Motor Fuels: California Gasoline Price Behavior (Letter Report,
04/28/2000, GAO/RCED-00-121).

Pursuant to a congressional request, GAO reviewed the behavior of
gasoline prices in California, focusing on: (1) the extent to which
retail gasoline prices spike more frequently and higher in California
than they do in the rest of the United States; (2) why gasoline prices
in California rise faster than they fall in response to increases and
decreases in the wholesale price of gasoline; and (3) the factors that
account for differences in the retail prices of gasoline between San
Francisco and Los Angeles.

GAO noted that: (1) according to GAO's analysis of gasoline price data,
from January 1995 through December 1999, retail gasoline prices spiked
no more frequently in California than they did in the rest of the United
States, but the spikes that did occur were generally higher in
California than elsewhere in the nation; (2) prices spiked seven times,
and during six of the spikes, the price increases were between 3 cents
and 31 cents per gallon higher in California than in the rest of the
United States; (3) many federal, state, and oil industry officials told
GAO that the higher price spikes in California were caused primarily by
unplanned refinery outages that disrupted the state's tight balance
between gasoline supply and demand; (4) according to the results of
statistical modeling by the Energy Information Administration (EIA),
retail gasoline prices in California rise faster than they fall in
response to a delayed pass-through of changes in the wholesale prices of
gasoline--a behavior that has been observed in other markets; (5) EIA
officials believe, however, that this price behavior has little or no
impact on consumers because their analysis shows that price increases
and decreases at the wholesale level are generally fully passed through
to the retail level, despite some delay; (6) oil industry officials and
experts GAO contacted also told GAO that retail prices generally fully
reflect changes in wholesale prices and that the observed price patterns
may be due to the way retail sellers react to these changes; (7) retail
gasoline prices are higher in San Francisco than in Los Angeles, in part
because of local supply and demand conditions; (8) retail gasoline
prices were, on average, about 11 cents higher in San Francisco than in
Los Angeles for the period from January 1992 through December 1999; (9)
among the local supply and demand conditions that are important in
explaining the price difference between the two cities are: (a) the
number and location of retail gasoline stations; (b) the costs of
building and operating gasoline stations; and (c) consumers' incomes;
(10) together, these conditions would be expected to lead to higher
retail gasoline prices in San Francisco than in Los Angeles, although
the exact magnitude of the effects on prices cannot be determined with
the available data; (11) the local supply and demand conditions GAO
identified may not entirely explain the price differences between the
two cities; and (12) other factors, such as competition at the refining
level, may help explain these differences, but GAO is unable to obtain
proprietary data that would have allowed GAO to explore this
possibility.

--------------------------- Indexing Terms -----------------------------

 REPORTNUM:  RCED-00-121
     TITLE:  Motor Fuels: California Gasoline Price Behavior
      DATE:  04/28/2000
   SUBJECT:  Petroleum prices
	     Gasoline
	     Comparative analysis
	     Fuel sales
	     Prices and pricing
	     Petroleum refining facilities
IDENTIFIER:  California
	     San Francisco (CA)
	     Los Angeles (CA)

******************************************************************
** This file contains an ASCII representation of the text of a  **
** GAO Testimony.                                               **
**                                                              **
** No attempt has been made to display graphic images, although **
** figure captions are reproduced.  Tables are included, but    **
** may not resemble those in the printed version.               **
**                                                              **
** Please see the PDF (Portable Document Format) file, when     **
** available, for a complete electronic file of the printed     **
** document's contents.                                         **
**                                                              **
******************************************************************

GAO/RCED-00-121

Appendix I: Comparison of Retail Gasoline Prices in California and Texas, a
Benchmark for the U.S. Gasoline Market

26

Figure 1: California's Average Daily Gasoline Consumption Compared With
Major World Markets, 1997 6

Figure 2: California's Gasoline Demand/Supply Network 8

Figure 3: California and U.S. Retail Gasoline Prices, Jan. 1, 1995,
Through Dec. 31, 1999 10

Figure 4: Comparison of California, Texas, and U.S. Retail Gasoline
Price Spikes, Jan. 1, 1995, Through Dec. 31, 1999 11

Figure 5: Retail Price Adjustments Relative to Changes in Wholesale
Prices 16

Figure 6: Retail Gasoline Prices in San Francisco and Los Angeles 19

Figure 7: California, Texas, and U.S. Retail Gasoline Prices, Jan. 1,
1995, Through Dec. 31, 1999 26

CARB California Air Resources Board

EIA Energy Information Administration

GAO General Accounting Office

Resources, Community, and
Economic Development Division

B-285102

April 28, 2000

The Honorable Dianne Feinstein
United States Senate

Dear Senator Feinstein:

Retail gasoline prices in the United States have risen sharply since early
1999, mostly in response to sharply rising world crude oil prices. Although
gasoline prices have, in general, been relatively low for U.S.
consumers--compared with both historical standards and the prices paid in
many other industrialized countries--sharply rising gasoline prices can
potentially have an adverse impact on U.S. consumers, as well as on the U.S.
economy. Moreover, during the second half of the 1990s, retail gasoline
prices throughout the United States have exhibited a high degree of
volatility and fairly frequent spikes. Particularly in California, where
consumers already generally pay higher average prices than they do elsewhere
in the United States, the spikes have raised questions about the behavior of
gasoline prices both within the state and between California and the rest of
the country.

