TECHNOLOGY Secondary processes key to gauging U.S. refining capability

Tancred Lidderdale, Nancy Masterson
Energy Information Administration
Washington, D.C.

Nicholas Dazzo Morgan
Guaranty Trust Co.
New York City

Utilization rates for crude oil distillation no longer correlate well with refining gross margins.

This relationship suggests that an alternative way to measure refining capacity and utilization is needed to properly assess the effect of growing demand on market prices and sources of supply.

A focus on the capacity and utilization of downstream processes may provide a better indicator of potential constraints in refining capacity. A more complex analysis could expand the scope to include foreign refiners in a global view of petroleum markets.

Utilization

U.S. crude oil refinery utilization rates have increased steadily since oil price and allocation decontrol in 1981.

The annual average atmospheric distillation utilization rate has increased from 68.6% in 1981 to 92.6% in 1994. The distillation utilization rate reached a peak of 96.4% in August 1994, the highest 1-month average rate in over 20 years.

This dramatic increase in refining capacity utilization has stimulated a growing interest in the ability of U.S. refineries to supply domestic requirements for finished petroleum products.

This article briefly reviews recent trends in domestic refining capacity utilization and examines in detail the differences in reported crude oil distillation capacities and utilization rates among different classes of refineries.

Historical capacity

U.S. crude oil distillation capacity grew steadily under the crude oil price and allocation programs set up in 1973 and 1974. The price and allocation programs benefited small refineries, resulting in growth in the number of refineries from 268 in 1973 to 324 in 1981.1

During that period, average annual operable distillation capacity increased from 13.6 million b/d to its all-time high of 18.6 million b/d. With the removal of price controls and allocations in early 1981 and the decline in petroleum demand in the early 1980s, many small refineries that sprang up in the 1970s, as well as older, inefficient plants, began to shut down.

Between January 1981 and January 1989, the U.S. refining industry experienced a net loss of 120 refineries and approximately 3 million b/d of operable capacity. Thereafter, capacity remained stable until 1992 when, pressured by the costs of environmental regulations and unfavorable refining economics, several marginal refineries shut down, while a few larger refiners idled or permanently shut down crude oil distillation units.

By Jan. 1, 1995, U.S. operable refineries numbered 175, with a total crude oil distillation capacity of 15.4 million b/cd.

Crude runs

While distillation capacity has declined during the last 15 years, refinery gross inputs have increased almost every year since 1983. Average gross inputs increased from 11.9 million b/d in 1983 to 14.0 million b/d in 1994.

This increase was the result of a steady rise in U.S. demand for petroleum products beginning in 1983. U.S. demand has exceeded U.S. crude distillation capacity every year since 1985.

Total petroleum product demand averaged 15.7 million b/d in 1985-0.4% higher than the average operable crude oil distillation capacity. By 1994, total petroleum demand exceeded operable capacity by 17%, as demand grew to 17.7 million b/d and operable capacity averaged 15.2 million b/cd.

With declining capacity and rising gross inputs, the amount of unused crude oil distillation capacity has shrunk significantly since 1981. Approximately 5.9 million b/d of available crude oil distillation capacity was unused in 1981, but by 1994 the difference between available capacity and gross inputs had narrowed to about 1.1 million b/d. (See related story, OGJ, Dec. 11, 1995, p. 29).

As a consequence, operable utilization rates for crude oil distillation units have increased fairly steadily since 1981.

Utilization rates increased nearly every year from 1981 to 1994, growing from an average of 68.6% in 1981 to 92.6% in 1994, the highest annual level since 1973. The refinery utilization rate reached a peak of 96.4% in August 1994-the highest 1-month average rate in more than 20 years. Since then, monthly utilization rates have remained between 87.7% (February 1995) and 95.6% (June 1995).

Downstream units

Since 1980, the refining industry's emphasis has shifted from increasing crude oil distillation capacity to investment in downstream (secondary) processing units, thereby increasing overall refinery complexity. This transition began several years before the passage of the Clean Air Act Amendments in 1990 as a result of increased demand for lighter, cleaner products that have to be produced from increasingly heavier and more sour crude oils.

In contrast to utilization rates for crude oil distillation, utilization rates for secondary processing units increased only slightly, or even declined, between 1987 and 1994 (Table 1 [55146 bytes]). The average calendar-day utilization rate for catalytic cracking increased every year through 1993 before decreasing slightly to 92.2% in 1994.

Cokers have had the highest utilization rate of the secondary processing units. In 1994, the average annual calendar-day utilization rate for cokers was 93.0% compared to 92.3% in 1987.

