|
Electricity Shortage in California: Issues for Petroleum and Natural Gas Supply |
|
5. Petroleum Product Prices and Supply Disruptions
A. SummaryWe cannot predict the impact of rotating electricity outages on petroleum prices in California this summer because we have no historical experience of such outages and their affect on the petroleum supply system. However, any loss of production from California refineries could increase product prices. The size of the increase is dependent on numerous factors that include the severity and length of the electricity outage. |
Although the past is not considered a good predictor of California petroleum prices this summer, it is informative to look at that history. This history shows that relatively small disruptions in supply can cause significant increases in prices. Past disruptions to California refinery operations have resulted in price spikes ranging from 7 to 52 cents per gallon (Figure 5-1). The price spikes have varied considerably because of differences in the number, magnitude, and expected duration of the refinery disruptions and the condition of the market at that time.
Three important factors that affect the size and duration of price spikes are:
The most notable characteristic of the west coast (PADD 5) petroleum product markets is that PADD 5 is almost self-sufficient in the refinery supply of finished motor gasoline (Figure 5-2) and low-sulfur diesel fuel (Figure 5-3) compared with other areas of the country. Product imports and shipments from refineries on the U.S. Gulf coast satisfy only about 6 percent of finished gasoline demand (Figure 5-4) and about 3 percent of low-sulfur diesel fuel demand (Figure 5-5).
Figure 5-2. Finished Gasoline Balance by PADD, 5-Year Averages (1996 - 2000).
Figure 5-3. Low-Sulfur Diesel Balance by PADD, 5-Year Averages, 1996 - 2000.
Figure 5-4. Sources of Finished Gasoline Supply by PADD, 5-Year Averages (1996 - 2000).
Figure 5-5. Sources of Low-Sulfur Diesel Supply by PADD, 5-Year Averages (1996 - 2000).
Although PADD 5 appears to have a good supply-demand balance compared with the U.S. East coast (PADD 1), the recent history of wholesale product prices reveals the opposite. Wholesale gasoline prices on the West coast have been the most volatile in the country (Figure 5-6). While monthly average conventional gasoline prices in PADD 1 (New York harbor) have risen to at most 5 cents per gallon over the U.S. Gulf Coast (Houston) during the last 6 years, the West Coast (Los Angeles) has experienced several price spikes of as much as 50 cents per gallon over Gulf Coast prices.
There are two principal reasons for the greater price volatility in California. First, only 13 refineries in California supply the gasoline and diesel fuel markets in that State. The unexpected loss of supply from one of the larger refineries represents the loss of a significant share of supply. Second, the California Air Resources Board (CARB) product quality requirements for both reformulated gasoline and diesel fuel are more stringent than any other in the United States and all other countries. Because California does not routinely receive product supply from outside the region, refiners in the U.S. Gulf Coast and other countries that may be able to supply CARB quality products do not maintain inventories of these products and any possible response to meeting California supply shortfalls will be delayed. Consequently any unexpected disruption of supply from a California refinery results in short-term price increases.
To evaluate the price response to an unexpected loss of supply of product for California we identified past refinery disruptions and examined the size and duration of the resulting price increases. We examined daily spot prices for CARB reformulated gasoline (RFG) in Los Angeles and San Francisco from 1996 to date and identified 20 separate instances of price "spikes," that can be linked to refinery problems that were reported in the trade press. Prices in either or both Los Angeles and San Francisco rose by 7 to 52 cents per gallon over a period of 3 to 47 days (Figure 5-7 and Table 5-1). In some of the identified spikes, a single incident appears to have started or sharply accelerated upward price movement, while in others a combination of existing circumstances and/or successive events contributed to the price run-up.
| Table 5-1. California Gasoline Price Spikes | |||||||||
| Start | Peak | Cents/Gallon | Days from Start to Peak |
||||||
| 03/29/96 | 04/23/96 | 30 | 25 | ||||||
| 05/16/96 | 05/23/96 | 8 | 7 | ||||||
| 11/20/96 | 01/06/97 | 22 | 47 | ||||||
| 01/20/97 | 01/23/97 | 7 | 3 | ||||||
| 03/03/97 | 03/25/97 | 16 | 22 | ||||||
| 08/08/97 | 09/03/97 | 20 | 26 | ||||||
| 04/14/98 | 04/28/98 | 28 | 14 | ||||||
| 02/22/99 | 03/02/99 | 18 | 8 | ||||||
| 03/16/99 | 03/29/99 | 52 | 13 | ||||||
| 05/20/99 | 06/11/99 | 36 | 22 | ||||||
| 06/22/99 | 07/13/99 | 42 | 21 | ||||||
| 07/21/99 | 08/03/99 | 27 | 13 | ||||||
| 02/16/00 | 03/10/00 | 43 | 23 | ||||||
| 04/25/00 | 05/02/00 | 26 | 7 | ||||||
| 06/21/00 | 06/27/00 | 17 | 6 | ||||||
| 07/25/00 | 08/09/00 | 22 | 15 | ||||||
| 08/17/00 | 09/05/00 | 52 | 19 | ||||||
| 10/16/00 | 11/02/00 | 28 | 17 | ||||||
| 01/23/01 | 02/08/01 | 33 | 16 | ||||||
| 03/19/01 | 04/10/01 | 33 | 22 | ||||||
| Sources Refinery Disruptions: Oil Price Information Service, "Refinery Watch" (Rockville, MD., various issues). Prices: Reuters Ltd., reported in Energy Information Administration, Weekly PetroleumStatus Report, Table 13 (Washington, DC, various issues) (http://www.eia.doe.gov/oil_gas/petroleum/data_publications/weekly_petroleum_status_report/wpsr.html) | |||||||||
Each spike had its own unique set of ambient conditions, including time of year, inventory levels, crude oil price levels, prices for gasoline in other markets, etc. These pre-existing conditions appear to be a very important factor in price spikes, given that they can determine whether a given incident creates a short-term supply shortage or merely reduces an existing surplus.