Concerned about the higher gasoline prices and the extent of price spikes in
California, you asked us to analyze the behavior of gasoline prices in the
state. Because we found no standard definition of a gasoline price spike,
for this report, we define a spike as an increase of at least 6 cents per
gallon in a 4- to 21-week period. As agreed with your office, this report
addresses the following questions: (1) To what extent do retail gasoline
prices spike more frequently and higher in California than they do in the
rest of the United States, and what factors account for any difference? (2)
Do retail gasoline prices in California rise faster than they fall in
response to increases and decreases in the wholesale price of gasoline and,
if so, why? (3) What factors account for differences in the retail prices of
gasoline between San Francisco and Los Angeles?

According to our analysis of gasoline price data, from January 1995 through
December 1999, retail gasoline prices spiked no more frequently in
California than they did in the rest of the United States, but the spikes
that did occur were generally higher in California than elsewhere in the
nation. Prices spiked seven times, and during six of the spikes, the price
increases (the differences between the low and high prices) were between 3
cents and 31 cents per gallon higher in California than in the rest of the
United States. Many federal, state, and oil industry officials told us that
the higher price spikes in California were caused primarily by unplanned
refinery outages that disrupted the state's tight balance between gasoline
supply and demand. Because California refineries produce at almost full
capacity, supply disruptions caused by refinery outages must be made up from
other sources, such as out-of-state providers. However, obtaining gasoline
from such providers is slow and costly because only a few out-of-state
refineries can produce gasoline that meets the state's stringent
emission-reducing standards and the gasoline must be shipped by tanker from
far-away locations. In contrast, some West Coast retailers told us that
reduced competition at the refinery and retail levels caused the higher
California spikes. The Federal Trade Commission is currently investigating
gasoline prices in California and other West Coast states.

According to the results of statistical modeling by the Department of
Energy's Energy Information Administration, retail gasoline prices in
California rise faster than they fall in response to a delayed pass-through
of changes in the wholesale prices of gasoline--a behavior that has been
observed in other markets. The model was not designed to explain the factors
that account for this price behavior. Energy Information Administration
officials believe, however, that this price behavior has little or no impact
on consumers because their analysis shows that price increases and decreases
at the wholesale level are generally fully passed through to the retail
level, despite some delay. Oil industry officials and experts we contacted
also told us that retail prices generally fully reflect changes in wholesale
prices and that the observed price patterns may be due to the way retail
sellers react to these changes. The officials and experts were uncertain
about what effect, if any, this behavior could have on consumers.

Retail gasoline prices are higher in San Francisco than in Los Angeles, in
part because of local supply and demand conditions. Retail gasoline prices
were, on average, about 11 cents higher in San Francisco than in Los Angeles
for the period from January 1992 through December 1999. Among the local
supply and demand conditions that are important in explaining the price
differences between the two cities are (1) the number and location of retail
gasoline stations, (2) the costs of building and operating gasoline
stations, and (3) consumers' incomes. Together, these conditions would be
expected to lead to higher retail gasoline prices in San Francisco than in
Los Angeles, although the exact magnitude of the effects on prices cannot be
determined with the available data. The local supply and demand conditions
we identified may not entirely explain the price differences between the two
cities. Other factors, such as competition at the refining level, may help
explain these differences, but we were unable to obtain proprietary data
that would have allowed us to explore this possibility.

Gasoline prices in California and the rest of the United States have risen
sharply over the past year, largely because of increases in the price of
world crude oil, which rose from a low of about $12 per barrel in February
1999 to a high of about $34 per barrel in early March 2000. In spite of this
sharp rise, gasoline prices are still lower in real terms than they were at
their peak in 1981. For example, in early March of this year, the average
price of gasoline in the United States was $1.50 per gallon, compared with
about $2.47 in 1981 (in 1999 dollars). In addition, the amount of oil the
U.S. economy uses per unit of gross domestic product has decreased since
1979. Finally, average fuel economy for the new vehicle fleet--including
light trucks and sport utility vehicles--has risen slightly since 1980,
although it has been declining since about 1988 with the increasing share of
light trucks and sport utility vehicles, which have a lower average fuel
economy than cars. If fuel efficiency continues to decline, the impact of
higher gasoline prices on consumers will also rise. Moreover, the increased
price volatility over the last year may have adverse effects on consumers
and the U.S. economy.

