Volume 39 Issue Supplement 2, August 2013, pp. S81-S94

The most effective policy to address environmental externalities from vehicular fuel use is an appropriate fuel tax. Instead of raising fuel taxes, Canada’s provincial and federal governments prefer to subsidize the purchase of fuel-efficient vehicles and the accelerated retirement of old vehicles. Are these programs effective? Can they be improved? We argue that subsidies for hybrids and electric vehicles are not cost-effective and instead recommend building on Canada’s brief and modest experience with ”feebates.” While British Columbia’s pioneering accelerated vehicle retirement program is cost-effective, its success rests significantly on inducing participants to switch from personal vehicles to alternative transportation modes. Policy refinements will be needed to adapt the lessons learnt from this program to less favourable conditions elsewhere in Canada.

Nous savons que la meilleure manière de réduire les coûts externes reliés à l’utilisation de véhicules fonctionnant au carburant est une taxe sur le carburant. Toutefois, plutôt que d’augmenter cette taxe, les gouvernements fédéral et provinciaux, au Canada, préfèrent subventionner l’achat de véhicules plus économiques en essence et la mise hors service plus rapide des vieux véhicules polluants. Mais ces mesures sont-elles efficaces ? Pourraient-elles être bonifiées ? Dans cet article, nous avançons que les subventions à l’achat de véhicules hybrides ou électriques ne sont pas rentables; nous recommandons plutôt la mise en place d’une combinaison de frais à l’achat de véhicules énergivores et d’une remise à l’achat de véhicules consommant moins de carburant (« feebate », en anglais). Par ailleurs, les mesures qui visent à retirer plus rapidement les vieux véhicules de la circulation, comme celles qu’a adoptées la Colombie-Britannique, sont rentables sur le plan environnemental, mais il faut préciser que leur succès repose sur le fait qu’elles encouragent les citoyens à utiliser des modes de transport plus verts que l’utilisation de leur véhicule personnel ; il faudra revoir ce type de politique pour l’adapter aux conditions moins favorables qui prévalent ailleurs au Canada.

Increasingly, governments rely on monetary incentives (direct subsidies or tax rebates) to encourage consumers to lower their environmental impact from personal transportation. Drawing on recent research, we review two prominent examples of market-based policies towards ”green” motoring in Canada: 1) subsidies towards the purchase of fuel-efficient vehicles, and 2) the accelerated retirement of older polluting vehicles. Do these programs work as intended? Are they a cost-effective use of scarce government resources? How can they be improved?

The most efficient economic intervention for mitigating an environmental externality is to price the externality directly and appropriately. A fuel tax is the most practical way to put a price on the environmental impact from fuel consumption. Vehicular burning of fuel emits the greenhouse gas carbon dioxide (CO2) and generates (mostly local) air pollution in the form of carbon monoxide (CO), nitrogen oxides (NOx), hydrocarbons (HC), sulfur dioxide (SO2), and particulate matter (PM).1 In the short run, a fuel tax raises the marginal cost of every kilometre driven, encouraging people to drive less in whichever car they own.2 This reduces emissions, and also reduces congestion and accident costs. In the long run, a fuel tax induces the purchase of more fuel-efficient vehicles. Fuel taxes are easy to implement, exist in almost all jurisdictions, and adjusting their level is cheaper than introducing and implementing new policies.

Nevertheless, governments rarely raise fuel taxes for environmental reasons. Instead they prefer to grant subsidies for purchasing fuel-efficient vehicles or retiring older vehicles—thus generating environmental credibility with small budgetary outlays.

We argue that subsidies towards purchasing fuel-efficient vehicles (specifically hybrid and electric vehicles) are not cost-effective. Instead, we recommend a ”feebate.” Recognized as the best alternative to a fuel tax, a feebate combines both a fee for fuel-inefficient vehicles and a rebate for fuel-efficient vehicles. A feebate approximates a fuel tax on vehicle purchases. It can be implemented at the level of consumer, producer, or retailer. Finally, it can be revenue-neutral, if desired.

Our study of British Columbia’s accelerated vehicle retirement program (BC SCRAP-IT) shows that this program is effective at reducing local emissions. It also shows that future programs of this kind are more likely to be cost-effective if focused on urban areas, relative to rural areas, because urban areas have a higher concentration of ambient pollution, a higher vehicle density, and a greater likelihood that participants will switch out of vehicle ownership entirely.

During the first decade of the 2000s, many governments in the United States and Canada introduced incentives encouraging hybrid electric vehicle (HEV) adoption (Gallagher and Muehlegger 2011). While some of these incentives have since lapsed, similar—although less numerous—incentives supporting electric vehicles (EVs) have taken their place.