It is important to recognize that a price spike actually reflects the perception of potential supply/demand impacts on the part of market participants, not necessarily consistent with the actual size or duration of the impact. If buyers or sellers believe that an event is likely to make product more scarce, and thus more expensive in the near future, they are likely to bid prices higher immediately, often making the run-up a sort of "self-fulfilling prophecy."
Market behavior in response to refinery incidents is not at all consistent. Depending upon previous conditions, price spikes can occur with no single discernible cause, and at other times, the size and/or duration of a spike can appear out of proportion to the observed drop in gasoline production (Figure 5-8).
The importance of pre-existing conditions in determining the magnitude of a price spike appears in the stronger correlation between California RFG stocks and spot prices than between production and prices (Figure 5-9). This presumably reflects the fact that the impact of one or more incidents on gasoline production, coupled with pre-existing conditions that may have resulted in rising or falling stock levels, are jointly reflected in the net stock changes in the period following the incident(s). Thus, if a given refinery incident occurs when stocks are otherwise rising, resulting in a less significant drop in inventories, it is likely to precipitate less price response than an incident with a similar production impact that occurs when stocks were already low and/or declining.
In general, retail gasoline prices in California exhibit a lagged and somewhat smoothed response to spot price movements. This lagged response pattern is observed in all other sections of the United States.[1]
As spot prices change at major buying centers in Los Angeles and San Francisco, prices further along the distribution chain will begin to reflect these changes. Changes in motor gasoline prices travel relatively quickly from the spot to retail markets. In a prior study, EIA has explored how prices at the retail end of the distribution chain reflect spot prices at the major buying center of Los Angeles. That analysis showed that 56 percent of a spot price change was passed through to retail within 4 weeks, with an additional 30 percent pass-through occurring in the next 4 weeks. That is, if the spot price rose 10 cents from one week to the next, retail customers would see about 6 cents of the increase 4 weeks later, and about 9 cents of the increase 8 weeks later.
Close examination of the price relationships in Figure 5-1 above reveals a number of important observations.
The time lag of the retail price response behind spot price movements is readily visible on price graphs such as Figure 5-1 that show the retail price rising slow and peaking later than spot prices. This lag effect, which means that the current retail price change is a moving average of prior spot price changes, creates a distortion that makes it seem as though the retail price is remaining at an elevated level longer than it should. Over a longer period of time, however, the average difference between spot and retail prices has not changed over the last 5 years.
California's problems are not isolated to that State. California refineries are important suppliers of product for the West Coast, and, as such, can impact prices in other PAD District 5 states. Figure 5-10 shows how West Coast prices tend to move more closely to each other than to the national average price. Thus, if the electricity outages in California this summer create any refinery problem that causes California gasoline prices to surge, the surrounding states could experience price increases as well.
If California prices are high enough relative to prices on the Gulf Coast, some Gulf Coast suppliers may ship gasoline components to the West Coast, which can increase price pressure in the Gulf Coast market as well.
EIA performed statistical tests to determine whether California spot CARB gasoline prices historically could be shown to consistently affect U.S. Gulf Coast RFG prices. The evaluation consisted of regressing weekly average U.S. Gulf Coast spot gasoline prices on the previous week's average Los Angeles spot prices and the previous week's U.S. Gulf Coast spot gasoline prices (Granger causality test). The Los Angeles prices provided no additional explanatory power above that provided by the Gulf Coast prices, which implies that California prices did not systematically "cause" Gulf Coast RFG prices to change during the last five years.
However, the methodology does not rule out the possibility that an occasional large price fluctuation in California can affect prices on the U.S. Gulf Coast. Nor does it rule out simultaneous price movements, in which price pressure in California could simultaneously cause price changes in California, U.S. Gulf Coast, and New York Harbor (which can also affect Gulf Coast) spot prices.
[1] See for example, Energy Information Administration, Assessment of Summer 1997 Motor Gasoline Price Increase, Appendix E, DOE/EIA-0621 (Washington, DC, May 1998) (ftp://ftp.eia.doe.gov/pub/oil_gas/petroleum/analysis_publications/assessment_1997_gasoline_price_increase/pdf/motor_gas_ass.pdf).
File last modified: June 12, 2001.
If you are having technical problems with this site, please contact the EIA Webmaster at wmaster@eia.doe.gov.