California consumed almost 1 million barrels of gasoline per day in 1999,
more than any other state in the country. Furthermore, from 1996 through
1999, California's gasoline demand grew at an annual rate of about 1.4
percent. To put these data into perspective, in 1997 (the last year data
were available for international comparisons) California was the third
largest consumer of gasoline in the world, behind only the United States and
Japan and ahead of such major countries as China, Germany, and Russia (see
fig. 1).

Figure 1: California's Average Daily Gasoline Consumption Compared With
Major World Markets, 1997
Consumption in thousands of barrels per day.

Source: Energy Information Administration (EIA).

According to oil industry, federal, and California officials, in general,
California's gasoline demand is met almost entirely by supply from refinery
production within the state. In 1999, 23 refineries in California made
gasoline: 11, owned by five large refiners, had the capacity to make almost
95 percent of the gasoline refined in the state, and 2, owned by independent
refiners, had the rest of the gasoline-producing capacity. Other refineries
made other petroleum products, such as asphalt and lubricants. Some
conventional gasoline and gasoline that met federal standards for
reformulated gasoline made by California refineries was shipped primarily to
other West Coast markets, such as Oregon, Arizona, and Nevada.1 Gasoline is
transported primarily by pipeline from the refineries to storage terminals
and then, typically, by truck from the storage terminals to retail gasoline
stations.

The remaining supply comes from using existing gasoline inventories and from
out-of-state providers. Gasoline brought into California from the U.S. Gulf
Coast or other out-of-state locations typically travels by water on tankers
or barges. Inventories and out-of-state providers generally play a minimal
role except during disruptions in refinery production, when they become
important supply sources. Events that substantially disrupt the supply of
gasoline through this system could have a significant impact on the prices
paid by consumers. Figure 2 shows California's gasoline demand/ supply
network.

Figure 2: California's Gasoline Demand/Supply Network
Note: Demand/supply data are for 1998.

Source: EIA for demand/supply data; an oil industry official for supply
percentages.

To improve its air quality, California established gasoline standards that
are more stringent than the federal standards and different from those of
any other state. On March 1, 1996, California implemented a program that
exceeded the federal requirements for states to use reformulated gasoline in
areas with serious ozone problems. The California reformulated gasoline
program is administered by the California Air Resources Board (CARB). To
make reformulated gasoline to meet the more stringent California standards,
referred to as CARB gasoline, California refiners invested billions of
dollars to modify their refineries to add sophisticated equipment and
processes needed to make such gasoline. According to several industry
officials and experts and CARB officials, some refiners, especially smaller
ones, that could not make the needed modifications, partly because of high
modification costs, shut down their refineries. This contributed to the
reduction in the number of refineries in California that can make gasoline.

the Rest of the Nation but Were Generally Higher, Primarily Because of
Refinery Outages

The retail price of regular gasoline spiked the same number of times in
California and in the rest of the United States from 1995 through 1999.2
However, all but one of the price spikes were higher in California than
elsewhere in the country. Many federal, state, and oil industry analysts and
officials believe that the California spikes were higher primarily because
unplanned refinery outages disrupted the state's tight balance between
gasoline supply and demand. In contrast, some West Coast gasoline retailers
believe that the higher California spikes resulted from reduced competition
at the refinery and retail levels.

Were Generally Higher

Regular gasoline retail prices spiked seven times in California and in the
rest of the United States (excluding California) from January 1, 1995,
through December 31, 1999, as shown in figure 3. Moreover, five of the seven
California spikes started at about the same time as the U.S. spikes, and six
California spikes overlapped the corresponding U.S. spikes by at least 4
weeks. Generally, these spikes coincided with increases in crude oil prices
and increases in the demand for gasoline during the spring and summer
driving seasons.3 Price spikes in California and in the rest of the United
States occurred at different times only between the fall of 1996 and the
spring of 1997. During this period, U.S. prices spiked in the fall of 1996,
and California prices spiked in the spring of 1997, with a 1-week overlap.

Figure 3: California and U.S. Retail Gasoline Prices, Jan. 1, 1995, Through
Dec. 31, 1999
Source: EIA.

In terms of size, retail gasoline price spikes were higher in California
than in the rest of the United States from January 1, 1995, through December
31, 1999, with one exception, as shown in figure 4. During six of the
spikes, the California price increases (the differences between the low and
high prices) were between 3 cents and 31 cents per gallon greater than the
corresponding price increases in the rest of the United States. The smallest
difference occurred in the summer of 1999, when California prices spiked 3
cents per gallon higher than prices in the rest of the United States (20
cents versus 17 cents). The largest difference occurred in the spring of
1999, when California prices spiked 31 cents per gallon higher (53 cents
versus 22 cents). Conversely, U.S. prices spiked 7 cents per gallon higher
than California prices (14 cents versus 7 cents) in the spring of 1995.

Figure 4: Comparison of California, Texas, and U.S. Retail Gasoline Price
Spikes, Jan. 1, 1995, Through Dec. 31, 1999

Source: EIA.

