4.3. Control of Emissions of Air Pollution from Highway Heavy-Duty Engines
The Final Regulatory Impact Analysis: Control of Emissions of Air Pollution from Highway Heavy-Duty Engines deals with the problem of rising use of highway diesel engines tending to offset the gains resulting from regulation:
Over the last 20 years, emissions from highway heavy-duty engines have been reduced as a result of changing emission standards and related requirements. Although previously promulgated standards for control of emissions of oxides of nitrogen (NOx) and hydrocarbons (HC) are expected to lead in the short term to reductions in emissions of these ozone precursors, there is concern that the fleets’ emission levels will increase in the future. … factors such as the growth in the number of vehicles and driving activity in the near future will likely result in total NOx emissions that will exceed current levels. (p. 1)
EPA’s remedy for this problem was to make a new rule and to sign a Statement of Principles (SOP) with the California Air Resources Board and the manufacturers of highway heavy-duty engines in 1995. The 2004 limits are more stringent than the 4 g/bhp-hr NOx standard that takes effect in 1998. The emissions limits beginning in 2004 are as follows:
Combined NMHC [non-methane hydrocarbons] plus NOx not to exceed 2.4 g/bhp-hr [grams per brake horsepower-hour],
Combined NMHC plus NOx not to exceed 2.5g/bhp-hr, with a cap of 0.5g/bhp-hr on NMHC emissions. (p. 1)
The SOP also provided for the development of averaging, banking and trading of emissions credits but regrettably these are not discussed in the document. The agreement clearly contemplated pushing the technology; the means to achieve the requisite emissions level were not at hand in 1995 but the document is optimistic that research in specific areas will succeed. The RIA contains extensive discussion of technologies that might be refined to meet the requirements but has no discussion of alternative policies.
The RIA identifies three primary technologies that EPA expects will be refined to meet the 2004 requirements: exhaust gas recirculation, combustion optimization and fuel system upgrades. Ibid., pp. 88-90. Cost estimates are made on a per vehicle basis for each of the three technologies for four vehicle types. Costs are further subdivided into capital and operating (production) costs incurred by the engine manufacturers – fixed and variable costs in the language of the paper – and incremental operating costs incurred by end users.
Capital costs are incurred for research and development, tooling and certification. To account for the fact that these costs are incurred before production, notional interest is added at the rate of 7% annually “and [costs] are then recovered with a five-year amortization at the same rate.” Ibid., p. 84. Ultimately, a present value calculation is made so this exercise this exercise in arbitrary parameters could be avoided with some reworking of the approach. The choice of 7% and five years affects the present value calculation, however. The exercise does permit costs to be stated year by year but the result is somewhat arbitrary.
With regard to fuel system upgrade technology, a substantial diminution in projected actual costs is made:
EPA believes it is not appropriate to assign the full cost of fuel system upgrades to the proposed emission standards. Much of the anticipated improvements will come independently of the 2004 model year standards and any remaining system improvements for 2004 and later model year vehicles will provide benefits beyond lower NOx emissions. In an effort to properly assess the cost of the fuel system upgrades, EPA estimates for light heavy-duty vehicles that 25 percent of the fuel system cost increase should be allocated to the proposed standards. For all other engines, half the fuel system cost increase is attributed to the proposed standards. (p. 90)
It is unclear where the 25% and 50% figures come from and no sensitivity analysis is performed on them. Incremental end user operating costs are incurred particularly with respect to exhaust gas recirculation. There may be some fuel penalty and additional costs for motor oil and engine rebuilding. Production costs are programmed in the document to decline:
The second modification is related to the effects of the manufacturing learning curve. This is a well documented and accepted phenomenon dating back to the 1930s. The general concept is that unit costs decrease as cumulative production increases. (p. 94)
In implementation, there are two cost reductions of 20% each. Some motivation is provided for the choice of 20% and sensitivity analyses are performed for greater and lesser values. Ibid., p. 96. Samples of the results follow. For light heavy-duty vehicles in 2004, the regulation raises purchase price by $258 and life-cycle operating costs by $7 whereas for heavy, heavy-duty vehicles in 2006, the purchase price is raised by $411 and operating costs by $131. Ibid., Table5-4, p. 98. Total annual costs are $242 million in 2004, $123 million in 2009 and $180 million in 2020. Ibid., Table 5-7, p. 100. Note that these figures are affected by the choice of how to amortize capital costs which explains why total cost plunges in 2009.
