International Aviation: Airline Alliances Produce Benefits, but Effect on
Competition Is Uncertain (Chapter Report, 04/06/95, GAO/RCED-95-99).
For U.S. airlines, growth in the international sector in recent years
has far outpaced domestic growth. Between 1987 and 1993, the number of
passengers traveling on U.S. airlines between the United States and
foreign destinations increased by 47 percent, while domestic traffic
increased by only six percent. The airlines' ability to respond to this
demand is limited, however, by intergovernmental restrictions and cost
constraints. As a result, the airlines have increasingly entered into
alliances with foreign airlines, rather than starting new service to
additional foreign cities. Likewise, foreign carriers have entered into
alliances with U.S. airlines to increase their access to the U.S.
market. This report (1) examines the effects of marketing alliances
between U.S. and foreign airlines on airlines' traffic flows and
revenues and on consumers and (2) identifies key issues concerning such
alliances that need to be addressed by the Transportation Department.
--------------------------- Indexing Terms -----------------------------
REPORTNUM: RCED-95-99
TITLE: International Aviation: Airline Alliances Produce Benefits,
but Effect on Competition Is Uncertain
DATE: 04/06/95
SUBJECT: Airline regulation
Airline industry
Computerized information systems
Airports
Management information systems
Antitrust law
Competition
Commercial aviation
Competition limitation
Regulatory agencies
IDENTIFIER: Heathrow Airport (United Kingdom)
Atlanta-Hartsfield International Airport (Atlanta, GA)
Memphis International Airport (Memphis, TN)
Dulles International Airport (VA)
Cincinnati (OH)
United Kingdom
Austria
Belgium
Iceland
Switzerland
Canada
Netherlands
Japan
France
European Union
**************************************************************************
* This file contains an ASCII representation of the text of a GAO *
* report. Delineations within the text indicating chapter titles, *
* headings, and bullets are preserved. Major divisions and subdivisions *
* of the text, such as Chapters, Sections, and Appendixes, are *
* identified by double and single lines. The numbers on the right end *
* of these lines indicate the position of each of the subsections in the *
* document outline. These numbers do NOT correspond with the page *
* numbers of the printed product. *
* *
* 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. *
* *
* A printed copy of this report may be obtained from the GAO Document *
* Distribution Facility by calling (202) 512-6000, by faxing your *
* request to (301) 258-4066, or by writing to P.O. Box 6015, *
* Gaithersburg, MD 20884-6015. We are unable to accept electronic orders *
* for printed documents at this time. *
**************************************************************************
Cover
================================================================ COVER
Report to Congressional Requesters
April 1995
INTERNATIONAL AVIATION - AIRLINE
ALLIANCES PRODUCE BENEFITS, BUT
EFFECT ON COMPETITION IS UNCERTAIN
GAO/RCED-95-99
International Airline Code-Sharing
Abbreviations
=============================================================== ABBREV
ASTA - American Society of Travel Agents
CRS - computer reservation system
DOT - Department of Transportation
EU - European Union
GAO - General Accounting Office
GRA - Gellman Research Associates, Inc.
Letter
=============================================================== LETTER
B-259645
April 6, 1995
The Honorable Larry Pressler
Chairman
The Honorable Ernest F. Hollings
Ranking Minority Member
Committee on Commerce, Science,
and Transportation
United States Senate
The Honorable John McCain
Chairman
The Honorable Wendell H. Ford
Ranking Minority Member
Subcommittee on Aviation
Committee on Commerce, Science,
and Transportation
United States Senate
This report (1) examines the effects of marketing alliances between
U.S. and foreign airlines on airlines' traffic flows and revenues
and on consumers and (2) identifies the key issues concerning such
alliances that need to be addressed by the Department of
Transportation (DOT).
As arranged with your offices, unless you publicly announce its
contents earlier, we plan no further distribution of this report
until 30 days after the date of this letter. We will then send
copies to the Secretary of Transportation; the Secretary of State;
the Director, Office of Management and Budget; and other interested
parties. We will also send copies to others upon request.
This work was performed under my direction, and I can be reached at
(202) 512-2834. Other major contributors are listed in appendix III.
Sincerely yours,
Kenneth M. Mead
Director, Transportation Issues
EXECUTIVE SUMMARY
============================================================ Chapter 0
PURPOSE
---------------------------------------------------------- Chapter 0:1
For U.S. airlines, growth in the international sector in recent
years has far outpaced growth domestically. Between 1987 and 1993,
the number of passengers traveling on U.S. airlines between the
United States and foreign destinations increased by 47 percent, while
domestic traffic increased by only 6 percent. The airlines' ability
to respond to this demand is limited, however, by intergovernmental
restrictions and cost constraints. As a result, the airlines have
increasingly entered into alliances with foreign airlines rather than
starting new service to additional foreign cities. Likewise, foreign
carriers have entered into alliances with U.S. airlines to obtain
increased access to the U.S. market.
The Chairmen and Ranking Minority Members of the Senate Committee on
Commerce, Science, and Transportation and its Subcommittee on
Aviation asked GAO to determine the (1) extent to which U.S. and
foreign airlines participating in alliances benefit from those
alliances in terms of added passengers and revenues and (2) effect
that alliances have on other U.S. airlines and consumers. They also
asked GAO to identify and examine key issues, if any, pertaining to
alliances that the Department of Transportation (DOT) did not address
in its November 1994 policy statement on international aviation or
its recently proposed rules aimed at ensuring that consumers are
notified, before purchasing a ticket, as to which airline partner
will actually be operating the flight.
BACKGROUND
---------------------------------------------------------- Chapter 0:2
International aviation is governed by bilateral agreements between
countries that often limit the number of cities that can be served
and airlines that can serve them. Historically, U.S. and foreign
airlines have entered into agreements to coordinate schedules and
ensure the efficient transfer of connecting passengers and baggage.
In part because of bilateral restrictions, however, they have
increasingly entered into closer partnerships called alliances. The
alliances involve one airline using its two-character designator code
(e.g., "NW" for Northwest Airlines) to advertise a flight as its own
in travel agents' computer reservation systems, even though the
flight is actually operated by its partner. Such "code-sharing"
allows airlines to connect traffic from foreign cities, which they do
not fly to, with their flights. Because one airline lists another
airline's flight as its own, that flight is listed twice in computer
reservation systems (once under each airline's code) and more times
if connections are involved.