Hydrocracking utilization rates consistently have been less than those of catalytic crackers and cokers. Average annual capacity for hydro cracking increased 14.6% between 1987 and 1994, while average annual inputs to hydrocrackers grew about 6.3% during these years.

Capacity calculations

Atmospheric crude distillation utilization rates reported by the Energy Information Administration (EIA) are calculated on the basis of calendar-day capacities.

Some refining industry analysts have made the mistake of assuming that utilization rates reported by EIA are based on stream-day capacities. This assumption leads to the belief that utilization rates of 94-95% probably are the best the industry can do on a sustained basis.

Because EIA's calculations are based on calendar-day capacities, utilization rates on the order of 100% are the best the industry can do on a sustained basis.

Operating factor

While stream-day capacity should be an objective design measure of a refinery's processing capability when operating, calendar-day capacity is a subjective measure that depends on an assumed operating factor.

Operating factors for U.S. atmospheric distillation units were calculated from the calendar-day and stream-day capacities reported by EIA in its Petroleum Supply Annual (Table 2 [27506 bytes]).

Most gasoline-producing refineries assume operating factors for atmospheric distillation of 94-98%. Calculated operating factors for refineries that do not produce gasoline (those that produce only asphalt, lube oil, waxes, and other heavy petroleum products) generally are less than 94%.

The lower operating factors assumed for these refineries may result from the different quality of crude they process and a higher frequency of turnarounds.

It follows, therefore, that refineries that produce only heavy products operate at significantly lower atmospheric distillation utilization rates than gasoline-producing refineries (Table 3 [24389 bytes]). During 1994, gross inputs to atmospheric distillation units at refineries not producing gasoline averaged 52% of calendar-day capacity, compared to 94% at gasoline-producing refineries.

Because of the low utilization rates at non-gasoline-producing refineries, these refineries are excluded from the analyses that follow in this report.

Complexity

The significant difference in capacity utilization rates between the different refinery configurations suggests differences in the economics of running the different types of refineries. The cost penalty for operating a non-gasoline-producing refinery at a low capacity utilization rate typically is much smaller than the cost of operating a complex refinery that produces gasoline.

Further disaggregation of complex refineries by type of secondary processing units (for example, so-called "coking" and "cracking" refineries) reveals no significant differences in utilization rates. (These results are not reported here.)

Utilization trends

The average industry utilization rate for distillation is, of course, highly seasonal, with peak rates occurring during the high-demand summer motor-gasoline season.

The maximum monthly average utilization rate in 1994 occurred in August (Table 4 [28310 bytes]). Only 15 of 129 individual gasoline-producing refineries, however, reported maximum 1994 utilization rates in August. Less than two thirds of U.S. refineries reported maximum 1994 operating rates during the summer months (between April and September).

The minimum industry atmospheric distillation operating rate in 1994 was recorded in March. Again, only about 15% of domestic refineries operated at minimum rates during March. About two thirds of the domestic refineries reported minimum 1994 operating rates during the winter months (between October and March).

Because refineries are less likely to take turnarounds during the high-demand summer gasoline season, it is not surprising that many refineries operate at rates in excess of reported calendar-day capacity during this period.

More than two thirds of the gasoline-producing refineries reported gross inputs to atmospheric distillation that exceeded their reported calendar-day capacity (Table 5 [19651 bytes]). What is surprising is that almost 40% of the domestic refineries reported a 1-month distillation utilization rate that exceeded their reported stream-day capacities.

The observation that almost 40% of the domestic refineries reported a 1-month utilization rate in excess of 100% of reported stream-day capacity implies that U.S. distillation capacity may be understated.

A "pseudocapacity" number could be derived by using observed processing rates for refineries that operated over 100% of stream-day capacity. The reported capacity of refineries operating at maximum rates less than 100% of stream-day capacity would not be changed.

The pseudocapacity of domestic atmospheric distillation is estimated to be 236,000 b/d more than the reported stream-day capacity (Table 6 [22064 bytes]).

Profitability measure

Gross margins for complex refineries have been relatively flat over the last several years, even as domestic distillation utilization rates have increased. Typically, higher capacity utilization would be expected to be associated with higher margins. (In the conventional economic short-run profit-maximization model of competitive markets, the marginal cost of production and the gross margin increase with capacity utilization.)

To explain this inconsistency, one must look beyond the standard measure of capacity utilization (crude distillation) toward a measure of refinery activity that more closely reflects the industry's economics.