In addition, the relationship between California and U.S. prices changed
after CARB gasoline requirements were implemented. The difference between
gasoline prices in California and in the rest of the United States increased
by about 6 cents per gallon--California prices were 10.5 cents per gallon
higher than U.S. prices before CARB and 16.9 cents per gallon higher after
CARB.

A comparison of the number and size of gasoline retail price spikes in
California and in Texas--a large refining state that is comparable to
California in terms of its role in the U.S. gasoline market--corroborated
our finding that spikes were no more frequent but were generally higher in
California than in the rest of the United States. (See app. I for details of
this related analysis.)

Refinery Outages, but Other Factors May Have Contributed

Many oil officials and analysts told us that refinery outages were the
primary reason California gasoline prices spiked higher than prices in the
rest of the United States. However, some West Coast retailers believe that
reduced competition was the primary reason.

Unplanned Refinery Outages Were the Primary Cause of Higher California Price
Spikes

California refineries had unplanned outages every year from 1995 through
1999. When such outages disrupted the California gasoline supply, oil
companies met demand with gasoline from other sources. They obtained
gasoline from out-of-state providers, used existing inventory, and increased
production at California refineries whose operations were not disrupted.
Obtaining gasoline from such sources was necessary when refinery outages
significantly disrupted California's supply, as they did in the following
instances:

ï¿½ On April 1, 1996, an explosion at the Shell refinery in northern
California virtually shut down the refinery's production, which amounted to
about 100,000 barrels of gasoline a day. Before the Shell refinery was fully
repaired, explosions and mechanical problems disrupted operations at several
other refineries. According to the Energy Information Administration (EIA),
these disruptions affected about 12 percent of the state's production for
several months. Our analysis showed that California gasoline prices spiked
about 39 cents per gallon that spring. The spike was primarily due to the
refinery disruptions, according to CARB and oil industry officials. Gasoline
was brought into California from as far away as Finland to make up for the
lost production.

ï¿½ An explosion at Tosco's northern California refinery in February 1999 and
subsequent outages in at least three other California refineries
significantly disrupted gasoline production for several months, adversely
affecting 12 to 15 percent of the state's production, according to EIA and
others. California Energy Commission officials and oil industry analysts
told us that these outages forced some oil companies to buy gasoline on the
spot market, driving up wholesale prices and, consequently, retail prices.
Our analysis showed that California retail prices spiked 53 cents per gallon
that spring. Gasoline from U.S. Gulf Coast, U.S. Virgin Islands, and foreign
refineries helped lower prices. However, additional problems at several
California refineries in the summer disrupted the state's supply again, and
these disruptions were exacerbated by a June 10 explosion that shut down
part of the Olympic Pipeline, which transports thousands of barrels of
gasoline a day from Washington State to Oregon. Federal, state, and oil
industry officials told us that the West Coast gasoline market is
interrelated and that a major supply disruption anywhere in the region
affects supply and prices throughout the region. Our analysis showed that
California retail prices spiked 20 cents per gallon that summer. According
to EIA, gasoline from the U.S. Gulf Coast and U.S. Virgin Islands was used
to meet California's gasoline demand.

Bringing gasoline into California is slow and costly because California is
isolated from out-of-state sources in two ways. First, only a few refineries
outside the state can make gasoline that meets the state's CARB gasoline
requirements. These few refineries are not set up to make CARB gasoline
routinely, and they have to reconfigure their refining operations to produce
it. Some oil industry officials told us that making the decision and
reconfiguring for CARB gasoline production takes up to a week and adds costs
for blending, storing, and segregating the gasoline. Second, because
California has no pipelines that can bring gasoline into the state, tankers
and other means must be used. According to oil industry analysts, CARB
gasoline has been brought into California by tankers from the U.S. Gulf
Coast, the U.S. Virgin Islands, and countries as far away as Finland,
Singapore, and South Korea. According to EIA and oil industry officials and
analysts, shipping gasoline into California from these locations takes
between 11 and 40 days and adds 3 to 12 cents per gallon to the retail
price.

To a limited extent, oil companies have also used gasoline in inventory and
have increased output at uninterrupted refineries to meet demand when some
California refineries' production has been disrupted. California inventories
offered little potential relief because oil companies maintain relatively
low inventories to avoid tying up resources. Similarly, California
refineries can increase their production to only a limited degree because
they are already operating at almost full capacity. According to CARB
officials, California refineries were operating at about 97 percent of
capacity in 1999.