Encouraging the Purchase of Hybrids in Canada

Starting with British Columbia, several provincial governments (Ontario, Manitoba, Quebec, and Prince Edward Island) introduced incentives for purchasing hybrid electric vehicles. In Table 1, we provide details of vehicle eligibility, rebates, limits, and the time horizons for these programs. Mostly, the provinces rebated their provincial sales tax up to a certain maximum. All but two of these programs have since been discontinued.

The cost-effectiveness of such programs depends on two elements: the existence of free riders and an improvement in fuel economy for those switching to hybrids. The term ”free rider” describes those buying HEVs irrespective of the incentive—people whose behaviour is unchanged by the subsidy. To be cost-effective, a program should induce change: it should both minimize free riding and induce improvements in fuel economy. Ideally, the subsidy should induce people to buy fuel-efficient, compact vehicles instead of large sport utility vehicles (SUVs).

Using monthly US state registration data, Diamond (2009) evaluates the impact of state

 Table 1 Tax Rebate Programs for Hybrid Vehicles in Canadian Provinces

Table 1 Tax Rebate Programs for Hybrid Vehicles in Canadian Provinces

Province Vehicle Eligibility Rebate Amount and Timing
British ColumbiaStart: August 2000End: July 2010PST rate: 7%, with graduated increases for vehicles over $55K All hybrid vehicles with regenerative breaking (cars and SUVs) 30% of tax paid up to$500 for vehicles bought before 31 July 2001; 30% of PST paid up to $1,000 after 31 July 2001;a point-of-sale reduction of all PST up to$2,000 after 16 February 2004. Additional rebates in PST (reductions in graduated increase of PST over 7%) for hybrid vehicles over $55K.a Prince Edward IslandStart: March 2004PST rate: 10% All hybrid vehicles PST paid up to$3,000 for vehicles bought after 30 March 2004.
OntarioStart: May 2001End: July 2010PST rate: 8% All hybrid passenger cars with regenerative braking; hybrid SUVs eligible 2002 PST rebate, up to $1,000, for cars bought after 10 May 2001. Hybrid SUVs and trucks included 18 June 2002. Point-of-sale reduction of all PST up to$2,000 after 23 March 2006.
QuebecStart: March 2006End: December 2013QST rate: 7.875% By model/yearb QST paid to a maximum of $1,000 for vehicles bought after 23 March 2006; QST paid to a maximum of$2,000 for vehicles bought after 22 February 2007, declining to $1,000 after 31 December 2011. ManitobaStart: November 2006End: October 2010PST rate: 7% By model/yearc Flat$2,000 rebate for all vehicles bought after 15 November 2006.

Notes: PST=provincial sales tax. QST=Quebec sales tax. SUV=sport utility vehicle.

