1. Development and Evaluation of Feebate Policies for California’s Effort to Limit Greenhouse Gas Emissions from Light-duty Vehicles David L. Greene Corporate Fellow, Oak Ridge National Laboratory Senior Fellow, Howard H. Baker, Jr. Center for Public Policy & Research USAEE Meetings October 10, 2011 Washington, DC

2. Feebates are a policy for encouraging car buyers to prefer more efficient, lower emission vehicles and manufacturers to design them. A fiscal policy combining graduated ▫FEEs on inefficient vehicles ▫ReBATEs on efficient vehicles. A “benchmark” defines who pays and who receives. (distribution) A “rate” determines the marginal costs and benefits. (efficient solution) Depending on the choice of benchmark, feebates can produce revenue, be revenue neutral or be a net subsidy to car purchases The dual of fuel economy standards?

3. Simplest feebate is linear in energy use or CO 2 emissions per mile. Benchmark is E 0. Rate is slope of the line. System is revenue neutral if benchmark is carefully chosen. E $ EoEo

4. Today, 16 countries have some form of CO 2 or fuel consumption (l/100km) tax on light-duty vehicles. New vehicle purchase or registration: ▫Austria, Canada, Finland, France, Ireland, Netherlands, Norway, Portugal, Spain, U.S.A. Annual or recurring registration fee: ▫Denmark, Germany, Ireland, Luxembourg, Portugal, Sweden, UK Source: N.A. Braathen, 2010. “Incentives for CO2 Emission Reductions in Current Motor Vehicle Taxes”, ENV/EPOC/WPNEP/T(2009)2/FINAL, Environment Directorate, OECD, Paris.

5. France’s Bonus/Malus had an immediate effect, lowering the average emissions of cars sold by 7 g/mi., entirely due to car buyers choosing lower emission vehicles.

6. In the ARB Feebates model, the technological potential to increase fuel economy/reduce emissions was represented by cost curves. Manufacturers redesign approximately one fifth of their models each year. 6

7. The NMNL vehicle choice model estimated market shares of 20 vehicle classes and more than 800 individual vehicle configurations, as well as to buy a new vehicle or not. 7

8. Most feebate systems analyzed were benchmarked to the CAFE footprint function. 8

9. 56 cases were analyzed, reflecting a comprehensive set of feebate designs in various contexts. 13 cases analyzed differences in ▫Feebate rate: $10/$20/$30 per gram per mile ▫Geographical coverage: CA, CA + opt-in states, All of US ▫Benchmark: footprint, single, car or truck ▫Functional form: linear or step function 22 sensitivity cases considered the effects of ▫Fuel prices ▫Technology costs ▫How consumers value fuel savings ▫Other parameters 17 additional cases considered alternative post 2016 standards 4 more cases assessed whether feebates could replace the CA standard, plus the effects of banking emissions credits. Consumers assumed to require a simple 3-year payback. 9

10. The reference case included the fuel economy/emissions standards to 2016 and two alternative paths to 2025. 10

11. The impact of a CA feebate system increases almost linearly with the size of the feebate rate. 11

12. Considering the full value of fuel savings, the full costs per ton of CO 2 avoided are negative. Why? 12

13. The greater the market coverage of the feebate system, the greater its impact on emissions. 13

14. For a California feebate, most of the impact is due to sales-mix shifts, less to increased use of technology. 14

15. For a nationwide feebate system, most of the impact is due to increased use of fuel economy technologies. 15

16. The footprint benchmark has the smallest impact on consumers’ surplus, a single point benchmark the largest. 16

17. Does the public support fiscal policies or not? 17

18. Key findings are… With emissions standards in effect, feebates will further reduce LDV GHG emissions unless the standards are so strict as to induce pricing by OEMs. Reductions of 5 to 20 g/mi occur immediately, depending on the size of the feebate rate ($10 to $30). Reductions are obtained at negative cost, assuming car buyers typically undervalue future fuel savings relative to expected value. Impacts will diminish over time if standards are very strict and technological progress is slow. Feebates could have a very large or very little impact on hybrid vehicle sales, depending on future costs of hybrids. A national feebate would likely have 3 times the impact of a California-only feebate. A linear feebate system will likely be easiest to manage. A footprint feebate will have a smaller impact on consumer satisfaction but also a smaller impact on GHG emissions. 18

19. Thank you. Full report available at: www.arb.ca.gov/research/apr/past/08-312main.pdf

20. A feebate can be viewed as a tax on future oil use or GHG emissions paid at time of purchase. Assuming: ▫14,000 km/year when new = K 0 ▫Decreasing at 4%/year = δ ▫Discount rate of 7%/year = r ▫Expected life of 14 years = L ▫Cost to society of oil use and GHG emissions = C

21. The immediate shift in sales was large and tended to favor French auto manufacturers.

22. Norway’s registration tax was based 50% on weight, 30% on engine displacement, 20% on power. In 2007 the displacement component was replaced by a CO 2 tax, with an immediate impact on emissions and fuel efficiency.

23. Denmark’s system is based on km/l and is equivalent to $320 US per MPG. There are different rates for fees and rebates.

24. Denmark’s experience was similar to that of the other states, an immediate improvement in l/100km.

25. The U.S. Gas Guzzler Tax (for passenger cars only, not light trucks and still on the books) is half a feebate system. $1,800/0.01gal/mi = approx. $20/g/mi Gallons per Mile

26. The US gas-guzzler tax has also been effective. No mass market vehicles have ever paid it. Gas Guzzler Limit

27. The gas guzzler function was adjusted several times. Would be necessary with feebates to achieve revenue neutrality.

28. France’s Bonus/Malus is roughly equivalent to € 150 /tCO 2, a rate similar to the U.S. gas guzzler tax.