-4

Figure 1. Conceptual Formulation of Price Impacts of Regulations in the API Report

Figure 2 reproduces the empirical results developed in the API Report. The figure shows the compliance in costs for individual refineries estimated by the PRISM refinery model developed by Baker and O'Brien, Inc.' As noted above, the AP] report concludes that the suppliers would face different compliance costs due to ultra-low-sulfur requirements, with the compliance cost per gallon ranging from about 2 cpg to more than 15 cpg depending upon the refinery. Note, however, that most of the refineries have compliance costs that are within a reasonably small range, about 4 cpg to 8 cpg. Indeed, only about 10 percent of the capacity in the API study are projected to have a compliance cost greater than 8 cpg.

Bera

-5

Figure 2. Compliance Cost Estimates for Refiners Reported in API Report

Figure 2 presents projected compliance costs for individual refiners, rather than the total of production costs and compliance costs. A footnote to Figure 8.1 in the API report (shown as Figure 1 above) provides the API report's rationale for focusing on compliance costs rather than on overall costs.

Note that the PRISM cost curves shown in the previous section give the
incremental cost associated with the regulations, which are shown in the figure
as the difference between the two supply curves. (API report, footnote 9, p. 66)

(.. continued)

3 The figure also includes the API estimates of 2007 demand conditions and possible import response. The implications of the API assumption regarding imports are discussed below.

6 Note that the compliance costs calculated in the API report are the total additional costs of producing ultra-low sulfur fuel that complies with EPA's standards. See API report p. 67.

nera

- 6

This interpretation of the empirical results is critical to the price projections in the API report. As discussed below. this statement is not correct except under the most extreme assumption that the refiners with the highest full compliance costs of meeting the ultra-low sulfur requirement are also the refiners with the highest full incremental costs of producing diesel fuel. Note that these two components of cost are separate. The two components can be summed to give the overall cost of producing ultra-low-sulfur diesel.

The actual change in the supply curve due to environmental regulations could be very different than the results given in the API report. The API report implicitly assumes that there is a positive correlation between production costs and compliance costs, i.e., that the refinery with the highest full costs of compliance also has the highest incremental costs of production." No empirical information is given in the API report to support this implicit assumption.

Figure 3 illustrates two extreme scenarios regarding the relationship between production costs and compliance costs.

In Supply A, production costs and compliance costs are positively correlated; i.e., lowproduction cost refiners are assumed to have low compliance cost and high-production cost refiners are assumed to have high compliance costs. Under this assumption, the price increase is the largest possible. This corresponds to the implicit assumption in the API report.

In Supply B, production costs and compliance costs are negatively correlated, i.e., lowproduction cost refiners are assumed to comply at high cost and high-production cost refiners are assumed to comply at low cost. Under this assumption, the price increase due to the environmental regulations is relatively small.

These illustrations do not prove that the API report assumptions are wrong, although they do indicate that the API assumptions are at one extreme. Under a less extreme assumption, the price increase due to environmental regulations would be substantially lower.

7 Incremental costs of production refer to the full costs that are incurred assuming that refineries have an opportunity to adjust their plant and equipment between now and 2007 This approach is consistent with the approach in the API report, which uses "full cost curves [that] approximate industry long-run marginal cost curves". See API report, p. 67.

nera

-7

Figure 3. Compliance Costs Could Affect the Diesel Supply Curve in Different Ways Depending Upon the Relationship Between Compliance Costs and Production Costs

It is possible to provide some indication of the potential change in results under a less extreme assumption. As Figure 2 indicates, according to the PRISM results, compliance cost is above 8 cpg for less than 10 percent of the total capacity. If the highest 10 percent of capacity based upon compliance costs were not also the highest production cost facilities—and the original supply curve were not nearly horizontal the likely price increase would be no more

than 8 cpg.

III. THE API COMPLIANCE COST ESTIMATES IGNORE MANY OF THE POTENTIAL U.S. SOURCES OF ADDITIONAL DIESEL SUPPLY

The diesel supply curve should reflect the cost of all potential suppliers of on-road diesel fuel, including those that currently participate in the market and those that would participate or increase participation if the price were to increase. The API detailed cost analysis appears to focus on one category of potential suppliers of ultra-low-sulfur diesel fuel

nera

-8

refineries that currently produce "low-sulfur" diesel fuel (500 ppm) for the on-road market-to the exclusion of other domestic suppliers. This section discusses the implications of excluding other U.S. sources from the supply analysis. (The section after this discusses the API report's treatment of imports.)

A. Sources Included in the Domestic Diesel Supply Curve

Potential U.S. suppliers of diesel fuel oil that complies with the environmental regulations can be divided into four categories:

1. Refineries that currently produce on-road (low sulfur) diesel fuel and could modify their facilities to produce ultra-low-sulfur diesel fuel.

2. Refineries that currently produce off-road (high sulfur) diesel fuel-but not on-road (low-sulfur) diesel fuel-and could modify their facilities to produce ultra-low-sulfur diesel fuel.

3. Refineries that currently do not produce diesel fuel but could convert some existing gasoline capacity to ultra-low-sulfur diesel fuel.

4. New refineries that might be built to produce the new ultra-low-sulfur diesel fuel. The empirical analysis in the AP] report appears to be limited to refineries in the first category, ignoring the possibility of expansion, conversion from existing refineries and supply from new facilities. In the face of the large potential price increases projected by the API report, this restrictive assumption seems implausible. History suggests that new investment responds to regulatory requirements. The current sulfur standards for on-road diesel fuel appear to have prompted such a high level of investment in low-sulfur (500 ppm cap) diesel fuel in the early 1990s that refineries sometimes sell this fuel for off-road uses even though off-road diesel uses do not currently require it.'0

9

No new U.S. refineries have opened in the last two decades largely because of low profitability in the industry.

See API report, p. 33, which provides the basis for PRISM investment assumptions.