aPeople buying light vehicles that are priced greater than $55,000 have to pay a higher PST rate.This rate increases by 1 percent for the first$1,000 over $55,000 and continues to increase by 1 percent for every additional$1,000 to a maximum of 10 percent (for vehicles costing more than $57,000). For hybrid vehicles the graduated increases come with an additional exemption of$7,000 on the threshold. This implies that the PST does not increase for hybrid vehicles till their price reaches $62,000. bCars eligible for a rebate in Quebec are: 2005–2006 Honda Insight; 2005–2007 Toyota Prius; 2007 Toyota Camry Hybrid; 2008 Ford Escape Hybrid (two-wheel drive); 2005–2007 Honda Civic Hybrid; 2005 Honda Accord Hybrid; 2007 Nissan Altima Hybrid. cCars eligible for a rebate in Manitoba are: Honda Insight; Lexus GS 450H; Lexus RX 400H; Toyota Camry Hybrid; Toyota Highlander Hybrid; Toyota Prius; Chevrolet Silverado 1500 LS Hybrid; Ford Escape Hybrid; GMC Sierra 1500 SLE Hybrid; Honda Accord Hybrid; Honda Civic Hybrid; Saturn VUE Green Line. Source: Numerous documents hosted on provincial government websites. incentives and gasoline prices on HEV adoption. He finds that while gasoline prices have a strong relationship with adoption, government incentives have a weaker link. Using quarterly sales data on HEV models for US states to study the relative effect of tax incentives, gasoline prices, social preferences, and other non-monetary incentives (for example, preferential access to high occupancy lanes and parking, etc.), Gallagher and Muehlegger (2011) attribute only 6 percent of existing hybrid sales to tax incentive schemes, approximately 33 percent to preferences, and 27 percent to rising gas prices. Beresteanu and Li (2011) evaluate the effect of gas prices and government incentives on the purchase of HEVs in 22 US Metropolitan Statistical Areas (MSAs). They find that rising gas prices from 1999 to 2006 explain 14–17 percent, and government subsidies, 27–32 percent of the diffusion of hybrid vehicles in these areas. In the most comprehensive study of Canada’s provincial rebates for HEVs, Chandra, Gulati, and Kandlikar (2010) estimate that one-quarter (26 percent) of the HEVs sold during the rebate programs were attributed to the rebate. This implies that there were no environmental benefits associated with three-quarters of the subsidy. While the highest rebate offered was$3,000, the sticker price of the largest selling hybrid in Canada—the Toyota Camry Hybrid—was approximately $32,000. At almost 10 percent of the sale price of the vehicle, the rebate induced one-quarter of a very small market to switch.3 In the absence of the rebate, Chandra, Gulati, and Kandlikar (2010) found that those induced to switch would have bought vehicles with similar characteristics anyway. In fact, HEV rebates induced fewer purchases of intermediate passenger cars, intermediate SUVs, and high-performance compact cars (for example, the Honda Civic Si series, the Corolla Sports versions, and Volkswagen GTI). Consumers buying intermediate cars obtained similar vehicle characteristics by choosing the Camry Hybrid, or Prius. Those wanting intermediate SUVs bought the Highlander Hybrid, or Escape Hybrid. Finally, those buying high-performance compact cars switched to a similarly sized hybrid at approximately the same price. Unsurprisingly, people did not give up the Hummer for a Prius, and consequently, the induced gasoline savings were small. Most HEV rebate programs subsidize consumers who would buy hybrids or other fuel-efficient vehicles in any case. Consequently, the cost per litre of gasoline saved, or per tonne of carbon dioxide saved, is very high. Chandra, Gulati, and Kandlikar (2010) estimate that a tonne of CO2 saved through these programs cost provincial governments$195. This compares unfavourably with benchmark prices for CO2 reduction. The average price of a futures contract for a tonne of CO2 at the European Climate Exchange in 2008 was 25.50 Euros (approximately 40 Canadian dollars at the exchange rate for 2008). Lutsey and Sperling (2009) estimated the abatement cost of reducing a tonne of carbon dioxide by converting 50 percent of the current (2009) vehicle fleet to hybrid electric vehicles by 2025 to be 42 US dollars.

Other benefits from hybrid rebates could justify the costs of the program. A rebate program might accelerate the diffusion of HEV technology and induce economies of scale in vehicle production. This argument can justify a program in a large vehicle-buying economy like the United States of America, but it is not valid for Canada.4 HEVs are associated with lower local air pollutants (Fontaras, Pistikopoulos, and Samaras 2008) and their associated benefits. Finally, local hybrid cars generate network externalities that provide valuable information for non-adopters in the same area. This can spur subsequent purchases of HEVs. We cannot quantify these benefits to ascertain their importance. However, one caveat remains. Unless the number of free riders drops significantly, these programs are likely to remain costly relative to their benefits.

Incentives for Electric Vehicles

In recent years the focus has switched from HEVs to EVs. Table 2 provides the details of EV programs in British Columbia, Ontario, and Quebec. While HEV programs relied on tax rebates, EV programs provide point-of-sale cash incentives.

