The Economic Effects of the Marathon - Ashland Joint Venture: The Importance of Industry Supply Shocks and Vertical Market Structure

The Economic Effects of the Marathon - Ashland Joint Venture: The
Importance of Industry Supply Shocks and Vertical Market Structure.
Christopher T. Taylor1
Federal Trade Commission
Daniel S. Hosken
Federal Trade Commission
Last Revised: May 7, 2004
Abstract: This study measures the retail and wholesale price effects in Louisville, Kentucky
resulting from the Marathon/Ashland (MAP) joint venture. MAP was an early transaction in the
recent era of petroleum mergers and it caused sizeable changes in concentration. We find no
evidence of increased retail prices resulting from MAP. Wholesale (rack) prices increased
significantly approximately 15 months following the transaction. This wholesale price (rack)
increase, however, was probably caused by a regional supply shock rather than the transaction.
These results suggest in this case that a significant petroleum merger in a moderately
concentrated market did not raise consumer prices.
JEL Classification: L1, L41, L71
Keywords: Merger retrospectives, Petroleum industry, Vertical market structure
1 The authors are economists at the Federal Trade Commission. Views and opinions expressed
in this paper are solely those of the authors and should not be interpreted as reflecting the views
of the Federal Trade Commission, any of its individual Commissioners, or other members of the
staff. Comments by Mary Coleman, Leslie Farber, Jeff Fischer, Mark Manusazak, Paul Pautler,
Louis Silvia, John Simpson, Shawn Ulrick, and participants at the Industrial Organization
Society Conference and excellent research assistance by Ryan Toone are appreciated.

---

I. Introduction
The U.S. petroleum industry has undergone substantial restructuring since the mid 1990's.
Among the major industry events were the creation of the Shell-Texaco and Marathon-Ashland
joint ventures, and the BP-Amoco, Exxon-Mobil, BP-ARCO, Chevron-Texaco, and Phillips-
Conoco mergers. Critics of the industry contend that the increase in concentration from these
transactions has led to higher prices. Some government officials have called for a moratorium on
petroleum mergers.2 In contrast, the industry contends that these mergers have led to
considerable costs savings. Before the Exxon-Mobil merger was completed the companies
predicted that they would save $2.8 billion a year in costs. Two years after the merger was
completed Exxon-Mobil stated they had achieved $4.6 billion dollars a year in savings.3
Despite the size of the petroleum industry and the controversy surrounding petroleum
mergers, there have been surprisingly few attempts to examine the effect of mergers on the price
of gasoline.4 The few papers examining petroleum mergers typically either estimate the effects
of a large number of mergers in a single study, or only examine one level of the industry,
typically wholesale (rack) pricing.5 The conventional approaches taken to study petroleum
2“I urge Congress to enact a moratorium of at least one year on any merger or acquisitions of
any major oil refiner, supplier or retailer, including cross-sector mergers and acquisitions, while
Congress, the FTC and the states work together to fashion a longer term remedy that helps
restore competitive forces and tempers the market dominance wielded by the few industry
giants.” Testimony of Connecticut Attorney General Richard Blumenthal Before the Permanent
Subcommittee on Investigations of the Senate Governmental Affairs Committee, May 2, 2002.
3ExxonMobil Corporation, Investor and Media Meeting, New York, Aug. 1, 2000, pp. 36-37.
4 There have been attempts to indirectly look at merger effects by examining changes in
concentration. (GAO, 1986) Simply using concentration as a proxy for merger effects is
problematic on a number of theoretical and practical levels, e.g. the difficulty of defining markets
correctly and controlling for endogenous market structure. (Evans et al., 1993)
5The most commonly examined wholesale price for gasoline is the rack price. The rack price
is the price posted at the truck rack at a terminal for trucks loading branded or unbranded
gasoline. The percentage of wholesale transactions taking place at the rack prices varies by
geography and by firm.
2