DOT requires that code-sharing alliances between U.S. and foreign
airlines be approved by the agency and periodically reapproved,
usually annually. Between 1992 and 1994, the number of these
alliances more than tripled, from 19 to 61. In 1992, DOT granted the
alliance between Northwest and KLM Royal Dutch Airlines immunity from
U.S. antitrust laws consistent with the accord with the Netherlands
that eliminated bilateral restrictions on air travel between the two
countries. In 1994, the agency issued a policy statement that
supported code-sharing alliances. It also proposed rules that would
require airlines and travel agents to notify customers, before
booking flights, which airline will be operating a code-share flight
and provide a written notice with the ticket naming the operating
airline.
RESULTS IN BRIEF
---------------------------------------------------------- Chapter 0:3
Alliances between U.S. and foreign airlines have in several cases
generated large gains for partners in terms of passengers and
revenues. In general, the more global the scope of the code-sharing
arrangement and the greater the degree of integration achieved by the
airlines in scheduling, operations, and frequent flyer programs, the
larger the benefits are for partners. Conversely, the impact on
other U.S. airlines in terms of reduced ridership and revenues
depends on an alliance's geographic scope and integration, the other
airlines' competitive responses, and the extent to which competition
between that alliance and the other airlines stimulates new traffic.
Although consumers benefit from the conveniences--such as decreased
layover times--that alliances provide, insufficient data exist to
determine (1) what effect alliances have had on fares in the short
term and (2) whether alliances will reduce or increase competition in
the long term and thereby lead to higher or lower fares.
Although DOT's policy statement notes the need to monitor the effects
of alliances on competition and the international competitiveness of
U.S. airlines, the agency has not required U.S. and foreign
airlines to report sufficient data to fully monitor these effects.
DOT also has not determined, in light of the Northwest/KLM
experience, whether antitrust immunity should be potentially
available for other alliances in markets that allow for significantly
increased access for U.S. airlines. Finally, although DOT has
proposed rules to ensure that consumers are told which airline
partner will actually operate a code-share flight, neither its
current regulations nor its proposed rules limit how often the same
flight can be listed in computer reservation systems. Multiple
listings of the same flight give airlines in an alliance a
competitive advantage. Recognizing this impact, the European Union
in 1993 limited to two the number of times a flight can be listed.
PRINCIPAL FINDINGS
---------------------------------------------------------- Chapter 0:4
ALLIANCES OFTEN PRODUCE
BENEFITS, BUT IMPACT ON
FARES IS UNCERTAIN
-------------------------------------------------------- Chapter 0:4.1
GAO's analysis of U.S. and foreign airlines' data indicates that
strategic alliances, which involve code-sharing on a vast number of
routes so as to strategically link airlines' flight networks, can
produce large traffic gains for partners. The three strategic
alliances entered into to date--Northwest/KLM (formed in 1992),
USAir/British Airways (1993), and United/Lufthansa (1994)--are
producing large increases in the number of passengers traveling on
these airlines because their alliances involve (1) code-sharing on
numerous routes covering a wide geographical area and (2) a great
degree of operating and marketing integration. Northwest and KLM
data show that their annual ridership has increased by about 350,000
as a result of their alliance, producing an increase in their
combined transatlantic market share from 7 percent in 1991 to 11.5
percent in 1994. Alliances that involve code-sharing on a more
limited number of routes, usually in one geographic region or between
a few cities, have also resulted in increased ridership in many
cases, though at much lower levels than the three strategic
alliances. Although the traffic gains achieved by airlines through
alliances have come largely at the expense of other U.S. and foreign
airlines, at least some of the gains have come from new traffic
stimulated by increased competition among alliances and between
alliances and other airlines, according to most U.S. and foreign
airline representatives and DOT officials that GAO interviewed.
Strategic alliances produce the largest revenue gains for partners.
On the basis of its analysis of Northwest's data, GAO estimates that
the alliance with KLM produced between $125 million and $175 million
in revenues for Northwest in 1994 (about one-third of its
transatlantic passenger revenues). By contrast, American Airlines'
alliance with South African Airways, which involves only flights
between New York and Johannesburg, generated less than 1 percent of
American's transatlantic revenues in 1994. Whether or not the U.S.
airline industry gains as a result of an alliance depends on the
specifics of each deal. Because they code-share on each other's
flights and split the revenues accordingly, Northwest and KLM gain
revenues roughly evenly, largely at the expense of both U.S. and
foreign airlines. Alternatively, British Airways gains revenues
primarily at the expense of U.S. airlines because its arrangement
with USAir allows only for it to code-share on USAir domestic flights
and keep most of the revenues. However, this effect must be
considered in the context of British Airways' $400 million investment
in USAir in 1993, which was of critical importance to the financially
struggling U.S. airline.
Alliances produce several benefits for consumers. For example, close
schedule coordination between partners often produces shorter layover
times between connections. DOT officials believe that competition
among alliances and between alliances and other airlines is resulting
in lower fares, thereby stimulating new traffic. However,
insufficient data exist to determine the effect of alliances on
fares.
BY RESOLVING KEY ISSUES, DOT
CAN BETTER ADDRESS IMPACTS
ON COMPETITION
-------------------------------------------------------- Chapter 0:4.2
Although DOT's policy statement held that alliances will likely
increase competition in the long term, the agency noted that it
needed to monitor them for potential harmful effects that could
result if competition decreased. Such effects could include
consumers facing higher fares if (1) strategic alliances lead to a
marketplace dominated by a handful of "mega-carriers" that are not
effectively competing with each other or are preventing other U.S.
carriers from entering international markets or (2) foreign countries
whose national airlines are in alliances fail to increase access to
their markets for other U.S. airlines.
To monitor developing trends, DOT created an economic analysis unit
in November 1994. Previously, DOT had approved and reapproved nearly
all code-sharing arrangements with little analysis. The unit's
efforts to monitor the effects of alliances will be hindered,
however, because the data reported by U.S. airlines to DOT from a
sample of their tickets do not identify (1) passengers who traveled
on code-share flights and (2) in some cases, which airline actually
operated a code-share flight. Likewise, because DOT does not collect
detailed data from foreign airlines' tickets, it lacks key data on
thousands of passengers traveling to and from the United States on
code-share flights. Currently, foreign airlines are required to
report data to DOT only on their overall traffic between gateway
cities (e.g., New York-London). According to DOT analysts, such data
are of limited use in analyzing the effects of alliances because they
do not, among other things, identify code-share traffic or provide
information on fares.