While the subsidies are sizable, the price of EVs remains significantly higher than that of comparable non-electric vehicles. For example, Nissan’s ”Leading, Environmentally friendly, Affordable, Family car (LEAF)” starts at $38,395 for the 2012 model year. Because of constraints imposed by long charging times and limited charging infrastructure,  Table 2 Subsidy Programs for Electric Vehicles in Canadian Provinces Table 2 Subsidy Programs for Electric Vehicles in Canadian Provinces Province Vehicle Eligibility Rebate Amount and Timing British ColumbiaStart: December 2011End: March 2013 All battery electric vehicles, plug-in vehicles, and fuel cell vehicles, purchased or leased Point-of-sale cash incentive from$2,500 to $5,000, depending on battery power and fuel used OntarioStart: July 2010End:open All conventional plug-in electric vehicles, purchased or leased Cash Incentive from$5,000 to $8,500, depending on battery power QuebecStart: January 2012End: December 2015 Plug-in hybrids and electric vehicles Point-of-sale cash incentive of$5,000 to $8,000, depending on battery power Source: Documents hosted on provincial government websites. most people are unlikely to choose an EV as their primary household vehicle. Its high price and limited usability likely restricts purchases to environmentally motivated, affluent buyers. We expect incentive uptake to be concentrated among free riders—affluent households with strong environmental motivation. As EVs are not appropriate for very high daily usage, reductions in fuel use from inducing a switch to an EV, if any, will also be small. Drawing from the literature on HEV subsidies, a high proportion of free riders coupled with low driving will likely imply low gains from the subsidy. There might even be an environmental loss from promoting an EV. Moreover, in regions that depend heavily on coal for electricity generation, well-to-wheel emissions from EVs may even be higher than those from HEVs.5 EV owners in Ontario also get green licence plates allowing unrestricted access to High Occupancy Vehicle (HOV) lanes. Bento et al. (2012) study a similar Clean Air Vehicle Stickers program in California. This program provides unrestricted HOV access to a limited number of hybrid vehicles. Even though the number of stickers is relatively small, they find significant congestion effects. Driving time in HOV lanes rises by 9 percent in morning peak hours. The authors argue that the combination of high congestion costs and restrictive benefits makes this policy very inefficient for transferring benefits to hybrid owners, approximately five times less efficient than cash incentives. While the tiny number of EVs on the road in Ontario might not induce congestion in HOV lanes, the specificity of benefits makes this policy less efficient than comparable cash incentives. Whether it be cash incentives funded by taxpayers, or reduced commuter times underwritten by carpoolers, the environmental case for subsidizing EVs is weak. Feebates Consumers desire a set of vehicle characteristics determined by family and individual needs. Medium-sized monetary incentives ($1,000–$3,000) induce a switch within a vehicles class (with similar characteristics), but are unlikely to switch people across different vehicle classes. However, if an incentive for fuel efficiency is coupled with a levy for fuel inefficiency, government expenditure on the program falls, and more people are induced to switch. In such cases cost-effectiveness can be high. The policy described above is a ”feebate.” Ontario introduced North America’s first automobile feebate in 1991. Born out of the 1989 Tax on Fuel-inefficient Vehicles, the Tax for Fuel Conservation (TFC) provided a rebate of$100 for passenger cars with fuel consumption of less than 6.0 litres per 100 kilometres (/100km), and an increasing tax for vehicles with fuel consumption above 6.0 litres/100km. The tax was higher for passenger vehicles than SUVs with an equal fuel consumption (see Table 3). It was a modest application of the concept. A large proportion (approximately 90 percent) of vehicles sold in Ontario were subject to a flat tax of $75 (Bregha and Moffet 1995). This tax and the rebate ($100) for fuel-efficient vehicles were small, relative to the price of new vehicles. They are unlikely to have influenced vehicle sales significantly.