---

mergers are problematic for two reasons. First, examining multiple mergers in a single study is a
virtually untenable task. The creation of boutique fuel specifications to comply with
environmental regulations has Balkanized gasoline distribution in the U.S.6 Each region of the
U.S. is subject to different idiosyncratic sources of price variation, such as supply outages, input
price fluctuations, seasonal changes in marginal supply and formulation changes. In order to
ascertain how prices changed as the result of a change in market structure, the researcher must
control for all of these complicating factors. Second, researchers should be careful about
measuring merger affects by examining wholesale (rack) prices alone. In any gasoline market,
there are multiple wholesale prices being charged to gasoline retailers, only some of which are
publicly observable.7 In addition, because petroleum mergers often affect the vertical structure of
a local gasoline market, any given transaction may affect the retail markup a retail outlet earns,
while having little effect on the retail price of gasoline.8
For these reasons, in this study we examine one transaction, the refining and marketing
joint venture of Marathon and Ashland to form Marathon Ashland Petroleum (MAP). The MAP
transaction proceeded with no antitrust challenge or divestiture. Testimony by various
participants before the Permanent Subcommittee on Investigations of the Senate Governmental
Affairs Committee, on May 2, 2002 suggested that the increased concentration from this merger,
and mergers in general, have led to higher or more volatile gasoline prices in the Midwest.9 In
6Before the changes in gasoline specifications brought about by the Clean Air Act there was
one gasoline specification in the country, now there are 18. Energy Information Administration,
Petroleum Supply Monthly, April 1999.
7Also the relationship between these different wholesale prices may change, often in response
to supply outages. For example, lessee dealer stations, a station owned by a major oil company
leased by an independent marketer, pay a “dealer-tank-wagon” or DTW price which is typically
higher than the posted rack price, but when refineries have supply problems, the DTW price is
often less than the posted rack price.
8There are a number of theoretical models that demonstrate how mergers, both horizontal and
vertical mergers may affect upstream (wholesale) but not downstream (retail) prices. For
examples of these types of models see, Ordover et al., (1990) and Froeb et al., (2002).
9“Increased concentration in the refining and distribution segment of the industry has
contributed to the exercise of market power by dominant industry actors to the detriment of
3

---

this paper we examine how the retail and wholesale prices of gasoline in arguably the most
potentially problematic area, Louisville, Kentucky, changed as a result of the joint venture. We
use the wholesale and retail price of gasoline in a number of cities as controls in estimating
whether the retail or wholesale price of gasoline changed in Louisville as a result of the joint
venture.
Retail gasoline prices in Louisville do not appear to increase as a result of the joint
venture. These findings are robust when comparing the retail price in Louisville to three control
markets. The wholesale (rack) prices of reformulated gasoline (RFG) increased 3-5 cents per
gallon approximately 15 months after the transaction. This wholesale price (rack) effect,
however, seem to be the result of a supply shock caused by St. Louis’s switch to RFG rather than
the joint venture. The difference in the retail and wholesale (rack) price changes demonstrates
that it is crucial to examine both retail and wholesale pricing when measuring the price effects of
a merger affecting gasoline markets. The finding that the wholesale price increase is not passed
through at retail is somewhat surprising. In this market, it appears that retailers directly supplied
by refiners, representing 30% of gasoline sales, did not experience a wholesale price increase in
1999. Apparently those stations facing the higher wholesale (rack price) were not able to pass
through enough of the price increase to affect the average market price because of competition
with stations directly supplied by refiners.10
The paper is organized as follows. The second section provides industry background and
then describes the structure of the MAP joint venture. Section three reviews the methodologies
used in merger retrospectives for various industries and those research papers that focus on
consumers.” and “Although not as large as the mergers referenced above on a national scale, the
most significant transactions in Michigan petroleum markets involve the merger of Marathon and
Ashland Petroleum and then later Marathon Ashland Petroleum’s acquisition of all the Ultramar
Diamond Shamrock assets in the State.” Testimony of the Michigan Attorney General Jennifer
Granholm before the Permanent Subcommittee on Investigations of the Senate Governmental
Affairs Committee, May 2, 2002.
10Competition from stations selling conventional gasoline which did not experience a
wholesale price increase directly across the Ohio River in Indiana or in Kentucky, outside the
RFG area, may also have limited the ability of rack supplied stations to pass thru the wholesale
price increase.
4

