal Results

We conducted NPV sensitivities to fuel prices and multiple externality costs that play a role in the societal perspective. As in the private case, societal NPVs are sensitive to the fuel price assumption—and in exactly the same way. Though each of the externality costs is quite uncertain, the assumptions have surprisingly little impact on overall NPV results. Since concerns about climate change and energy security have motivated a national discussion about which engines and fuels best serve our interests, we report the results of these sensitivity analyses below.

The damages from greenhouse gas emissions are not known with any certainty. In Fig. 2, we report passenger-car NPVs using widely different estimates of the per ton damages from carbon dioxide. The rank ordering of diesels, hybrids, and E85 holds even as the per ton cost of carbon dioxide rises dramatically (our nominal assumption begins at $5.50 per ton and rises to $15 per ton in 2030). Even at the high end of the sensitivity range, NPV for diesels and hybrids is only approximately $1,000 and E85 is still negative. For SUV and pickup applications, the rank ordering is even more robust with respect to differing environmental inputs.

The reductions in greenhouse gas emissions from diesels and hybrids are due to the higher fuel economy, which decreases the amount of fuel used by the driver and the associated emissions that occur “upstream” (extraction, transport, and refining of oil). E85 greenhouse gas benefits are moderated by two facets of the current corn-based ethanol production process. A significant amount of emissions are associated with the fertilizer used to grow corn, and many of the ethanol plants rely heavily on coal as a fuel.⁽²⁹⁾

The nominal estimate of the “energy security” externality is 32 cents per gallon and is based on the economic ramifications of U.S. oil dependence. This figure could easily be too high or too low by 30%, but differences of this magnitude do not change the rank ordering of the four technologies. In Fig. 3, we show the sensitivity analysis of energy security costs for passenger cars.

Fig. 3 shows that the energy security externality can significantly impact the results for E85. As the energy security externality rises, the NPV of all three technologies improves, but the NPV of E85 improves disproportionately. At a value of $0.80 per gallon, E-85 has a positive NPV (for the passenger car). With an externality cost of $1.14 per gallon, E85 NPV exceeds both hybrids and diesels (see Table VIII). These externality costs exceed most estimates of this value, but not all.⁽³⁰⁾ The figure also shows that, even with externality costs approaching $1.75 per gallon, hybrids do not gain an advantage over diesels.

Table VIII. Switch Points in NPV from Energy Security Externality Cost

	Hybrid	Diesel	E85	E85 Best Option
Passenger car	62	0	80	114
SUV	0	0	87	163
Pickup truck	20	0	89	–

In contrast to the results on greenhouse gas emissions, Fig. 3 shows that E85 NPV improves faster than the other alternatives with increases in the energy security externality, and indicates E85 can have considerable energy security benefits. E85 reduces oil consumption with only limited impacts on greenhouse gas emissions because producing corn-based ethanol is an energy-intensive process where most of the energy comes from natural gas and coal. Some petroleum-based energy is used in growing corn and in transportation during intermediate and final stages of the product; however, petroleum use is still relatively small compared to consumption of other fossil fuels.⁽²⁹⁾

Our approach to the energy security externality does not account for any effect of U.S. oil consumption on U.S. military expenditures, U.S. flexibility in foreign policy, the financing of terrorists by oil-exporting regimes, or the financing of insurgents in Iraq by oil-exporting regimes. Some analysts argue that lowering U.S. oil consumption would drive down the price of world oil, which would reduce oil revenues to countries such as Iran, which are believed to be financing terrorists, insurgents in Iraq, and instability in the Middle East.^(31–34) A counterargument is that lowered world oil prices would also hurt moderate regimes that are sympathetic to U.S. interests in the Middle East and the world (e.g., Saudi Arabia and Mexico). Moreover, terrorism is such a low-cost activity that it is not clear that marginal reductions in U.S. oil consumption would have any discernible effect on the financing of terror, the strength of the insurgency in Iraq, the rate of U.S. military expenditures, or the conduct of U.S. foreign policy.^(23,35)

With regard to conventional air pollution, damage estimates for each pollutant reflect mid-range literature estimates, but we also examined how much NPV changes when we altered the input values for damages from conventional pollutants. The results (unreported) do not reveal much sensitivity to pollution damage estimates. Qualitatively, larger damage estimates for particulate matter and nitrogen dioxide hurt the diesel engine relative to the hybrid engine.⁽³⁶⁾ Given that new vehicles emit relatively few emissions for all four engine/fuel combinations, it is unlikely that alternative damage estimates would play a significant role in reordering the technologies.

In our societal analysis, we captured what were in our judgment the key social impacts from the technologies. However, as in any complex analysis, we could not include all possible effects. For instance, we have not quantified any water-pollution (or aquifer depletion) impacts that may occur from growing corn for ethanol production or mining nickel for use in hybrid batteries. Even in cases in which our NPV estimates for technologies are positive, there may be other uses of the consumed resources that would produce an even higher return for society (e.g., improved vehicle safety or engine performance). It has been observed that, historically, major improvements in vehicle fuel efficiency have been traded in vehicle design for enhancements to performance, vehicle size and weight, and vehicle features that consumers value.⁽³⁷⁾

5. RECENT MARKET AND POLICY DEVELOPMENTS

At current fuel prices, our benefit-cost model ranks the alternatives as follows: diesels, hybrids, gasoline, and E85. In this section, we briefly summarize what is actually happening in the marketplace and in public policy.

6. PORTFOLIO STRATEGY

Given the large degree of uncertainties and different consumer preferences revealed in this article, it is not necessary for decisionmakers to choose only one of the alternative technologies. The United States is already pursuing a portfolio strategy, where each of the technologies is being implemented to some extent, and new decisions will be made as experience accumulates. Our findings show that consumers internalize the majority of the benefits and costs of these technologies. Therefore, public policy should ideally focus on setting the correct incentives for market participants and allowing the best portfolio of technologies to emerge through competitive market responses. More research is needed on the full range of government policies that are favoring or retarding diffusion of these technologies.^(54–57) Our results do show that diesels and hybrids provide net benefits to society under a considerable portion of the range of uncertainties considered. E85, though, had largely negative results and decisionmakers should reconsider current strong support for this costly alternative. However, if technological breakthroughs can significantly reduce the cost of producing ethanol, E85 may have the potential to compete in the market and provide net benefits to society.