The Vehicle Efficiency Initiative (VEI) from March 2007, the first national automotive feebate in Canada, combined a fee for fuel-inefficient vehicles (the greenLEVY) with rebates for fuel-efficient vehicles (the ecoAUTO rebate). The ecoAUTO program offered between $1,000 and$2,000 with a combined city/highway fuel consumption lower than than 6.5 litres/100km for cars, and 8.3 litres/100km for light trucks. A total of 19 vehicle models (cars and truck) were eligible from model year 2006 and 2007, and 28 were eligible from model year 2008. Flex-fuel vehicles, running on gasoline and ethanol (E85), were also rebated if their fuel consumption was rated as less than 13 litres/100km. Four vehicles from model years 2006, 2007, and 2008 were eligible. The ecoAUTO program ended in December 2008. The greenLEVY on Fuel-inefficient vehicles imposes a tax starting at $1,000 for combined fuel consumption ratings between 13 and 14 litres/100km. This increases by$1,000 for every additional litre consumed up to 16 litres/100km, and vehicles above this threshold pay a maximum of $4,000. The greenLEVY is still in place; however, most vehicles sold are not subject to it. While rebates  Table 3 Ontario's Tax for Fuel Conservation Table 3 Ontario's Tax for Fuel Conservation Highway Fuel Rating (Litres/100 km) Tax (Rebate) New Passenger Cars New Sport Utility Vehicles Under 6.0 ($100) $0 6.0 to 7.9$75 $0 8.0 to 8.9$75 $75 9.0 to 9.4$250 $200 9.5 to 12.0$1,200 $400 12.1 to 15.0$2,400 $800 15.1 to 18.0$4,400 $1,600 Over 18.0$7,000 $3,200 Source: Ontario Ministry of Finance (2001). under the VEI were larger than under the Ontario feebate, the overall scope of the program was still relatively small. We illustrate the concept of a feebate in Figure 1.6 The pivot point illustrated is chosen for revenue neutrality, and thus the areas for the fee and rebate are equal. All consumers buying vehicles with fuel consumption below the pivot point receive a rebate, and vehicles with fuel consumption greater than the pivot point need to pay a fee. The slope of the line passing through the pivot point determines the shadow price for fuel consumption. If this slope is steeper, there are greater incentives to purchase fuel-efficient vehicles. A feebate approximates a fuel tax on vehicle purchases.7 Feebates simultaneously reward virtuous behaviour and penalize profligate behaviour. If the pivot point is set appropriately, feebates can generate revenue, be revenue-neutral, or have a subsidy outlay.8 Anderson et al. (2011) argue that a feebate can be adjusted for uncertainty of fuel prices. Further, a feebate can be implemented at consumer, producer, or retailer levels to optimize administrative simplicity. A comprehensive feebate proportional to fuel consumption induces the equality of marginal compliance costs for vehicle fuel economy across producers. This cannot be achieved with the fuel economy standards recently adopted by Canada. By being comprehensive and encouraging optimal vehicle choice within each class, the feebate will have fewer free riders. If the feebate is proportional to fuel consumption, the feebate will raise the price of fuel consumption and thus encourage fuel savings. Lastly, as the feebate need not imply a subsidy outlay, it will almost certainly be more cost-effective relative to subsidies for fuel-efficient vehicles. Policy-makers in Canada should revisit feebates as a policy to promote fuel efficiency. They should aim to improve the Ontario TFC and the federal VEI by Figure 1 Revenue-Neutral Feebate Source: Illustration designed by authors. doing away with their multiple pivot points, discrete rebate, and tax brackets. Including one pivot with proportional fees and rebates can widen the scope of the program beyond the marginal vehicles targeted earlier. The pivot point can be reviewed periodically to ensure the revenue neutrality of the program and provide vehicle manufacturers with continuing incentives to innovate and improve fuel efficiency. Voluntary accelerated vehicle retirement (AVR) programs, or scrapping programs, are used in jurisdictions across North America. Primarily intended to reduce emissions, the programs were also used during the recent recession to stimulate vehicle demand. Vehicles emit a variety of primarily local air pollutants (CO, HCs, NOx, and SO2 and PM from diesel combustion), impacting both public health and the environment. Recently there has been a greater focus on CO2 due to its role as a greenhouse gas contributing to climate change. Except for CO2, vehicle emissions have declined rapidly due to advances in vehicle technology. The removal of lead from gasoline and the introduction of catalytic converters were key milestones. Modern three-way catalytic converters (introduced in the early 1980s) reduce emissions of CO, HC, and NOx by converting them into carbon dioxide, nitrogen, and water. Electronic engine control systems further reduce emissions by fine-tuning the fuel-oxygen mix and ensuring a more thorough combustion. Onboard computers alert drivers to failures in these systems. Compared to their newer counterparts, older (pre-1995) vehicles have less sophisticated emission-control systems. While some jurisdictions (especially in urban areas) have mandatory testing of vehicle exhaust systems, this only ensures that the most polluting vehicles are forced to undergo repairs. As regulators tighten emission standards for new vehicles, older vehicles—effectively grandfathered—continue to generate emissions at higher rates. This gap is substantial. A recent study of 133 randomly selected scrapped vehicles and 15 new vehicles conducted by the AirCare program in the Greater Vancouver area of British Columbia (see Table 4) shows that scrapped vehicles have emissions factors eight to ten times greater than new cars for CO and HC, and about 50 times greater for NOx. Clearly monetary incentives to retire old vehicles generate positive environmental benefits. However, to determine the benefit of AVR programs, we need to know two things. First, how many more years would the scrapped vehicle be driven in the absence of the incentive? This is not easily observed and needs to be estimated. Second, what replaces the old vehicle: another old car, a new car, or a different mode of transportation?  Tabl5e 4 Vehicle Emission Factors from AirCare Sample Tabl5e 4 Vehicle Emission Factors from AirCare Sample [Gram/Km] CO HC NOx Scrapped vehicles 2.6203 0.1808 0.4412 New vehicles 0.2487 0.0209 0.0089 Source: Authors' calculations. Pioneering studies by Alberini, Harrington, and McConnell (1995, 1996) provide the theoretical and empirical foundation for such research. They find that as the incentive level is raised, program participants increasingly offer more valuable vehicles: that whereas low incentives attract vehicles in the poorest condition, with short expected lifespans, higher incentives induce the scrapping of vehicles with longer remaining lives. Programs in the United States and Canada In the United States, the Car Allowance Rebate System (CARS), or ”Cash for Clunkers,” was a short-term program in July/August 2009 designed to stimulate the automotive industry. Vehicles scrapped had to be less than 25 years old and to have a combined (45/55 percent highway/city) fuel economy of 18 miles per gallon or less (or fuel consumption of 13.1 litres/100km or more). Vehicles bought had to have a combined fuel economy of at least 22 miles per gallon (10.7 litres/100km or less). Eligible participants received between$3,500 and $4,000, and with a budget of$3 billion, over 690,000 vehicles were retired. Sivak and Schoettle (2009) report that the program had a small positive effect on fuel efficiency of new vehicles (about 0.6–0.7 miles per gallon). Some US states offer their own programs. For example, California’s Bureau of Automotive Repair administers an AVR program providing $1,000 per vehicle ($1,500 for low-income individuals) for any licenced and registered vehicle passing a visual inspection for roadworthiness.