Reduced Competition May Have Contributed to Higher California Spikes

Several West Coast retailers we contacted and others believe that reduced
competition at the refinery and retail levels was the primary reason why
California spikes were higher than U.S. spikes from 1995 through 1999.
According to the executive director of the California Service Stations and
Automotive Repair Association, which represents about 850 gasoline service
stations, a lack of competition in California caused the spring 1999 spike.
Testifying before the California legislature on April 28, 1999, he noted
that consumption in the state increased 30 percent between 1982 and 1999,
while the number of refineries decreased from 43 to 23 and the number of
service stations decreased from 14,687 to 9,513. Similarly, representatives
from the Automotive Trade Organizations of California, which represents the
owners of over 2,000 service stations, repair facilities, and related
businesses, told us that reduced competition was the major cause of the
California price spikes. Additionally, the executive director of the
Automotive United Trade Organization, which is headquartered in Washington
State, attributed the generally higher gasoline price spikes in California
and other West Coast states to the limited competition facing large oil
companies in these states at both the refiner and the retail levels. In
addition, referring to a 1999 preliminary report on California gasoline
prices, the California attorney general issued a press release expressing
concern that the relative lack of competition in California contributed to
the state's high gasoline prices.4 We could not confirm that reduced
competition was the primary reason for the higher California price spikes
because the information needed to do so, such as oil companies' pricing
formulas, plans, or policies, was not readily available.

The Federal Trade Commission is investigating gasoline prices in West Coast
states, including California, Oregon, and Washington. This investigation was
prompted by allegations of anticompetitive behavior by oil companies. As of
January 2000, Commission officials had reached no conclusions on the matter.

Response to Changes in Wholesale Prices

According to statistical modeling completed by EIA at our request, from
April 1996 through July 1999, retail gasoline prices in California rose
faster than they fell in response to changes in the wholesale prices of
gasoline. In other words, the pattern of retail price adjustments to
increases and decreases in the wholesale prices of gasoline was asymmetric.5
The modeling did not separately estimate how much faster retail prices
increased versus decreased in response to wholesale price changes. A
previous study by EIA also found this asymmetric pattern for price
adjustments in the Midwest, and the pattern is consistent with the findings
of several other studies cited in the EIA report.6 Figure 5 illustrates an
asymmetric pattern of retail price adjustments to wholesale price increases
and decreases using California price data over a 10-week period during the
price spike in the spring of 1999.

Figure 5: Retail Price Adjustments Relative to Changes in Wholesale Prices
Source: EIA and the Oil Price Information Service.

As the figure illustrates, when wholesale prices rose, retail prices
adjusted to this increase by also rising, but with a lag. However, the
figure also shows that before retail prices reached their peak, wholesale
prices began to fall. Again, retail prices responded to wholesale prices by
falling, but with a lag. The figure also shows that retail prices rose at a
faster rate than they fell. Although retail prices did not reach the peak of
the increase in wholesale prices before the latter started falling, they
stayed up longer, or fell more slowly, than the decline in wholesale prices.
In this example of an actual price spike, retail prices rose for 4 weeks and
fell for 6.

The finding that retail gasoline prices have risen faster than they have
fallen in response to wholesale price changes may have little or no
implication for gasoline consumers apart from a consideration of the extent
to which the wholesale price changes were passed on to consumers. EIA
officials told us that their analyses of the data for California and other
U.S. markets have shown that while the time taken for wholesale price
changes to be fully reflected at the retail level varies among markets, all
increases and decreases in wholesale prices were completely passed through
to the retail level. According to these officials, because both increases
and decreases in wholesale prices are ultimately fully passed through to the
retail level, the pattern of the pass-through has little or no adverse
impact on consumers. Many of the oil industry officials and experts we
contacted also believe that, in general, because of competition at the
retail level, retail prices fully reflect wholesale price changes, although
they do so with a lag when prices are changing.

To understand why retail gasoline prices may rise faster than they fall in
response to wholesale price changes, and because there is no consensus in
the economic literature as to why, we discussed this asymmetric price
pattern with EIA and other oil industry officials and experts. EIA officials
said that the observed pattern is almost entirely driven by the way retail
prices respond with a lag to changes in wholesale prices. In general, this
explanation is consistent with the description of the price patterns
depicted in figure 5 above.

In our discussions with oil industry officials and experts, several of them
said that retail gasoline prices probably rise faster than they fall in
response to wholesale price changes because retailers try to make up, during
falling prices, for revenues lost when wholesale prices were rising.
According to some oil industry officials and experts, although retail prices
may rise fairly quickly in response to increases in wholesale prices, the
increases in retail prices may not always fully reflect the wholesale price
increases. They said that retailers exercise caution in raising their prices
when wholesale prices are rising to avoid decreased sales and to forestall
any backlash from consumers and public officials. Therefore, they explained,
when wholesale prices fall, retailers lower prices more slowly in an attempt
to recoup revenues lost when prices were rising. Furthermore, some argued
that while retail prices may be slow to follow when wholesale prices fall,
competition eventually forces retail prices down. The officials and experts
we contacted said they did not know what impact this pattern of price
adjustments would have on consumers. Furthermore, they pointed out that such
price patterns are generally short lived and are not typical of long-run
price behaviors.