Finally, DOT's rules do not limit the number of times a flight can be
listed on computer reservation systems. Computer reservation systems
often list the same code-share flight option several times. For
example, GAO found a Lufthansa flight from Berlin to Frankfurt that
connects with a United flight from Frankfurt to Chicago listed as (1)
Lufthansa throughout, (2) United throughout, and (3) Lufthansa to
Frankfurt and United to Chicago. GAO also found that such listings
consumed much of the computer reservation system's first display
screen in nearly 20 percent of the cases it reviewed, thereby
"crowding out" competing flight options to lower screens. This
situation limits competition because industry studies have shown that
travel agents--who are responsible for 80 percent of all airline
bookings--book flights that are listed on the computer reservation
system's first screen as often as 90 percent of the time.
RECOMMENDATIONS
---------------------------------------------------------- Chapter 0:5
GAO recommends that the Secretary of Transportation (1) require that
U.S. airlines, as part of their regular reporting of traffic data to
DOT, identify which passengers traveled on code-share flights and
that they take steps to ensure that they report which airlines
actually operated those flights; (2) require, either by regulation or
by making it a condition of approving code-sharing alliances, that
foreign airlines involved in such alliances with U.S. airlines
report data on their code-share traffic to DOT; (3) direct the
agency's new economic unit to analyze DOT's existing data and the
data obtained as described above to determine if the U.S. airline
industry or consumers have been negatively affected before
reapproving all strategic alliances and any other alliance that the
Secretary deems significant; (4) examine, in light of the
Northwest/KLM experience, whether immunity from U.S. antitrust laws
should be potentially available for other alliances in markets that
allow for significantly increased access for U.S. airlines; and (5)
prohibit more than two listings of the same code-share flight in
computer reservation systems.
AGENCY COMMENTS
---------------------------------------------------------- Chapter 0:6
GAO discussed a draft of this report with senior DOT and State
Department officials, including DOT's Acting Assistant Secretary for
Aviation and International Affairs. They emphasized that airlines
primarily enter into code-sharing alliances as a result of market
forces stemming from the airlines' efforts to efficiently expand
their international operations and that such alliances produce
benefits for partners and consumers. Likewise, they believe that
alliances increase competition. These officials agreed, however,
that DOT needed more detailed data from U.S. airlines and additional
data from foreign airlines to better track alliances' long-term
impacts on competition. They further stated that in some cases
code-sharing rights are exchanged in bilateral agreements and that
because of resource constraints, it would not be practicable for the
new economic unit to analyze smaller, noncontroversial arrangements
before DOT reapproves them. On the basis of their comments, GAO
revised its proposed recommendation concerning the new unit.
DOT officials agreed that the agency has not determined whether
immunity should be potentially available for other alliances in
markets that allow for significantly increased access for U.S.
airlines. Noting that this issue is sensitive and that DOT is in
negotiations with several countries, they declined to comment
further. Finally, they noted that American Airlines and TWA,
supported by the American Society of Travel Agents, have petitioned
DOT to pass rules limiting how often a flight can be listed in
computer reservation systems. They stated that DOT is analyzing the
petitions and therefore declined to comment on GAO's recommendation.
As requested, GAO did not obtain written comments on a draft of this
report.
INTRODUCTION
============================================================ Chapter 1
The number of passengers traveling between the United States and
foreign destinations has increased dramatically since 1980, and the
rate of growth in the international sector for U.S. airlines has far
exceeded the rate for domestic air travel over the last several
years. Unlike the domestic market, however, international air travel
is heavily regulated. Airlines are often limited in the routes they
can fly, how often they can serve those routes, and the fares they
can charge. Because of these restrictions and cost constraints, U.S.
airlines have increasingly entered into alliances with foreign
airlines rather than starting up new service. Likewise, foreign
carriers have entered into such alliances to obtain increased access
to the U.S. market. The Department of Transportation (DOT) requires
that agreements between U.S. and foreign airlines be approved by the
agency when one airline markets another airline's flight as its own.
Under U.S. law, the Secretary of Transportation has the authority,
in certain circumstances, to grant immunity from U.S. antitrust laws
to an agreement in foreign air transportation.
INTERNATIONAL SECTOR IS A KEY
GROWTH AREA FOR U.S. AIRLINES
---------------------------------------------------------- Chapter 1:1
Largely because of the growth of international tourism and the
globalization of economic activity, the demand for air travel between
the United States and the rest of the world has grown rapidly since
1980. Total passenger traffic between the United States and foreign
destinations increased by 134 percent from 1980 through 1993--from
39.5 million passengers to 92.6 million. The International Air
Transport Association estimates that this number will increase to 226
million passengers by 2010.
U.S. airlines have captured an increasing share of this growing
sector. In 1993, U.S. airlines carried 54 percent of the passengers
traveling between the United States and other nations, as compared to
49 percent in 1980. In addition, U.S. scheduled airlines'
international traffic has grown at a faster pace than their domestic
traffic in recent years (fig. 1.1).
Figure 1.1: Growth in U.S.
Scheduled Airlines'
International and Domestic
Passenger Traffic, 1987-93
(See figure in printed
edition.)
Source: Air Transport Association.
UNLIKE DOMESTIC TRAVEL,
INTERNATIONAL AIR TRAVEL IS
HEAVILY REGULATED
---------------------------------------------------------- Chapter 1:2
Despite increasing demand, the international aviation market--unlike
the U.S. domestic market--remains heavily regulated. Under a
framework established by the United States and 51 other nations in
1944, international air travel is largely governed by bilateral
agreements. Two countries negotiate the air services between them
and award their airlines the right to offer those services. In
general, bilateral agreements define (1) which routes can be served
between the countries and to third countries; (2) whether the fares
airlines charge need government approval; and in some cases (3) how
frequently flights can be offered and (4) how many airlines from each
country can fly the routes.
As of February 1995, the United States was party to 72 bilateral
agreements. These accords often greatly restrict U.S. airlines.
The U.S. accord with the United Kingdom, for example, specifies that
only two U.S. airlines--currently American Airlines and United
Airlines--can serve London's Heathrow Airport, which is a key gateway
for traffic traveling between the United States and both Europe and
the Middle East. Many agreements also limit U.S. airlines' ability
to change their fares and the number of flights that can be operated.
These restrictions constrain the airlines' ability to respond to
market demand and thereby prevent the public from obtaining better
airline service. By the same token, these agreements generally
restrict the extent to which foreign airlines can serve U.S.
destinations. For example, the U.S. agreement with the Philippines
specifies that that country's airlines can serve only eight cities in
the United States.