---

potential price effects of petroleum mergers. The fourth section describes the data used in the
analysis. The fifth and sixth sections discuss the results of the analysis and the interpretation of
the results, and the last section discusses conclusions.
II. Industry Background and the Marathon/Ashland Transaction
A. Industry Background
Empirical analysis of gasoline pricing in the United States requires some familiarity with
the institutional structure of the gasoline refining and distribution system that affect the pricing of
gasoline. This section discusses the institutional structure that is of particular relevance for
estimating the price effects of MAP. There are five main components of retail gasoline prices
(costs): crude oil acquisition cost, refining costs, distribution and marketing costs, and taxes.
The size and volatility of refining, wholesaling and marketing costs in different regions of the
United States are affected by the myriad of gasoline formulations used in various regions and the
multiple sources of supply to a given region. In addition to conventional gasoline, other fuel
specifications exist which are designed to reduce emissions and air pollution. Such
specifications are usually some form of oxygengenated or reformulated gasoline (RFG). The
federal government has developed specifications for RFG, and there are different specifications
for the North and South and some areas uses a different oxygenate, either MTBE (methyl
tertiary-butyl ether) or ethanol. Some areas have their own formulations to satisfy federal clean
air requirements without using RFG. These “boutique” gasoline formulations tend to cost less to
produce, on average. However, in periods of supply disruption, e.g., a refinery outage, it can be
difficult for refiners to ship gasoline to an affected region quickly because alternative supplies of
that region’s specific type of gasoline may not be readily available.
A city’s source of gasoline supply varies significantly throughout the United States. The
eastern half of the United States is linked by a network of pipelines and waterways which
connect large refining areas in the Gulf Coast, the upper Midwest and the Northeast. While most
regions of the country receive some of their gasoline from local refineries, the source of marginal
supply varies across the U.S. and may change during the year. The Gulf Coast of the U.S.
5

---

(refineries in Texas and Louisiana) produces much more gasoline than it consumes, and ships
gasoline to the Midwest and East Coast. The eastern region of the U.S. is a net importer of
gasoline, with marginal supply coming from the Gulf via pipeline and from Canada, Europe and
the Caribbean via ports around New York City. Most of the gasoline consumed in the upper
Midwest, e.g., Illinois or Minnesota, is refined locally, but the region receives marginal supply
from the Gulf.
Not only does the method of supply vary by geography, but vertical integration among
levels of the petroleum industry- crude exploration, refining, wholesaling and marketing- vary by
firm and geography as well. Some firms, such as Exxon-Mobil, are vertically integrated from the
exploration and production of crude oil through refining, wholesaling and marketing. Other
firms, such as Tesoro, concentrate on refining and marketing, and other firms concentrate on
simply refining, such as Koch, or marketing, such as Sheetz or Racetrac.
Further complicating the vertical market structure in the industry, there are also different
vertical relationships between the wholesale and retail levels of the industry.11 A branded
gasoline station, e.g. Exxon or Shell, may be owned and operated by an oil company (company
op), owned by the oil company and leased to an independent operator (lessee dealer), or owned
and operated by an independent operator (open dealer). It is important to note that each of these
retail/wholesale vertical relationships results in a potentially different wholesale price. The
company owned and operated station pays an unobserved transfer price for gasoline, the lessee
dealer typically pays a dealer tank wagon price which can vary by station and which is difficult to
observe, and the open dealer typically pays the rack price plus delivery and possibly a markup to
the delivery firm which is somewhat observable. The percentage of branded stations of each
vertical type varies dramatically by brand and geography.12 While this is a very abbreviated
summary of some important facts about the petroleum industry, it serves to outline those factors
11The vertical market structure is impacted in a number of states by divorcement regulations,
restrictions on petroleum companies owning gasoline stations. See, Vita (2000) and Blass and
Carlton (2001) for a description, and the estimated economic impact, of divorcement.
12For a more detailed description of the wholesale gasoline markets and DTW and rack
pricing see Borenstein and Shepard (1994).
6