Conventional measures of capacity utilization are based on crude distillation capacity because it is still by far the largest component of the typical refinery's capacity.

Even in the U.S., where refiners have added mostly secondary processing capacity, atmospheric distillation is still equivalent to almost 300% of catalytic cracking capacity. By contrast, in Europe and the Far East, atmospheric distillation capacity is more than 700% of catalytic cracking capacity.3 For this reason, inputs to distillation units provide a useful indication of refinery output despite the advance of refinery complexity.

But crude distillation is insufficient for analyzing refinery profitability.

Complexity has increased to meet growing demand for refined products. In addition, U.S. refineries have had to comply with governmental regulation of fuel quality. These regulations have further encouraged growth in secondary processing capacity, which, in recent years, has exceeded demand growth.

As a result of these trends, distillation capacity utilization has increased while utilization of secondary processing capacity has remained relatively flat.

Since 1990, refining gross margins have generally declined (OGJ, Dec. 11, 1995, p. 29). When this trend is compared to trends in distillation capacity utilization vs. secondary processing capacity utilization, it appears that secondary processing capacity may be a more appropriate tool for analyzing refinery profitability.

While distillation capacity utilization is a poor predictor of refinery profitability, it should be noted that total secondary processing capacity utilization is not much better.

Monthly average gross margins for U.S. Gulf Coast and West Coast refiners are poorly correlated with both of these measures of utilization in Petroleum Administration for Defense Districts (PADDs) 3 and 5 (respectively, the Gulf and West Coasts). One important explanation for this is the importance of foreign trade.

Because the U.S. is just a part-albeit a large one-of the broader Atlantic Basin market for petroleum products, domestic refiners are subject to supply/demand balances outside U.S. borders.

Western European refiners have raised their catalytic cracking capacity sharply during the past 2 years (between Jan. 1, 1993, and Jan. 1, 1995).4 This increase in European conversion capacity implies that refiners expected European motor gasoline demand to grow at a rate close to the 2.7%/year average seen during the previous 5 years.3

This forecast has proven overly optimistic, and a resultant increase in exports (or even just the availability of exports) to the U.S. has been a depressing factor on U.S. Gulf Coast refinery profitability.5

Cracking margins

This article has shown that measures of refinery capacity utilization have several shortcomings, and that 100% stream-day capacity for atmospheric distillation units is not a clear limit of U.S. refinery throughput. Moreover, 100% stream-day capacity does not represent an unsurpassable limit, but rather a slope of increasing steepness.

Capacity constraints relate to cost functions. (How fast do the marginal costs of production-either operating costs or product-quality degradation-increase as operating rates increase?) Thus, rather than focus on reported utilization rates, it may be more useful to study how marginal costs increase as utilization rates rise. Marginal costs should dictate gross margins for crude oil refining and market prices of petroleum products.

The marginal costs of increasing throughputs in secondary processing units, such as cokers and catalytic crackers, is likely to be greater than in atmospheric distillation units. It is the high marginal cost of running secondary processing units that causes these units generally to represent the constraining capacity in complex refineries.

In the U.S., cracking margins should provide the first signal that a refining capacity constraint is being reached. The trend in cracking margins over the last 5 years does not indicate that an industry capacity constraint has been reached.

It should be noted that cracking margins also reflect crude and product markets outside the U.S. and are subject to changes in refining capacity in the broader global market. While catalytic cracking and hydrocracking capacities have been relatively fixed in the U.S. since 1990, these capacities have increased in the broader Atlantic Basin, and show signs of continued rapid growth among key U.S. trading partners in the Pacific.

References

1. Annual Energy Review 1994, Energy Information Administration, DOE/EIA-0384(94), Washington, D.C., July 1995, p. 155.

2. Maples, Robert E., Petroleum Refinery Economics, PennWell Books, Tulsa, 1993, pp. 332-35.

3. International Energy Annual 1993, Energy Information Administration, DOE/EIA-0219(93), Washington, D.C., May 1995, pp. 51, 54-55 (and earlier issues).

4. Williamson, Michelle, Worldwide Refining Survey, OGJ, Dec. 19, 1994, pp. 54-55 (and earlier issues).

5. Petroleum Supply Annual 1994, Energy Information Administration, DOE/EIA-0340(94)/1, Washington, D.C., May 1995, p. 17.

6. Energy Information Administration publications: Petroleum Supply Monthly, Petroleum Supply Annual, Monthly Refinery Report, Annual Refinery Report, 1981-1995.

The Authors