The heavily regulated international marketplace contrasts greatly
with the deregulated U.S. domestic market. As a result of the
Airline Deregulation Act of 1978, U.S. airlines can generally choose
which routes they fly within the United States, the frequency of
flights, and the fares charged. Since the late 1970s, DOT and its
predecessor, the Civil Aeronautics Board, have attempted to "export"
this deregulated environment by working with foreign governments to
eliminate bilateral restrictions. The agencies have achieved mixed
results. For example, the United States reached agreements with
Austria, Belgium, Canada, Iceland, Israel, Jamaica, Korea,
Luxembourg, the Netherlands, Singapore, and Switzerland that reduce
or eliminate the restrictions. However, many countries, including
the United Kingdom and Japan, have maintained--and in some cases
added--extensive limitations on U.S. airlines' access to and beyond
their markets. Others, such as France and Thailand, have renounced
their accords with the United States. These countries have taken
such actions principally to protect their national carriers from
competition with U.S. airlines, which often have much lower
operating costs.\1 A study by the European Union (EU), for example,
found that the operating costs of major European airlines were about
50 percent higher than the operating costs of major U.S. airlines in
1992.
--------------------
\1 We recently reported that U.S. airlines serving key European and
Pacific Rim airports often face--in addition to bilateral
restrictions--obstacles, such as inadequate terminal facilities, that
foreign airlines operating in the United States experience to a much
lesser extent. See International Aviation: DOT Needs More
Information to Address U.S. Airlines' Problems in Doing Business
Abroad (GAO/RCED-95-24, Nov. 29, 1994).
U.S. AND FOREIGN AIRLINES HAVE
INCREASINGLY ENTERED INTO
ALLIANCES BECAUSE OF COST
CONSTRAINTS AND BILATERAL
RESTRICTIONS
---------------------------------------------------------- Chapter 1:3
Historically, U.S. and foreign airlines have entered into agreements
to coordinate schedules and other activities. In the last few years,
however, U.S. airlines have increasingly entered into more extensive
partnerships with foreign airlines, often called alliances. The
alliances generally involve U.S. airlines marketing foreign
airlines' flights as their own rather than serving these destinations
directly. U.S. airlines have entered such alliances primarily
because it is uneconomical for them to serve many foreign cities with
their own aircraft. In addition, bilateral restrictions limit their
ability to serve many foreign markets, thus making alliances more
attractive. Alliances generally allow a U.S. airline to connect
passengers from foreign cities with its flights. Likewise, foreign
airlines have entered into such alliances to connect passengers from
U.S. cities, which they do not fly to, with their flights. These
alliances require DOT's approval.
U.S. AND FOREIGN AIRLINES
COMMONLY ENTER INTO
AGREEMENTS TO COORDINATE
ACTIVITIES
-------------------------------------------------------- Chapter 1:3.1
U.S. and foreign airlines coordinate schedules and attempt to ensure
the efficient transfer of connecting passengers, baggage, and cargo
through standard agreements, commonly referred to as "interline
agreements." Interline agreements provide for the mutual acceptance
by the participating airlines of passenger tickets, baggage checks,
and cargo waybills, as well as establish uniform procedures in these
areas. These agreements are common, and DOT has traditionally not
required that they be filed for approval. In addition, airlines are
also increasingly entering into simple marketing arrangements, such
as linking frequent flyer programs or sharing airport facilities,
which are designed to enhance the benefits of interline agreements
for passengers and the participating airlines. As with interline
agreements, DOT has not traditionally required that simple marketing
arrangements be filed for approval.
THE NUMBER OF MARKETING
ALLIANCES THAT INVOLVE
"CODE- SHARING" HAVE TRIPLED
SINCE 1992
-------------------------------------------------------- Chapter 1:3.2
Over the last few years, U.S. and foreign airlines have increasingly
entered into alliances that are more extensive than interline
agreements and simple marketing arrangements. These alliances
involve "code-sharing"--the practice of two airlines each placing its
two-character designator code (e.g., "NW" for Northwest Airlines) on
the same flight when listing that flight in computer reservation
systems (CRS) used by travel agents to book flights. Airline
designator codes are assigned to individual airlines by the
International Air Transport Association and are used in reservations,
schedules, and ticketing.
Code-sharing occurs when an airline, by agreement, uses its
designator code to market flights operated by another carrier as its
own. Code-sharing is most often used to show connecting flights as
occurring on one airline. In displaying connecting flights as being
on one airline, airlines are listing them as "on-line" (same airline)
rather than "interline" (two airlines). In doing so, they are
responding to consumers' preferences for booking connecting flights
on the same airline. Prior studies by DOT and others have shown that
consumers generally prefer on-line over interline connections. DOT
found that consumers generally believe that same carrier connections
(1) involve shorter distances between gates in the terminal, thus
making transfers to connecting flights easier, and (2) are less
likely to result in lost luggage. In addition, airlines prefer to
offer connecting flights as on-line because some CRSs list on-line
flights before interline connections, and travel agents tend to book
customers on flights listed higher on the CRS screen.
In 1987, DOT began requiring that code-sharing arrangements between
U.S. and foreign airlines be filed with the agency for approval.
Between January 1, 1992, and December 31, 1994, the number of
code-sharing alliances approved by DOT more than tripled from 19 to
61. (App. I lists the 61 alliances and the year in which DOT
approved them.) DOT also requires that code-sharing arrangements be
reapproved after a specified period of time, usually annually. In
November 1994, the agency issued the U.S. International Aviation
Policy Statement, in which it supported the practice of code-sharing.
In December 1994, a DOT contractor, Gellman Research Associates, Inc.
(GRA), issued a report that also generally supported code-sharing
alliances.\2
Finally, to ensure that consumers know the nature of services they
are purchasing, DOT in 1994 proposed rules to strengthen current
requirements that airlines and travel agents, before making
reservations for passengers on a code-share flight, tell customers
which airline will actually be operating the flight. The proposed
rules would require travel agents in the United States and ticket
agents for U.S. and foreign airlines to, among other things, provide
written notice at the time of sale naming the airline that will
operate the flight for tickets sold in the United States.\3
--------------------
\2 A Study of International Airline Code Sharing, Gellman Research
Associates, Inc., Dec. 1994.
\3 As of February 1995, DOT was reviewing the public comments
received on its proposed rules.
U.S. AIRLINES SEEK ACCESS
TO MORE FOREIGN DESTINATIONS
THROUGH CODE-SHARING
-------------------------------------------------------- Chapter 1:3.3
U.S. airlines have generally entered into code-sharing alliances
with foreign airlines to "feed" their international flights with
passengers traveling to and from foreign cities that the U.S.
airlines do not serve with their own aircraft. The airlines often do
not fly to these cities because the cost of providing nonstop or
direct service is too high relative to passenger demand. Bilateral
restrictions also sometimes limit their ability to expand their
international service. In general, foreign governments have been
more willing to grant U.S. airlines authority for code-sharing than
to remove restrictions on U.S. airlines' ability to directly serve
their markets.