---

that affect the wholesale and retail price of gasoline. In particular, given the different
relationships between suppliers and retailers, it is important to understand the vertical structure of
local markets and the pricing at different levels when examining the potential effects of any
consummated merger.
B. The Marathon-Ashland Joint Venture
The MAP joint venture affected both the wholesale and retail distribution of gasoline in
the Midwest. This was one of the first major transactions in the most recent era of petroleum
mergers and it caused significant changes in concentration. Many subsequent mergers did not
cause important changes in concentration because of substantial divestitures required by
regulators. In May of 1997, USX-Marathon and Ashland Inc, announced that they planned to
combine their downstream operations into a refining and marketing company. The joint venture
included 930,000 barrels per day of refining capacity at seven refineries, and 5,400 retail outlets.
The joint venture was owned 62 percent by Marathon and 38 percent by Ashland. The refineries
from Marathon were in Garyville, Louisiana, Robinson, Illinois, Texas City, Texas, and Detroit,
Michigan. The refineries from Ashland were in Catlettsburg, Kentucky, St. Paul, Minnesota, and
Canton, Ohio.
In addition, Marathon contributed 51 terminals and Ashland contributed 33 terminals.
Marathon contributed 3,980 retail outlets in 17 states and Ashland contributed 1,420 retail outlets
in 11 states. The combined firm has a retail presence in 20 states. Marathon also contributed
5,000 miles of pipelines to the joint venture (Platt’s Oilgram News, May 16, 1997). Marathon
and Ashland signed the definitive joint venture agreement in December 1997, and consummated
the joint venture on January 1, 1998.
Marathon and Ashland acknowledged that the Federal Trade Commission (FTC) was
reviewing the transaction and that they had received a second request for information. A
Prudential Securities report in October of 1997 stated that Ashland had completed its FTC
document request and anticipated approval in “four to six weeks.” A December 1997 news story
commented that the FTC had signed off on the merger and did not mandate any divestiture
(Platt’s Oilgram News, December 15, 1997). The FTC does not usually publicly acknowledge
7

---

that it is conducting a particular merger investigation nor does it issue statements about closed
investigations. There were no FTC announcements concerning the MAP joint venture.
There were three levels of the petroleum industry where anticompetitive effects were
possible as a result of this merger: refining and the wholesale and retail distribution of gasoline in
the area. Five of the joint venture’s refineries were located in the Midwest, and two were located
in the Gulf Coast (where the market was not concentrated). Gasoline consumed in the Midwest
comes from refineries in the area and from pipelines and barges that shipped gasoline from the
Gulf Coast to the Midwest. While Marathon and Ashland competed throughout the Midwest,
given their respective refinery locations, Ashland had a much larger presence in the eastern and
northwestern portions of the Midwest and Marathon had a larger market presence in the central
portion of Midwest.
At the wholesale level, Marathon or Ashland were among the top four suppliers in nine
states in 1996 and 1997, according to Department of Energy, Energy Information Administration
data. These nine states were Kentucky, Ohio, West Virginia, Indiana, Illinois, Michigan,
Minnesota, South Carolina, and North Dakota. There was only one state, Kentucky, where both
Marathon and Ashland were among the top four wholesale suppliers. The wholesale HHI, sum of
squared market shares, for Kentucky (the narrowest region for which we can calculate an HHI
with publicly available data) increased by about 800 points from 1477 in 1997 to 2263 in
1998.13,14
Figure 1 shows the Marathon and Ashland refineries in the central Midwest as well as the
other refineries and pipelines in the area where the largest wholesale overlap occurred. Not
surprisingly, like the wholesale overlaps, the highest retail market shares from the joint venture
were in Kentucky (26%), Ohio (26%) and Indiana (27%). In Indiana almost the entire market
share was from Marathon. In the other two states, Kentucky and Ohio, the HHI increased by
13Department of Energy, Energy Information Administration, Petroleum Supply Annual,
1997-1998.
14Implicit in this calculation is the assumption that the state of Kentucky is a market. For the
region we study, Louisville, the number of firms posting a rack price for conventional gasoline
went from eight to seven and for RFG from four to three.
8

