• Valuationof fuel savings. A central assumption of the modelis that consumers undervalue fuel savings. Althoughthere is evidence to this effect, there is no informationon the magnitude of the undervaluation. Sensitivityanalysis is done to investigate the effect.   
• Consumerelasticities. These values determine the extent towhich consumers respond to price signals. Since thereis no data on Canadian elasticities, our approachhas been to use the elasticities proposed by Greeneet al. but to halve them, as a way of approximating long-run responses and to better represent assumed Canadian circumstances. We also undertake a sensitivity analysis.   
• Hybridand diesel technologies are not included. As mentionedabove, hybrids and diesels are expected to play asignificant role in improving fuel economy. Unfortunately,the current version of the model lacks the informationnecessary to include these options. As a result, the  effectiveness of feebates is underestimated.   
• Effectson used vehicle markets are not modeled. As mentionedabove, the used vehicle market represents more thanhalf of vehicle sales. However, the model assumesthat consumers primarily respond by shifting purchasesto other new vehicles. As a result, the effectiveness of feebates is overestimated.   

• Most scenarios result in significant fuel savings and GHGreductions.      
• Most scenarios produce a net economic benefit, mostly inthe form of unvalued fuel savings.      
• Most scenarios produce a significant shift in sales buttechnology still accounts for more than two thirdsof the improvement.     
• GHG reductions and the shift in sales increase relativelylinearly with an increasing feebate rate.      
• Benefits are positive for all rates but level off between $500and $1000.      
• A cap removes incentives for highly inefficient vehiclesto improve, since a fixed fee is paid on them.      
• A deadband removes incentives to improve fuel economyfor vehicles near the pivot point.      
• If consumers are assumed to fully value fuel savings,the base case becomes more advantageous and the benefitsof feebates are correspondingly reduced. Fuel economystill improves but there is a net cost per tonne ofGHG emission reductions.   
• Differentiating pivot points for cars and trucks means a lower feeor even a rebate as larger vehicles are assessed onlyagainst their cohorts. This discourages shifting tosmaller vehicles (less change in market shares) andmeans less improvement in consumer surplus and fewer GHG reductions. Going to 11 classes has little additional effect.   
• With North American implementation, all vehicles improveaccording to the full change in consumer willingnessto pay. As a result, GHG reductions are larger andthe change in Canadian surplus is larger.    
• With Greene’s original elasticities, consumers aremore sensitive to price changes. As a result, thesales mix changes more easily but policies are lesscostly, since consumers take greater advantage ofthe option to purchase other vehicle types. This doubles the size of the sales shifts and the GHG reductions.   
• A higher fuel price means that there is more incentivefor fuel economy present in the base case. Thus, theincremental of the feebate on fuel and GHG savingsis reduced. On the other hand, the unvalued fuel savingsare worth more, so the overall benefit is higher and the benefit per tonne is greater. The price of fuel has little impact on the sales mix.   

• Feebates will encourage additional investment in fuel-efficiencytechnology and shift the market towards more fuel-efficientvehicles (trucks to cars, large cars to small cars,more fuel-efficient cars in the same class).    
• Over time, this will improve the fuel efficiency of thevehicle stock and will reduce GHG emissions.      
• The investment in fuel-efficiency technology will raisethe cost of individual vehicles and reduce consumersurplus accordingly.     
• To the extent that consumers undervalue fuel savings,feebates will capture savings that would otherwisenot have been realized. If the undervaluation is significant,over the life of the vehicle fuel savings are likelyto exceed the added cost to vehicles, resulting in  a net economic benefit to society.   
• Higher prices will depress vehicles sales.       
• The shift towards more fuel-efficient vehicles will alsoreduce overall revenues.      
• In a single-class feebate, General Motors (GM), Ford,and DaimlerChrysler (DCX) will lose additional marketshare and bear a disproportionate share of the adjustmentcosts. This could be alleviated by adopting separateclasses for trucks and cars, though doing so would reduce GHG savings and economic benefits.   
• The extent of the shifts is determined by the elasticitiesof demand. If elasticities are greater than expected,the environmental and economic benefits will be greaterbut so will the adjustment costs. Conversely, if elasticitiesare less than expected, the environmental and economic benefits will be reduced, as will the burden on manufacturers.   

Assessment of Feebate Options

• EnvironmentalEffectiveness. Feebates are less well-targetedthan alternatives such as fuel taxes but the mainconcern, the rebound effect, is expected to be nomore than 23 percent, and probably less. Consumersare expected to respond by switching from trucks to cars, from larger vehicles to smaller vehicles, and to more fuel-efficient vehicles within a given class.The car share of the vehicle market is expected to increase by 1–6 percent, depending on the rate. Manufacturers will respond primarily by investing in cost-effective technologies; however, some less fuel-efficient models may be dropped. The combination of technology improvements and shifts in purchasing is expected to yield fuel consumption improvements of 0.2 litres per 100 km to 0.8 litres per 100 km. Corresponding GHG reductions are expected to range from 1.5 Mt per year to 6.2 Mt per year, with 3.0 Mt per year for a $500 per litre per 100 km feebate. Adopting two or more separate classes would significantly lower the GHG reductions (because these are cumulative).The environmental effectiveness (as well as the revenue neutrality) of feebates could be compromised by the possibility of vehicle “arbitrage”—the import of relatively new large vehicles from the US and the export of relatively new smaller vehicles to the US. 