Angeles, in Part Because of Local Supply and Demand Conditions

Retail gasoline prices are higher in San Francisco than in Los Angeles, and
these differences have increased since the introduction of CARB gasoline in
1996. The price differences between the two cities are explained in part by
local supply and demand conditions, including (1) the number and location of
retail gasoline stations, (2) the costs of building and operating gasoline
stations, and (3) consumers' incomes. These local supply and demand
conditions may not entirely explain the price differences between the two
cities. Other factors, such as competition at the refining level, may help
explain these differences, but we were unable to obtain proprietary data
that would have allowed us to explore this possibility.

Since at Least 1992

We examined retail gasoline prices from January 1992 through December 1999
and found that with few exceptions, San Francisco had higher prices than Los
Angeles.7 According to data from the Oil and Gas Journal, the average
difference was about 11 cents over the entire period. Moreover, as discussed
below, the average difference grew larger in the second half of the period,
from March 1996 through 1999, and grew still further in 1999.8 Figure 6
shows retail gasoline prices in the two cities from January 1992 through
December 1999.

Figure 6: Retail Gasoline Prices in San Francisco and Los Angeles
Source: Oil and Gas Journal.

From January 1992 through February 1996, the retail price of regular
gasoline was, on average, about 6 cents higher in San Francisco than in Los
Angeles, but this difference increased to about 17 cents over the period
from March 1996 through December 1999. In addition, refinery shutdowns in
the Bay Area in the summer of 1999 further increased the price difference
between the two cities to about 38 cents in August and September--the
highest difference between 1992 and 1999. The price difference remained
above 26 cents through December 1999.

San Francisco and Los Angeles

In general, local supply and demand factors help explain why retail gasoline
prices are higher in San Francisco than in Los Angeles. On the supply side,
according to some experts and industry officials we interviewed, one key
factor explaining the price differences is that consumers have fewer places
to buy gasoline in San Francisco than in Los Angeles. In 1996, for example,
there were about 19 gasoline stations in San Francisco for every 100,000
people, compared with about 25 stations in Los Angeles. One explanation for
why there are fewer gasoline stations per capita in San Francisco than in
Los Angeles is that land is relatively more developed in San Francisco,
which raises the cost of acquiring a site for a gasoline station. For
example, according to a recent study, gasoline station development
costs--real estate and construction costs--are about 50 percent higher in
San Francisco than in Los Angeles.9 In addition, zoning and other
regulations make it harder for station owners in San Francisco to operate
convenience stores on the same property as gasoline stations and therefore
eliminate profitable secondary sales. Being unable to spread high land costs
over gasoline and convenience store sales would tend to make the costs of
selling gasoline and also its price higher in San Francisco than in Los
Angeles.

On the demand side, the annual per-capita consumption of gasoline is higher
in San Francisco than in Los Angeles--in 1996, about 520 and 390 gallons,
respectively. Tourists consume part of the gasoline, and on a per-capita
basis, more tourists visit San Francisco than Los Angeles. For example,
according to a survey done in 1998 for the California Department of
Commerce, San Francisco County--with a population of around 783,000--was
host to about 10 million leisure visitors, or about 13 per capita. In
contrast, Los Angeles County--with a population of about 9,587,000--had
about 24 million visitors, or about 2.5 per capita. Travelers to San
Francisco County were also more likely to rent cars--about 12 percent of the
leisure visitors to San Francisco rented cars compared with about 8 percent
for Los Angeles. Besides increasing the total demand for gasoline, tourists
may be less inclined than residents to search for low prices, enabling
gasoline stations in areas with high tourist traffic to charge higher
prices. Per-capita incomes are also higher in San Francisco than in Los
Angeles, which may make the demand for gasoline in San Francisco less
sensitive to price. Finally, the market structure of the two areas may play
a role in explaining the price differences. For example, ARCO−a
self-declared and commonly recognized seller of low-priced gasoline at the
retail level−has a bigger market share in southern California than in
northern California, potentially contributing to the lower prices in Los
Angeles.

The local supply and demand conditions that led to generally higher prices
in San Francisco than in Los Angeles may also explain why these price
differences rose when CARB gasoline was introduced in 1996 and why they rose
further during refinery outages in 1999. The introduction of CARB gasoline
in March 1996 caused prices to rise in California relative to the rest of
the United States, in part by raising the cost of refining gasoline. At the
same time, the supply of gasoline in California became more sensitive to
supply disruptions because no outside source of CARB gasoline is readily
available. When the cost of producing gasoline rose, refiners would have
passed at least some of the cost on to retailers in the form of higher
wholesale gasoline prices, in turn causing retail prices to rise. However,
both wholesale and retail prices apparently increased more in San Francisco
than in Los Angeles--the gasoline spot price (a wholesale price) rose about
2 cents more in San Francisco, and the difference in retail prices between
the two cities increased from 6 cents to 17 cents, an increase of 11
cents.10 There is no consensus among experts and industry officials as to
why higher price increases occurred in San Francisco. One explanation
offered is that higher refining costs are easier to pass on to consumers in
San Francisco because of its local supply and demand conditions. Another is
that the new fuel requirements might have tightened the gasoline supply and
demand balance more in the northern part of the state than in the southern
part. Consensus is also lacking as to why the refinery shutdowns in 1999
caused such a large increase in the retail price difference between the two
cities, particularly since gasoline can be shipped by barge between San
Francisco and Los Angeles for between 2 and 4 cents per gallon. However, as
noted, local supply and demand conditions may make it easier to pass on
refinery costs and wholesale price increases in San Francisco than in Los
Angeles.