---

over 250 points to 1500 to 1600 range. These retail market shares are based on sales of gasoline
by brand. Since some of the branded stations are independently owned and could switch brands,
these market shares overstate the concentration.
Given the concentration measures, both at wholesale and at retail, and the location of the
refineries, Kentucky appears to be the area most likely place to experience an anticompetitive
effect from the MAP joint venture. We are agnostic as to whether the possible anticompetitive
problem may be at the refining, wholesaling or retailing level.
Within Kentucky, we concentrate our analysis on Louisville for a number of reasons.
First, it is the largest major metropolitan area in Kentucky.15 Second, Louisville is directly
between the two refineries (Robinson and Cattletsburg), see Figure 1, and both Marathon and
Ashland had a large retail and wholesale presence. Third, Louisville uses RFG, which makes
arbitrage from nearby regions (that use conventional gasoline) more difficult.16 In other words,
while in most parts of the Midwest one or the other of the firms had a significant presence pre-
joint venture, Louisville is the place where they most directly overlapped and where each had a
major presence.
III. Literature Review
This section reviews the methodology used in studies that use pre- and post-merger
pricing data to estimate merger effects and the results of papers that examine the effects of
petroleum mergers. While many papers discuss merger effects, there is not a large literature on
the estimated price effects of mergers outside of historically regulated industries, e.g., banking,
15See Figure II for a map of the Louisville MSA and the gasoline station locations.
16We also analyzed conventional gasoline prices at the Louisville rack and at retail in the area
surrounding the RFG area in Louisville relative to the control cities. There was no change in the
price of conventional gasoline at the Louisville rack or in the surrounding retail areas. Figure VII
shows the price of conventional gasoline at the Louisville rack relative the Chicago rack. There is
no change in the price of conventional gasoline.
9

---

health care, and airlines.17 Most merger event studies that examine product prices before and after
a merger use one of three types of reduced form regressions.
In the first type of regression (see Barton and Sherman (1984) and Kim and Singal
(1993)), the price of the product affected by the merger is compared to the price of a product that
faces similar demand and cost conditions but is unaffected by the merger. Specifically, the price
of the product of the merged firm is regressed on the price of the control product(s) and controls
for time or seasonality and a merger dummy variable. In the second type of regression (see
Schumann et al., (1992) and (1997)), the price of the merged firm’s product (or market price) is
regressed on demand and supply/cost shifters plus a merger dummy. A third approach combines
elements of both approaches. In their study of a hospital merger, Vita and Sacher (2001)
examine the price of the merged firm relative to the price of a control group of firms unaffected
by the merger that should be affected by the same demand and supply factors. They then
regressed these relative prices on relative demand shifters, relative cost shifters, and the merger
event to gauge the effect of the merger.
The second approach, which relies completely on demand and supply variables, is
problematic in this case. There are few variables that are available on a weekly or monthly basis
at the city level to help explain wholesale or retail gasoline price variation. The most promising
approach for gasoline markets is the control city approach with possibly additional variables to
check for marginal supply changes. These marginal supply changes are likely seasonal, caused
by peak capacity of pipelines or refineries.
One published paper and three recent working papers have estimated merger effects in
gasoline markets by either calculating the actual effect of consummated petroleum mergers on
gasoline prices or simulating the projected price effects from proposed mergers that were not
actually consummated.18 These papers are representative of the wide range of the methodologies
17For a review of the literature on the multitude of methodologies used in examining the
effects of mergers, including those papers that attempt to directly estimate the price effects see,
Pautler (2003).
18In addition to the recent working papers discussed in the text, a government report by the
U.S. General Accounting Office (GAO, 1986) examined gasoline prices from the time period
10

---

Document Outline

• MAP_JIE_Submission.pdf
  • Page 1
  • Page 2
  • Page 3
  • Page 4
  • Page 5
  • Page 6
  • Page 7
  • Page 8
  • Page 9
  • Page 10
  • Page 11
  • Page 12
  • Page 13
  • Page 14
  • Page 15
  • Page 16
  • Page 17
  • Page 18
  • Page 19
  • Page 20
  • Page 21
  • Page 22
  • Page 23
  • Page 24
  • Page 25
  • Page 26
  • Page 27
  • Page 28
  • Page 29
  • Page 30
  • Page 31
  • Page 32
  • Page 33
  • Page 34
  • Page 35

---