In Canada, a nationwide program called Retire Your Ride (RYR)—introduced in 2007—provided a modest $300 cash incentive to scrap vehicles of model-year 1995 and earlier. Designed to reduce vehicle emissions, the program was delivered by Summerhill Impact, a not-for-profit organization.9 It ended on 31 March 2011. The most innovative Canadian AVR program operates in British Columbia. In effect since 1996, the BC SCRAP-IT program offers a uniquely wide range of post-retirement options. Unlike traditional programs that provide cash or subsidize the purchase of a new vehicle, the BC SCRAP-IT program encourages participants to choose alternative forms of transportation. Participants can receive subsidies towards public transit use, membership in ride-share or car-share programs, or the purchase of a bicycle. In addition, incentives for choosing transit are higher than for vehicle replacements and cash.10 Estimating Emission Savings In Antweiler and Gulati (2012), we study the emission savings from the BC SCRAP-IT program. Three individual components of our estimate are discussed below. First, how many years will a car survive if it is not scrapped? To estimate the remaining lifetime of vehicles, we use detailed vehicle fleet inventory data to construct survival probabilities using different parametric functions.11 We estimate a residual vehicle lifetime of approximately nine years. This is considerably higher than other studies (for example, 3.4 years in Sandler 2012). British Columbia’s unique conditions might explain this: the climate is milder than the rest of Canada and much of the United States where vehicles suffer greater wear and tear, and vehicles are driven significantly less. In addition, programs in other jurisdictions have lower incentives, and thus attract vehicles with shorter expected lifetimes. Our estimate’s second component concerns how many kilometres the retired vehicle has been driven, and what the intensity of use is for its replacement. Unlike previous studies, estimating vehicular emissions and kilometrage using overall fleet data (Hahn 1995; Li, Linn, and Spiller 2010; Knittel 2009; Sandler 2012), we have access to end-of-life kilometrage and emissions for individual vehicles retired through British Columbia’s AirCare program. Past a certain age, all vehicles in the BC Lower Mainland are tested annually or biennially for vehicle emissions through AirCare. We observe the average vehicle kilometres travelled between the two last mandatory inspections. This information is projected onto the replacement vehicle, assuming that it is driven at the same annual average as the scrapped vehicle. Third, what is the per kilometre difference in emissions between the scrapped vehicle and its replacement? This replacement could be a car or alternative transportation. For replacement vehicles, the difference in emissions derives from the last mandatory emission inspection of the retired vehicle, and an average for new vehicles from Table 4. For those switching to public transit, the process is a little more complicated. Additional emissions from higher utilization rates of public transit are negligible. However, not everyone opting for public transit continues using it. A transit incentive may also be combined with a vehicle purchased outside the program. We assume that approximately half of previous vehicle kilometres are replaced by public transit, if chosen in the program. This assumption derives from a survey of BC SCRAP-IT participants conducted in 2007, where 58 percent and 32 percent of those opting for transit indicate they are either ”very likely” or ”somewhat likely” to continue using transit once program-funded passes run out. Fuel efficiency gains between new and old vehicles are relatively small.12 Most technological gains are eroded by additional weight and power in newer vehicles. This implies that CO2 savings from AVR programs are not substantial. In contrast, as shown in Table 4, savings in CO, HC, and NOx are significant. The results of our analysis are reported in Table 5 for the Lower Mainland of British Columbia—the region for which we have the most comprehensive data. Accelerated vehicle retirement generates, on average, 11.2 tonnes reduction in CO2 over the remaining vehicle lifetime. Much of these savings come from transportation mode switching. CO emissions drop by 466 kilograms (kg), HC by 32 kg, and NOx by 81 kg. Our results for CO and HC are similar to Sandler (2012), but are much higher for NOx. It is possible to assign economic values to emission reductions using estimated valuations of the various pollutants.13 We estimate that the benefits of the program range from about$750 for replacing an old vehicle with a new vehicle to about $1,100 for participants switching to public transit use. Alternative Incentives The BC Scrap-It program is unique in its variety of options for alternative transportation. While purchasing a replacement vehicle was the most popular option (56 percent) during the program’s most popular phase phase (2008–2009), public transit and bicycle-transit combo packages accounted for the second largest share (37 percent). In Antweiler and Gulati (2013), we analyze the decision to participate in the program by linking the participants’ postal codes to census data and thus to a variety of socio-demographic characteristics. We find that urban vehicle owners are much more  Table 5 Calculation of Program Benefits (Lower Mainland of British Columbia Only) Table 5 Calculation of Program Benefits (Lower Mainland of British Columbia Only) Average New Vehicle Public Transit Pollutant Price Reduction Value Reduction Value Reduction Value Carbon Dioxide 30$/mt 11.2 mt $336 5.80 mt$174 14.9 mt $447 Carbon Monoxide 0.50$/kg 466. kg $233 418. kg$209 499. kg $250 Hydrocarbons 3.50$/kg 31.9 kg $112 28.3 kg$99 34.3 kg $120 Nitrogen Oxides 3.50$/kg 81.0 kg $284 76.0 kg$266 84.5 kg $296 Total$965 $748$1,113