As shown in figures 1.2 and 1.3, Northwest Airlines' alliance with
KLM increases Northwest's access to Europe and the Middle East by
allowing it to market services through CRSs and direct advertisements
to over 30 cities in Europe and the Middle East, when it actually
flies to only 4 cities. By listing KLM's flights between Amsterdam
and 30 cities as its own and connecting these flights with
Northwest's flights between the United States and Amsterdam,
Northwest can advertise that it serves these 30 cities in addition to
the 4 cities to which it actually flies. Thus, Northwest can more
effectively attract passengers who want to travel between the 30
cities and the United States than through a standard interline
agreement.
Figure 1.2: Northwest's
Flights to Europe and the
Middle East Prior to Alliance
With KLM
(See figure in printed
edition.)
Source: GAO's illustration of information provided by Northwest.
Figure 1.3: Northwest's
Flights to Europe and the
Middle East and Code-Share
Flights Operated by KLM as a
Result of Their Alliance
(See figure in printed
edition.)
Note: Northwest flies passengers between the United States (via
Boston and Minneapolis hubs) and Amsterdam, and KLM flies passengers
between Amsterdam and the other cities. However, through
code-sharing, Northwest is able to market in CRSs service between the
United States and these foreign destinations. Finally, Northwest
also markets KLM's flights between Detroit and Amsterdam as its own.
Source: GAO's illustration of information provided by Northwest.
FOREIGN AIRLINES SEEK
INCREASED ACCESS TO U.S.
DOMESTIC MARKET THROUGH
CODE-SHARING
-------------------------------------------------------- Chapter 1:3.4
Similarly, foreign airlines have generally entered into code-sharing
alliances to "feed" their international flights with passengers
traveling to and from U.S. cities that those airlines do not serve
with their own aircraft. Through its alliance with USAir, British
Airways markets service to 52 U.S. cities that it actually does not
fly to. By listing USAir's flights to and from these cities as its
own, British Airways can more effectively feed its flights across the
Atlantic with passengers who want to travel between those cities and
London (or points beyond London) than it could under interline
agreements with U.S. airlines.
Figure 1.4: British Airways'
Flights Between London and the
United States Prior to Alliance
With USAir
(See figure in printed
edition.)
Source: GAO's illustration of information provided by British
Airways.
Figure 1.5: British Airways'
Flights Between London and the
United States and Code-Share
Flights Operated by USAir
Within the United States as a
Result of Their Alliance
(See figure in printed
edition.)
Source: GAO's illustration of information provided by USAir.
GEOGRAPHIC SCOPE OF
CODE-SHARING ALLIANCES
VARIES AND IS OFTEN
COMPLEMENTED BY OTHER TYPES
OF INTEGRATION
-------------------------------------------------------- Chapter 1:3.5
Code-sharing alliances between U.S. and foreign airlines vary in
their scope. Three of the 61 alliances--Northwest/KLM, USAir/British
Airways, and United/Lufthansa--are "strategic" alliances in that they
involve code-sharing on a vast number of routes so as to
strategically link both airlines' flight networks. Eight "regional"
alliances involve code-sharing between airlines on several routes to
and from a specific region. United's alliance with Ansett Australia,
for example, allows it to code-share on Ansett flights within
Australia and connect those flights with United's flights between
Australia and the United States. Finally, 50 alliances involve
code-sharing on flights between a small number of cities (referred to
in this report as "point-specific" alliances). These alliances often
involve one airline's purchasing blocks of seats on another airline's
flights and then reselling them (referred to as a blocked-space
agreement).\4 (App. II lists the strategic, regional, and
point-specific alliances.)
Code-sharing alliances often involve additional cooperation between
the airlines, ranging from schedule coordination to joint operations
to equity investments. Figure 1.6 summarizes the varying degrees of
integration that are possible and denotes when DOT's approval has
traditionally been required. U.S. antitrust laws limit the level of
integration that competing airlines can achieve.\5 However, Northwest
and KLM can integrate their operations in such areas as pricing
without fear of legal challenge from competitors because, as
discussed below, DOT granted that alliance antitrust immunity
("merger" model). In granting immunity in this case, though, DOT
stated that it believed the antitrust laws would not bar the carriers
from integrating their operations as planned because their
cooperation would not result in a substantial lessening of
competition since they were not significant competitors on most
routes served by the alliance. However, the agency granted immunity,
finding that "the parties are unlikely to proceed with the Agreement
without antitrust immunity."
Finally, some airlines have made equity investments in other airlines
in an effort to own a large portion of another airline ("investor"
model). For example, in addition to its code-sharing arrangement
with USAir, British Airways invested $400 million in USAir and now
owns just under 25 percent of that airline and holds 3 seats on
USAir's 16-member board of directors. U.S. law limits the voting
interest that a foreign airline can have in a U.S. airline to 25
percent and requires that control of the airline be exercised by U.S.
citizens.
Figure 1.6: Levels of
Integration Between U.S. and
Foreign Airlines
(See figure in printed
edition.)
Note: DOT has traditionally not required that interline and simple
marketing agreements be filed with the agency for approval.
--------------------
\4 The airline that purchases the block of seats also lists the
flight in CRSs under its own designator code.
\5 U.S. antitrust laws do not prevent two carriers that are not
significant competitors from integrating their services. For
example, U.S. airlines commonly integrate their operations with
their commuter partners in the domestic market and have not
sought--nor do they need--antitrust immunity, according to DOT and
Justice Department officials.
DOT HAS AUTHORITY TO GRANT
ALLIANCES IMMUNITY FROM U.S.
ANTITRUST LAWS
---------------------------------------------------------- Chapter 1:4
U.S. law gives the Secretary of Transportation the authority to
grant immunity from U.S. antitrust laws to agreements in foreign air
transportation. In general, the antitrust laws are designed to
protect consumers by prohibiting competitors from colluding and
engaging in such anticompetitive behavior as jointly setting prices
(commonly referred to as "price fixing"). The Secretary may grant
immunity if an agreement is in the public interest and is necessary
to permit implementation of an approved cooperative agreement. If
the Secretary finds that a cooperative agreement will substantially
reduce or eliminate competition, however, the Secretary may only
approve it if (1) the agreement is necessary to meet a serious
transportation need or to achieve important public benefits,
including international comity and foreign policy considerations, and
(2) that transportation need or those public benefits cannot be
achieved by reasonably available alternatives that are less
anticompetitive.