The benefits analysis in the document begins with a simplifying assumption:
While the standards are combined NMHC plus NOx … it was necessary to consider the NMHC and NOx emission impacts separately. … it is reasonable to model the … new standards as being equivalent to a 2.0 g/bhp-hr NOx standard and a 0.4 g/bhp-hr NMHC standard. (p. 103)
The bulk of the benefits seem to be concentrated in the defacto NOx standard. The document estimates a NOx decrement of 106 thousand short tons in 2005 increasing to 1,066 short tons in 2020 attributable to the regulation. Ibid., Table 6-6, p. 107. The reductions are with respect to a base case of no further controls beyond the 1998 limits.
The per-vehicle average lifetime emission reduction of NOx is 540 pounds for light heavy-duty vehicles, 2,350 pounds for medium heavy-duty vehicles and 7,040 pounds for heavy heavy-duty vehicles.
The RIA summarizes the environmental impacts of the regulation as follows:
NOx reductions are projected to exceed 1.2 million tons per year in 2020, which would be a five percent reduction in the total NOx inventory. NMHC reductions are projected to be much smaller, about 25,000 tons per year in 2020, which would be much less than one percent of the national NMHC (or VOC) inventory. These emission reductions are expected to contribute very significantly towards reducing and controlling ambient ozone levels in the future, counteracting the expected effects of new sources and growth in vehicle miles traveled. The new controls would also result in benefits with respect to nitrate particulates, visibility, acid deposition, and estuarine eutrophication. (p. 121)
With year-by-year estimates of costs and emissions benefits already calculated, the document proceeds to make cost-effectiveness computations. Discounting of costs and benefits is performed using a 3% interest rate. Ibid., p. 125. Although not stated, the modest 3% rate is probably a real rate because there are no inflation factors applied to costs over time. Note that this is a different rate than the 7% used to defer and amortize capital costs. No explanation of the difference or either rate is provided. The effect of using different rates is not analyzed.
There are two cost-effectiveness scenarios. Ibid., p. 125. In the first, the divisor is the total emissions decrement for NOx and NMHC. In the second, the divisor is the emissions decrement in attainment problem areas: the states that border on the Mississippi River, all of the states east of the Mississippi River, Texas, California, and any remaining ozone nonattainment areas west of the Mississippi. That is equivalent to reducing the divisor by 13%.
Cost-effectiveness results are presented for (1) three vintages of vehicles corresponding to when costs are programmed to change because of the learning curve effect and the expiration of capital cost amortization, (2) two scenarios and (3) three vehicle classes. The results range from a low of $100/ton for heavy heavy-duty vehicles beginning in 2009 for either scenario to a high of $1200/ton in 2004-2005 in the regional scenario. Ibid., pp. 127-128. However, these results are sensitive to the choice made as to how to amortize capital costs.
Additionally, cost-effectiveness numbers are provided for an aggregate of the first 30 affected model years.
These cost-effectiveness numbers are calculated by weighting the various model year per-vehicle cost-effectiveness results by the fraction of the total 30 model year sales they represent. The sales … were assumed to grow at a linear rate of two percent from the 1995 levels. (p. 126)
Here, there are three vehicle classes and two scenarios, the vintages having been aggregated. The results range from $100/ton for heavy heavy-duty vehicles in either scenario to $700/ton for light heavy-duty vehicles in the regional scenario. Ibid., pp. 127-128.
There is no original effort to monetize benefits; most of the brief economic analysis is cost-effectiveness. However, the results of a study of studies of monetized benefits of NOx reduction are listed in the document. Ibid., p. 130. These benefits, including human mortality and morbidity, are described as secondary because the primary benefit is construed as reduced ozone production. The secondary benefits total $967/metric ton. The computed cost-effectiveness numbers seem to be on the order of the secondary benefits alone suggesting that overall benefits exceed costs. The paper, however, says only that the benefits numbers
… are presented in this chapter for informational purposes only. (p. 130.)
Finally, for purposes of comparison, the cost-effectiveness numbers for two other NOx reduction programs are presented. They are $1300-1500 per ton for the Clean Fuel Fleet Vehicle Program and $5000 per ton for Reformulated Gasoline – Phase II. Ibid., p. 134.