Prices

The 1999 preliminary report on California gasoline prices for the California
attorney general concluded that there is less competition at the refiner
level in California than in the rest of the United States. The report stated
that refiners engage in the practice of zone pricing, which enables them to
charge different wholesale prices to different retail dealers according to
what the market will bear. The report stated that retail dealers pay higher
wholesale prices in San Francisco (17 cents higher for the first 9 months of
1999) than in Los Angeles and that these differences in wholesale prices
explain most of the differences in retail prices between the two cities.
Although zone pricing is not unique to California, this practice could be a
significant cause of retail price differences between San Francisco and Los
Angeles. However, we were unable to obtain proprietary data on the actual
wholesale prices paid by specific retail dealers, and without this
information, we could not explore this possibility. Moreover, the ability of
refiners to engage in and benefit from zone pricing depends to a large
extent on other factors we have addressed in this report. For example,
refiners may not be able to charge higher wholesale prices for gasoline if
competition among retail dealers will preclude them from passing the higher
prices on to consumers.

We provided a draft of this report to the Department of Energy and EIA for
review and comment. We discussed the report with EIA officials, including
the Director, Petroleum Division. EIA agreed with the report and provided
clarifying comments that we incorporated, where appropriate.

To determine the extent to which retail gasoline prices spike more
frequently and higher in California than in the rest of the United States,
we obtained and analyzed average weekly price data from EIA for selected
retail regular gasoline markets for the period from January 1, 1995, through
December 31, 1999. Specifically, we compared price data for California
reformulated gasoline with price data for all formulations of U.S. gasoline
(excluding California). We also compared price data for California
reformulated gasoline with price data for all formulations of Texas
gasoline, using Texas as a benchmark state for U.S. prices. Specifically, we
calculated the differences between the low and high gasoline prices in
California, Texas, and the rest of the United States during the periods we
identified as spikes--when the California and U.S. prices increased at least
6 cents per gallon in a 4- to 21-week period. To ascertain the reasons for
the differences, we reviewed expert studies and relevant federal and state
records, and we interviewed officials and experts in the oil industry
(selected oil companies, consulting firms, and trade organizations) and at
EIA and the Federal Trade Commission, the California Energy Commission and
CARB, and the University of California at Berkley and Purdue University.

To determine whether California retail gasoline prices rise faster than they
fall in response to changes in wholesale gasoline prices, we worked with EIA
to develop and interpret an econometric model. This type of model is
generally used by energy analysts to determine whether the prices of
petroleum products, such as gasoline and home-heating oil, rise at a
different rate than they fall in response to wholesale or even crude oil
price changes--a phenomenon commonly referred to by analysts as price
asymmetry. We used this model to analyze the response of retail prices to
wholesale price changes from April 1996 through July 1999. To the extent
possible, we used data in EIA's database, which we supplemented with data
purchased from the Oil Price Information Service (a private vendor).
However, these purchased data were not available at the level of detail
needed to fully explain price behavior. To determine the reasons for the
gasoline price asymmetry, we interviewed officials and experts in the oil
industry, EIA, state agencies, and academia. We also reviewed existing
studies and economic literature on gasoline markets.

To determine the extent to which gasoline prices were higher in San
Francisco than in Los Angeles, we analyzed data on retail and wholesale
gasoline prices for the San Francisco/Bay Area and Los Angeles. To determine
the reasons for the differences in the gasoline prices for these two areas,
we assessed the potential effects on gasoline prices of such factors as
geographical characteristics, barriers to market entry, and cost
differences. A lack of data prevented us from fully describing the
importance of all these variables. We also interviewed officials from the
oil industry, EIA, the California Energy Commission, and academia.

We conducted our work between June 1999 and March 2000 in accordance with
generally accepted government auditing standards.

Unless you publicly announce its contents earlier, we plan no further
distribution of this report until 7 days after the date of this letter. At
that time, we will send copies to appropriate congressional committees and
interested Members of Congress. We will also make copies available to others
upon request.

If you have any questions about this report or need additional information,
please call me at (202) 512-3841. Key contributors to this report included
Daniel Haas, Godwin Agbara, Byron Galloway, and Frank Rusco.