Only one code-sharing alliance between a U.S. and foreign airline
approved by DOT since 1987 has applied for antitrust immunity, and in
that case DOT granted it. In November 1992, DOT, working in
conjunction with the Department of Justice, approved the application
of Northwest and KLM--an action closely linked to the September 1992
"open skies" bilateral agreement between the United States and the
Netherlands.\6 As stated earlier, in granting immunity to the
alliance, DOT said that it believed that the antitrust laws would not
bar the carriers from integrating their operations as planned.
However, DOT granted immunity on the basis of its finding that the
agreement was in the public interest and that it was unlikely,
without antitrust immunity, that the parties would proceed with the
agreement for fear of legal challenge from competitors.
--------------------
\6 The open skies accord between the United States and the
Netherlands removed all restrictions on air travel between the two
countries, thereby allowing any U.S. carrier to serve any point in
the Netherlands and beyond from any point in the United States and
allowing any Dutch carrier to do the same.
OBJECTIVES, SCOPE, AND
METHODOLOGY
---------------------------------------------------------- Chapter 1:5
Citing the increasing number of code-sharing alliances between U.S.
and foreign airlines, the Chairmen and Ranking Minority Members of
the Senate Committee on Commerce, Science, and Transportation and its
Subcommittee on Aviation asked us to determine the (1) extent to
which U.S. and foreign airlines participating in alliances benefit
from those alliances in terms of additional passengers and revenues
and (2) effect that alliances have on other U.S. airlines and on
consumers. They also asked us to identify and examine key areas of
concern, if any, pertaining to alliances that were not addressed by
DOT's recent policy statement or proposed regulatory actions.
To determine the extent to which U.S. and foreign airlines
participating in alliances benefit from those alliances and the
alliances' impact on other U.S. airlines, we analyzed data provided
by U.S. and foreign airlines on passenger traffic and revenues.
Because DOT's traffic data are not sufficiently detailed to fully
analyze such effects, we relied heavily on airlines' internal data in
conducting most of our analyses (ch. 3 discusses the limitations of
DOT's data). We also interviewed DOT's Director, Office of Aviation
and International Economics, as well as analysts in that office and
reviewed GRA's study of 3 months of limited DOT data on the
Northwest/KLM and USAir/British Airways alliances. In addition, we
interviewed representatives of the major U.S. airlines that fly
internationally--American, Continental Airlines, Delta Air Lines,
Northwest, TWA, United, and USAir. Likewise, we interviewed
representatives of the following foreign airlines: Ansett Australia,
British Airways, British Midland, Cathay Pacific, China Airlines,
Lufthansa, KLM, Philippine Airlines, Qantas, Singapore International
Airlines, Swissair, Thai International Airways, and Virgin Atlantic.
By selecting these 7 U.S. and 13 foreign airlines, we were able to
collect information on 85 percent of the 61 code-sharing alliances
approved by DOT since 1987.
To assess the impacts of the alliances on consumers, we interviewed
officials from the Justice Department, including the Chief of the
Transportation, Energy, and Agriculture Section of Justice's
Antitrust Division, and the Executive Director of the International
Airline Passengers Association to obtain their views on the potential
long-term effect of alliances on competition and fares. We also
interviewed nine airport representatives from the organization known
as U.S. Airports for Better International Air Service to obtain
their perspectives on the effect of code-sharing alliances on
consumers in their communities. The limitations of DOT's traffic
data, however, prevented us from determining the effect that
alliances have had on fares.
To identify and examine key areas of concern, if any, pertaining to
alliances that were not addressed by DOT's recent policy statement or
proposed regulatory actions concerning consumer notification, we
analyzed DOT's policy statement and proposed rules, examined DOT's
past orders approving code-sharing alliances, and reviewed relevant
U.S. laws and regulations. We also discussed the implications of
our analyses with DOT and State Department officials. To obtain
foreign perspectives, we interviewed officials from the
transportation departments and civil aviation authorities of several
European and Pacific Rim nations. In Europe, we interviewed
officials from Germany, the Netherlands, Switzerland, and the United
Kingdom. In the Pacific Rim, we interviewed officials from
Australia, Hong Kong, the Philippines, Singapore, Taiwan, and
Thailand.
To examine issues relating to the listing of code-share flights in
CRSs, we interviewed DOT officials and reviewed DOT's regulations
governing CRS displays. We also interviewed the Assistant Director,
Industry Affairs for the American Society of Travel Agents (ASTA),
representatives of seven judgmentally selected travel agencies in the
United States, and representatives of several foreign travel agency
associations, such as the Australian Federation of Travel Agents. We
reviewed CRS listings of code-share flights and discussed with agents
code-sharing's impact on their work. We also interviewed
representatives of the EU to determine how it regulates the listing
of code-share flights in European CRSs.
We discussed a draft of this report with senior DOT and State
Department officials, including DOT's Acting Assistant Secretary for
Aviation and International Affairs and the State Department's
Director, Office of Aviation Programs and Policy. On the basis of
their comments, we revised the report where appropriate. We have
included a detailed discussion of their comments and our changes at
the end of chapter 3. As requested, however, we did not obtain
written comments on a draft of this report. Finally, U.S. and
foreign airline representatives reviewed relevant sections of a draft
of this report relating to their airlines. We incorporated their
comments and suggested revisions where appropriate. We conducted our
work from February 1994 to February 1995 in accordance with generally
accepted government auditing standards.
ALLIANCES' BENEFITS FOR PARTNERS
AND EFFECTS ON OTHER AIRLINES
VARY, AND IMPACT ON FARES IS
UNCERTAIN
============================================================ Chapter 2
The extent to which airlines participating in alliances benefit from
them varies greatly and depends on the (1) geographic scope of the
code-sharing arrangement, (2) level of operating and marketing
integration achieved by the airlines, and (3) agreement between the
airlines on how to divide revenues. Conversely, the impact on other
U.S. airlines in terms of reduced ridership and revenues depends on
an alliance's scope and integration, the other airlines' competitive
responses, and the extent to which competition between that alliance
and the other airlines leads to lower fares and stimulates new
traffic. We were unable, however, to obtain sufficient data from the
airlines--and DOT's data are insufficient--to determine what effect
alliances have had on fares in the short term and whether alliances
will reduce or enhance competition in the long term and thereby lead
to higher or lower fares.