Notes: kg=kilogram(s). mt=metric tonne(s).

Source: Authors' calculations.

likely to participate in the program than their rural counterparts. The urban-rural divide also determines the choice among the different incentive options. Population density is the strongest predictor of adopting public transit, as public transit networks are denser and service is more frequent in densely populated urban areas. We also find that transit is favoured more strongly in areas with a younger population as well as in areas with a disproportionate share of high-income households. Public transit is also favoured by owners of younger (higher-value) vehicles, which is consistent with the link between incentive level and participation rates identified in Alberini, Harrington, and McConnell (1996).

Discussion

Accelerated vehicle retirement programs can make an effective contribution to reducing emissions. The BC SCRAP-IT program appears to be cost-effective due to a combination of favourable conditions in British Columbia (higher remaining vehicle lifetime) and the program’s focus on promoting public transit use. The program suggests important lessons for similar programs elsewhere.

First, accelerated vehicle retirement is most effective when there are large discrete changes in vehicle technology, such as the introduction of catalytic converters or electronic engine control systems. Targeting vehicle vintages on the far side of such discrete changes generates greater environmental benefits than indiscriminate programs. This can create substantial reductions in local air pollutants such as CO, HC, and NOx. However, as the gap in fuel efficiency between new and old cars is small, CO2 savings are not significant.

Second, such programs are better focused on urban areas. Because of lower average speeds and traffic congestion, emissions per kilometre driven tend to be higher in urban areas than in rural areas. Ambient pollution concentrations are also higher in urban areas, and their health impacts are aggravated by the non-linear dose-response function of major air pollutants. Consequently, limiting AVR programs to high-impact urban areas generates greater environmental and health benefits.

The innovative scrappage program in British Columbia pioneers a third and novel dimension. Rather than merely incentivizing switching from old to new vehicles, the BC SCRAP-IT program encourages switching to public transit and car-share programs. Both public transit and car-share programs are more readily available in urban areas, and bicycling distances are shorter. This strengthens the case for targeting urban areas. More importantly, a large share of the environmental gains in the BC SCRAP-IT program comes from incentivizing the switch to public transit (and to a lesser extent, to bicycling and car-sharing). This implies a large differential between incentive levels for new cars and alternative transportation modes. Given the uniquely beneficial aspects of promoting public transit use in urban areas, the most cost-effective use of accelerated vehicle retirement programs is likely a strong—or even exclusive—focus on alternative transportation modes.14

Despite the great attention garnered by hybrid and electric vehicle incentive programs, they are not cost-effective. There are two reasons. First, the programs induce only a tiny share of vehicle owners to buy targeted vehicles. Most of the subsidies are paid to those who would buy these vehicles anyway. Second, fuel savings from those induced to switch are small. Most consumers buying HEVs and EVs would buy fuel-efficient vehicles even in the absence of the subsidy. To make these programs cost-effective, it would be necessary to induce much larger participation and to persuade the owners of gas guzzlers to switch to HEVs and EVs.

There is a better alternative that may be politically viable. Recognized as the best alternative to a fuel tax, feebates combine a fee for vehicles with fuel consumption per kilometre above a specified value (the pivot point) and a rebate for vehicles with fuel consumption below the same value (Greene et al. 2005; Fischer 2008). The province of Ontario employed North America’s first automobile feebate in 1991. The VEI, a federal feebate, was briefly implemented from 2007–2008. Both programs were somewhat ineffective due to their small scope. However, if the fee and rebate are proportional to fuel consumption, and the policy is comprehensive, it will approximate a fuel tax on vehicle purchases. Moreover, feebates can be implemented at consumer, producer, or retailer level to optimize administrative simplicity. They also induce the equality of marginal compliance costs for vehicle fuel economy across producers, and that makes them more efficient than conventional fuel economy standards. Lastly, feebates can be administered in a revenue-neutral way if desired. This policy will be more cost-effective than subsidies and can improve on the benefits of recently implemented fuel efficiency standards.