Sincerely yours,
Barry T. Hill
Associate Director, Energy,
Resources, and Science Issues

Comparison of Retail Gasoline Prices in California and Texas, a Benchmark
for the U.S. Gasoline Market

We compared gasoline prices in California and Texas to determine whether
California prices were different from prices in the rest of the United
States. We selected Texas as a benchmark for the comparison because Texas,
like California, played a major role in the U.S. gasoline market from
January 1, 1995, through December 31, 1999. Texas and California,
respectively, were the first and third largest refining states and the
second and first largest consuming states.

Gasoline prices in Texas followed a pattern similar to prices in the rest of
the United States (excluding California and Texas), increasing seven times,
as shown in figure 7. However, two Texas price increases, which averaged
less than 6 cents per gallon in the fall of 1996 and the spring of 1998, did
not meet the criteria for a spike that we applied to California price
increases.

Figure 7: California, Texas, and U.S. Retail Gasoline Prices, Jan. 1, 1995,
Through Dec. 31, 1999
Source: Energy Information Administration.

The size of the retail gasoline price spikes was greater in California than
in Texas. Six of the seven California price spikes were between 4 and 34
cents per gallon higher than the corresponding Texas price spikes or
increases. The smallest difference occurred the summer of 1999 (20 cents per
gallon in California versus 16 cents per gallon in Texas), and the largest
difference occurred in the spring of 1999 (53 cents per gallon in California
versus 19 cents per gallon in Texas). Once, in the spring of 1995, Texas
prices spiked 6 cents per gallon higher than California prices (13 cents
versus 7 cents per gallon).

(141367)

Figure 1: California's Average Daily Gasoline Consumption Compared With
Major World Markets, 1997 6

Figure 2: California's Gasoline Demand/Supply Network 8

Figure 3: California and U.S. Retail Gasoline Prices, Jan. 1, 1995,
Through Dec. 31, 1999 10

Figure 4: Comparison of California, Texas, and U.S. Retail Gasoline
Price Spikes, Jan. 1, 1995, Through Dec. 31, 1999 11

Figure 5: Retail Price Adjustments Relative to Changes in Wholesale
Prices 16

Figure 6: Retail Gasoline Prices in San Francisco and Los Angeles 19

Figure 7: California, Texas, and U.S. Retail Gasoline Prices, Jan. 1,
1995, Through Dec. 31, 1999 26
  

1. Reformulated gasoline is designed to reduce harmful exhaust emissions
that cause smog.

2. Because we found no standard definition of a gasoline price spike, we
analyzed gasoline prices in California and in the rest of the United States
(excluding California) between Jan. 1, 1995, and Dec. 31, 1999, to identify
apparent spikes. During that time, there were seven periods when California
and U.S. prices increased by at least 6 cents per gallon in a relatively
short period of time--from 4 to 21 weeks. For this report, we refer to these
increases as spikes.

3. The director of EIA's Petroleum Division testified on Mar. 9, 2000, that
gasoline retail prices in the United States were about $1.50 per
gallon--about 23 cents higher than at the beginning of the year--citing
increases in crude oil prices as a major contributing factor.

4. Keith Leffler and Barry Pulliam, Preliminary Report to the Attorney
General Regarding California Gasoline Prices (Nov. 22, 1999).

5. EIA's statistical modeling also tested for the pattern of gasoline retail
price increases and decreases in response to changes in the wholesale price
in Texas, using Texas data for the same period. The result showed that
retail prices in Texas also rose faster than they fell in response to
changes in wholesale prices.

6. See Price Changes in the Gasoline Market: Are Midwestern Gasoline Prices
Downward Sticky? DOE/EIA, Washington, D.C. (Mar. 1999).

7. We chose the period from 1992 through 1999 so that we could examine the
price differences between San Francisco and Los Angeles before and after
CARB gasoline was introduced.

8. We used data from the Oil and Gas Journal to calculate the differences in
retail gasoline prices between San Francisco and Los Angeles. We compared
these data with data from the Lundberg survey that we received from an
industry source. The Lundberg data showed a similar pattern of price
differences between San Francisco and Los Angeles, but the differences were
about 3 cents per gallon smaller. For example, the average difference
between prices from Jan. 1992 through Feb. 2000 was about 8 cents using
Lundberg data compared with 11 cents using Oil and Gas Journal data.
Similarly, according to Lundberg data, the average differences before and
after CARB were about 3 cents and 14 cents, respectively, compared with 6
cents and 17 cents using Oil and Gas Journal data.

9. Gasoline Station Development Issues in San Francisco, prepared for the
Western States Petroleum Association by Kosmont & Associates, Inc. (Jan.
1998).

10. We were unable to gather data on the two other principal wholesale
prices--"rack" and "dealer tank wagon"--so we cannot determine the extent to
which wholesale prices in general changed when CARB gasoline was introduced.
*** End of document. ***