STRATEGIC ALLIANCES GREATLY
BENEFIT PARTICIPATING CARRIERS
AND REDUCE TRAFFIC AND REVENUES
FOR OTHER AIRLINES
---------------------------------------------------------- Chapter 2:1
Of the 61 alliances between U.S. and foreign airlines, 3 employ
code-sharing on a vast number of routes so as to strategically link
both airlines' flight networks. These strategic
alliances--Northwest/KLM (formed in 1992), USAir/British Airways
(1993), and United/Lufthansa (1994)--are producing large increases in
the number of passengers traveling on these airlines. This effect is
occurring because of the (1) broad nature of the code-sharing
arrangements and (2) great degree of integration achieved by the
carriers in scheduling, operations, advertising, and frequent flyer
programs. However, the extent to which each airline in these
alliances is benefiting in terms of added revenues varies depending
on the details of each agreement. Limited data indicate that
alliances' traffic and revenue gains are generally coming at the
expense of other U.S. and foreign airlines, although airline
representatives and DOT officials we interviewed contend that some
gains have come from traffic stimulated by increased competition
among the alliances and between alliances and other airlines.
ALLIANCE BETWEEN NORTHWEST
AND KLM IS PRODUCING SIZABLE
BENEFITS FOR BOTH AIRLINES
-------------------------------------------------------- Chapter 2:1.1
As a result of their strategic alliance, both Northwest's and KLM's
riderships have increased dramatically over the last few years.
Northwest's data indicate that for the year ended June 1994, over
353,000 passengers traveled on Northwest aircraft as part of the
alliance, compared to 164,450 passengers traveling on connecting
Northwest and KLM interline flights in 1991.\1 In addition to this
increase of nearly 200,000 passengers on Northwest aircraft, KLM
representatives estimated that about 150,000 passengers traveled on
code-share flights in which only a KLM aircraft was involved during
this period.
Northwest and KLM representatives emphasized that although improving
economic conditions in the United States and Europe since 1991 have
helped increase their riderships, the alliance has been a key factor
in their traffic growth. Northwest representatives pointed out, for
example, that Northwest has never served the 30 overseas cities that
they now serve by code-sharing on KLM's flights. Thus, traffic
connected from these cities to Northwest's flights between Amsterdam
and the United States is primarily additional traffic caused by
code-sharing, not improved economic conditions. For example, they
noted that it would require an investment of several airplanes and
millions of dollars for Northwest to serve Oslo, Norway, via its
Minneapolis hub. However, through the alliance, Northwest has added
to its system over 30 passengers per day who fly on KLM's flights
between Oslo and Amsterdam and connect to Northwest's flights between
Amsterdam and the United States.
Because the airlines (1) divide the resulting revenues on the basis
of an agreed prorated formula that accounts for the miles each
airline flies under the alliance and (2) both airlines fly numerous
long-haul routes as part of the alliance, the increased ridership
resulting from the alliance has had a significant impact on both
airlines' financial performances.\2 Likewise, increased interline
traffic from non-code-share cities and cost savings have benefited
both airlines. On the basis of our discussions with Northwest
representatives and analysis of Northwest's traffic and confidential
data, we estimate that the alliance produced between $125 million and
$175 million in added revenues for the airline in 1994. These
revenues represent about one-third of Northwest's $455 million in
transatlantic passenger revenues and about 5 percent of its $3
billion in total international passenger revenues in 1994. These
added revenues helped Northwest post a company record $830 million
operating profit in 1994 as opposed to a loss of $60.1 million in
1991 and $141.7 million in 1990. Similarly, we estimate that KLM
earned approximately $100 million in added revenues as a result of
the alliance during 1994. The added revenues constitute 18 percent
of KLM's transatlantic passenger revenues and 3 percent of its
overall international passenger revenues.
The alliance's success is due to the broad scope of the code-sharing
network and the degree of integration the airlines have achieved.
First, they have scheduled flights to take advantage of Northwest's
hubs (Boston, Detroit, and Minneapolis) and KLM's Amsterdam hub. By
doing so, they link Northwest's domestic service from 88 interior
U.S. cities with 30 cities in Europe and the Middle East (fig.
2.1).
Figure 2.1: Northwest/ KLM's
Code-Sharing Network as of Dec.
31, 1994
(See figure in printed
edition.)
Legend: KL--KLM; NW--Northwest.
Notes: Northwest also code-shares on KLM's flights between Amsterdam
and eight KLM gateway cities in the United States.
Source: GAO's illustration of Northwest's data.
Second, antitrust immunity has allowed Northwest and KLM to achieve a
high level of integration without fear of legal challenges from
competitors. Northwest and KLM representatives stated that immunity
allows them to jointly develop fares for routes served by the
alliance. Without immunity, airlines that are significant
competitors cannot discuss pricing issues and must develop prorate
agreements in "arm's length" negotiations to divide revenues, a
cumbersome process when thousands of city-pairs are involved. With
immunity, Northwest and KLM can develop formulas to set fares in all
markets and, according to Northwest and KLM representatives, quickly
enact fare reductions to attract traffic. Antitrust immunity has
also allowed the carriers to develop, without fear of legal reprisal,
(1) a joint identity by operating under the same service mark, which
features the names of both airlines, and (2) common incentives for
their sales forces so that they market the flights of both airlines
throughout the world.
DOT and Justice Department officials noted, however, that the high
degree of integration that the two carriers have achieved would not
violate antitrust laws if the carriers did not have immunity because
before the alliance the airlines were not significant competitors on
most routes. These officials stated that they believed the key
benefit of immunity in this case is the protection from legal
challenge by other airlines, thereby allowing Northwest and KLM to
more closely integrate their operations and marketing than they
otherwise would for fear of legal reprisal. DOT officials agreed
with Northwest's contention that the two airlines would not have
pursued the existing high level of integration--especially in the
area of pricing--without immunity because of this fear.
In addition to the areas discussed above, Northwest and KLM have
integrated in other areas since 1992. For example, they
created marketing products, such as World Business Class (a special
section of seats and service for business travelers), that are
common to both Northwest's and KLM's flights so as to attract
international business travelers;
contracted for the same branding of airplane exteriors and
interiors, uniforms, vehicles, and stationary (e.g., same style,
color-scheme), and as a result, such things as the pitch to
which seats recline and the type of dinner plates and napkins
are the same for Northwest and KLM airplanes, thereby reducing
purchase costs and highlighting for passengers the level of
integration achieved; and
produced television ads emphasizing their integration and the
resulting benefits for consumers (e.g., better service, reduced
layover times).
Our discussions with U.S. and foreign airline representatives
indicate that much of the alliances' traffic gains have come at the
expense of other U.S. and foreign airlines, although we were unable
to obtain sufficient data to precisely quantify these impacts.