• FiscalImpact.By definition, the measure will berevenue neutral. However, there will be administrationcosts and reductions in fuel taxes to consider. Thesecould be recovered from the feebate, but this couldbe perceived as a tax increase. It will also be necessary to consider options on how to handle the uncertainty associated with having to predict the overall balance between fees and rebates.   
• EconomicEfficiency. To the extent that consumersfail to value fuel savings correctly—this issupported by the market research in Canada, thoughthe extent is unknown—feebates provide a meansof correcting this perception. Feebates also provide an indirect means of giving value to GHG reductions. Feebates impose costs which rise as the rate increases, but the reduction in consumer surplus is more than compensated for by unvalued fuel savings that are realized. The benefits are positive for all rates up to $1000 but marginal costs begin to outweigh benefits between $500 and $1000. Adopting two or more classes reduces the benefits significantly while creating a relative subsidy for larger vehicles. Because of the unvalued fuel savings, feebates produce economic benefits as opposed to costs. These range from $40 per tonne for a $250 per litre per 100 km feebate, to $10 per tonne for a $1000 per litre per 100 km feebate. If it is assumed that consumers already fully value fuel savings, then there are no unvalued fuel savings and the costs are in the range of $10 per tonne. By selectively targeting fuel economy, feebatesimpose opportunity costs. Feebates will only affect vehicle sales in Canada, so there should be no impact on exports. Furthermore, feebates should have no impact on the environment for manufacturing. However, as noted previously, an environment interpreted as hostile to the product could affect investment decisions. Overall vehicle sales are expected to decline slightly(at most 6000 or approximately 0.5 percent of annual  sales for a $1000 feebate). Of greater importance is the shift to less expensive models, which overall would reduce revenues by approximately $1.5 billion per year. (Note: these results are very sensitive to elasticity assumptions.) Although net sales may only decrease slightly, the employment impacts could be greater if imports are substantially increased. On the other hand, a large proportion of the North American adjustment may occur in the US. Given the overall economic benefit, the loss of jobs in this industry should be more than offset by job gains elsewhere in the economy.  
• Fairness.In terms of market share, the main impact is furtherloss in market share for GM/Ford/DCX. The shift increasesas the rate increases, reaching 4 percent for a $1000per litre per 100 km feebate. This shift can be significantlymitigated by segmenting the market into two classes.(Having 11 classes does not make much additional difference.)  As far as profitability is concerned, the assumption is that all costs and savings are passed on, and so profits are unchanged. However, since there will be a shift to smaller vehicles, and historically these vehicles have had lower profit margins, it is reasonable to conclude that profits will be adversely affected. As far as parts suppliers and retailers are concerned, they will be affected in proportion to their exposure to GM/Ford/DCX. For individuals, the key issue is price. The price of each individual vehicle will rise to pay for new technology. However, consumers are expected to shift to lower priced models within classes and to lower priced classes overall, so average prices will decline. Certain consumers who are unable or unwilling to shift will bear a greater burden. For example, the estimated 50 percent of consumers who use trucks for commercial purposes may not be able to avoid the higher fees. Similarly, larger families may be restricted in shifting to smaller vehicles. Regions and areas that have a greater preference for larger vehicles (western Canada and rural areas, for example) will find that their traditional choices are more costly. Conversely, consumers who would have purchased fuel-efficient vehicles anyway will gain a windfall. Because lower income households tend to purchase smaller vehicles, the measure is progressive overall. (Feebates will eventually influence prices in the used car market as well as new cars.)  
• Simplicity.The size of the transfers will range from approximately$300 million per year to $1.1 billion per year, whereasthe number of transactions will be equal to the numberof new vehicle sales (1.5 million per year). A singleclass would clearly be the simplest approach, whereas11 classes could be cumbersome to manage. Because definitions are unclear, anything more than one class creates the potential for gaming (artificially changing features to move vehicles into a different class). Similarly, the use of a cap, plateau or deadband would introduce added complexity and induce responses that would reduce the effectiveness of the measure. In terms of administrative practicality and costs, the measure could be similar to the Goods and Services Tax (GST). Retailers would need to collect the fees, pay the rebate, and submit the appropriate paperwork on a regular basis. Overall, given the experience of the GST, it would be anticipated that costs would be significant at first but would fall substantially after the initial implementation.  

Conclusions

• Feebatescan be designed to be environmentally effective andeconomically efficient. Although other measures suchas fuel taxes may be better targeted, feebates area legitimate alternative should other measures notbe feasible.   
• Theimposition of feebates will involve difficult adjustmentsfor automobile manufacturers at a time when the industryis faced with the challenge of oversupply. GM, Fordand DCX will bear most of the burden.    
• Themeasure is administratively feasible and can be designedto be fiscally neutral.      
• Thereare significant uncertainties and risks that affectthe magnitude of the benefits as well as the marketshifts involved.