Voluntary AVR programs are used to reduce emissions (and occasionally to stimulate vehicle demand during recessions). Due to discrete improvements in vehicle emission-control technologies, newer vehicles are significantly cleaner than their older (pre-1995) counterparts. This is particularly true for most local air pollutants. While recently there is much greater focus on carbon dioxide as a contributor to climate change, the gap in fuel efficiency (and therefore CO2 intensity) between new and old cars is small. Gains in fuel efficiency from improved engine technology are offset by increases in vehicle mass and power.

AVR programs are more cost-effective if focused on urban areas. Firstly, they generate greater environmental (health) benefits in areas with higher ambient pollution concentrations due to higher vehicle density. The future success of AVR programs thus depends critically on their ability to switch participants out of vehicle ownership completely. Secondly, as most vehicles on the road employ modern emission-control technologies, replacing already low-emission vehicles with newer vehicles in the future is unlikely to generate large environmental benefits. This implies switching participants to a new mode of transportation: bicycling, car-sharing, and most importantly, public transit. This is also more likely in urban areas. For the BC SCRAP-IT program, a sizable portion of its environmental gains derive from encouraging participants to switch to public transit. The dense and frequent public transit network in the Vancouver region contributes significantly to the success of the BC SCRAP-IT program. Future AVR programs should also focus on areas where they can piggyback on existing transit networks to reduce individual car ownership.

Notes

1 The ideal policy is to price each pollutant at its marginal damage to society—often an impractical solution. Consider a driver generating CO that reduces the respiratory capacity of an infant residing close to her route. Ideally, the driver should pay for the damage caused to the infant’s health. This would involve accurately measuring emissions and exposure to individual pollutants, establishing cause and effect on health outcomes, and finally valuing the outcome. While not ideal, a fuel tax can closely approximate this relationship because emissions are linked closely to the combustion of fuel.

2 A uniform fuel tax is appropriate for CO2, as the marginal damage is global in scope. Due to regional variation of ambient concentrations of local air pollutants (as well as non-linear dose-response), a fuel tax should be higher in densely-populated than in sparsely-populated areas.

3 The share of hybrid vehicles in total light vehicle sales in Canada for 2005 was 0.35 percent.

4 Overall sales of HEVs in Canada for the year 2006 were only 8,924, dwarfed by sales in the United States, where 263,271 HEVs were sold in the same year.

5 The Union of Concerned Scientists evaluates emissions from electric vehicles across the United States. Anair and Mahmassani (2012) find that nearly half of Americans live in regions where charging an EV on the grid emits less CO2 than driving hybrid vehicles. Regions with a high share of coal-fired power plants perform worst. The results in Samaras and Meisterling (2008) are less favourable. Comparing life-cycle GHG emissions, they find that plug-in hybrids or EVs reduce GHG emissions by 32 percent compared to conventional vehicles, but have small reductions compared to traditional hybrids. Both studies only focus on CO2 and do not include other types of pollutants.

6 We adopted this figure as a result of comments received from our referee, whom we wish to thank and acknowledge for contributions to this paper.

7 Note that a feebate cannot increase the marginal cost of driving per kilometre. That is still determined by the price of fuel.

8 As only the greenLEVY remains of the VEI, the program is generating tax revenue. At the time of its implementation, the Ontario Tax on Fuel Conservation was projected to raise $45 Million (Bregha and Moffet 1995). 9 Founded in 2000 through a partnership of governments, industry, and non-governmental organizations, the organization was originally known as the Clean Air Foundation, and subsequently morphed into Summerhill Impact. 10 Participants get$200 in cash or $300–$1,000 towards purchasing a new vehicle—the amount determined by the difference in fuel economy of the new and retired vehicle. The value of the transit alternative is approximately $1,350. 11 A major difficulty is accounting for a possible selection bias for program participation: scrapped vehicles are likely to be of poorer quality than vehicles of similar make and age. However, due to high incentives offered previously in this program, we do not expect this to be much of an issue. 12 Fuel efficiency data for new and old cars is either readily available (e.g., from the US EPA), or is estimated by information on engine displacement, fuel, and transmission type. 13 We use the BC carbon tax of$30/tonne as a lower bound for the negative externality of carbon dioxide, although the meta study by Yohe (2007), makes a strong case for $50/tonne. For NOx,$3.50/kg is the value of NOx given by Krupnick et al. (2005). HC and CO match the NOx price in 1:1 and 1:7 ratios respectively.

14 In addition, if technological improvements in emissions-control technology have plateaued, this may be the only sensible focus for the future.