Northwest and KLM representatives stated that the alliance has
increased their combined transatlantic market share from 7 percent
before the alliance to 11.5 percent in 1994. Our analysis of data
provided by Northwest documents the increasing share (from 1.2
percent in 1991 to 3.3 percent in 1994) of passengers traveling on
Northwest and KLM between 34 U.S. interior cities and 30 European
and Middle Eastern cities that are key to the alliance (table 2.1).\3
Table 2.1
Number of Northwest/KLM Passengers
Traveling Between 34 U.S. and 30
European and Middle Eastern Interior
Cities, and Alliance's Share of Those
Markets, 1991-94
6/93 to 6/
1991 1992 1993 94
------------ ---------- ---------- ---------- ----------
Northwest/ 17,150 23,260 52,510 60,630
KLM
All carriers 1,488,160 1,688,570 1,744,090 1,810,780
Northwest/ 1.2 1.4 3.0 3.3
KLM market
share
(percent)
------------------------------------------------------------
Source: GAO's analysis of data provided by Northwest.
Representatives of the other six U.S. airlines that fly
internationally stated that their airlines had lost traffic and
revenues to the Northwest/KLM alliance. For example, Continental
representatives estimated that the airline lost about $1 million in
revenues in 1994 because traffic it would normally fly between the
United States and Europe shifted to the Northwest/KLM alliance
(approximately 0.3 percent of that carrier's $325 million in
transatlantic passenger revenues for that year). Representatives of
Continental, which did not make a profit in 1994, emphasized that
because of the small profit margins in the airline industry, such
revenue is important. Likewise, representatives of several foreign
carriers emphasized that they have lost traffic and revenues to
Northwest and KLM. Most U.S. and foreign airlines did not have or
would not provide data, however, that would allow us to determine the
extent of those losses or whether U.S. airlines were losing more
than foreign airlines.
Likewise, DOT's data are not sufficiently detailed to allow such a
determination. However, in examining the agency's data for the first
3 months of 1994, DOT's contractor, GRA, concluded that although some
U.S. carriers had lost traffic to the Northwest/KLM alliance, the
U.S. industry overall was receiving a small net gain in revenues in
light of the benefits accruing to Northwest. In reaching its
conclusion, GRA acknowledged that the limitations of DOT's data
caused it to make "important theoretical and computational
compromises."
--------------------
\1 The latest available data Northwest had were for the year ended
June 1994.
\2 If one carrier flies more of the long-haul routes, it generally
accrues more of the resulting revenues.
\3 These "interior" cities are cities other than Northwest or KLM
gateway cities (e.g., Minneapolis, Amsterdam). Examples of interior
U.S. cities in this analysis are Des Moines, Iowa, and Albuquerque,
New Mexico. Examples of interior European cities used in this
analysis are Hamburg, Germany, and Milan, Italy.
USAIR/BRITISH AIRWAYS
ALLIANCE IS YIELDING
INCREASING BENEFITS TO
BRITISH AIRWAYS
-------------------------------------------------------- Chapter 2:1.2
Whereas Northwest and KLM code-share on each other's routes, which
yield roughly equivalent benefits for the partners, the USAir/British
Airways alliance, which began in May 1993, involves only code-sharing
by British Airways on USAir's flights within the United States.
Under this arrangement, USAir does not list British Airways' flights
as its own (fig. 2.2).\4 Because it does all of the long-haul flying
across the Atlantic, British Airways under its prorate agreement with
USAir keeps most of the revenues resulting from the code-sharing
arrangement. In addition to the 7 percent dividend paid quarterly to
British Airways by USAir on the British carrier's $400 million
investment, the revenues from the code-sharing arrangement are a
return on that investment for British Airways.\5 Through such equity
investments in other airlines, British Airways seeks to create a
global network ("investor" model in fig. 1.6). Although USAir's
main benefit from the alliance was the $400 million
cash-infusion--capital that was critical to the viability of the
financially struggling airline--USAir also benefits from some added
revenues due to the (1) code-sharing arrangement, (2) increased
interline traffic resulting from frequent flyer links with British
Airways, and (3) "wet leasing" of three aircraft to British Airways
for transatlantic operations.\6
Figure 2.2: USAir/British
Airways' Code-Sharing Network
as of Dec. 31, 1994
(See figure in printed
edition.)
Legend: BA--British Airways.
Source: GAO's illustration of USAir's data.
Our analysis of British Airways' data indicates that its alliance
with USAir, despite its more limited scope, is attracting an
increasing number of passengers. Between May 1993, when the
code-sharing arrangement was implemented, and March 1994 (11 months),
14,300 passengers traveled on USAir/British Airways' code-share
flights. Between April and December 1994 (9 months), 47,749
passengers traveled on those flights.
USAir's data confirm this increase in the number of code-share
passengers. Their data are based on passenger bookings, which
include "no shows," and as a result, are somewhat higher numbers than
British Airways' data on actual ridership presented above. Table 2.2
presents USAir's data and indicates that the vast majority of
bookings have come in the last 9 months of 1994.
Table 2.2
Number of Passengers Booked to Travel on
USAir/British Airways' Code-Share
Flights, May 1993-Dec. 1994
Number of code- Number of code-
share passengers share passengers
Quarter 1993 1994
-------------------- ------------------ ------------------
I (Jan-Feb-Mar) N/A 9,189
II (Apr-May-June) 914 20,058
III (July-Aug-Sep) 3,113 21,500
IV (Oct-Nov-Dec) 4,412 16,846
============================================================
Total 8,439 67,593
------------------------------------------------------------
Notes:
1. Code-sharing arrangement did not begin until May 1993; thus, there
are no data for the first quarter of 1993.
2. Bookings are as of day of flight.
Source: GAO's analysis of USAir's data.
Most passengers traveling on USAir/British Airways' code-share
flights represent new traffic for British Airways because the airline
did not previously serve the 52 code-share cities, having instead
interline agreements with U.S. airlines. In addition, USAir and
British Airways representatives stated that they believed these
increasing figures were the result of an increased level of
coordination and integration between the airlines. For example, they
noted that as the result of the airlines' marketing efforts, USAir
passengers had a growing awareness that they can use their frequent
flyer miles to earn free trips on British Airways' flights.
An additional benefit of the alliance for British Airways has been a
substantial increase in its interline traffic with USAir from U.S.
cities other than the 52 code-share cities. For example, a
comparison of British Airways' traffic data for April through
December 1994 with the same time period a year earlier shows that the
number of USAir/British Airways interline passengers has increased by
60 percent, from 36,396 to 58,164.
The code-share and interline traffic gains have produced sizable
revenues for British Airways in the transatlantic market, although
they are small compared to British Airways' overall international
operating revenues. British Airways representatives estimated that
between April 1994 and March 1995, the alliance will produce $100
million in revenues for the airline--$45 million from the code-share
traffic and $55 million from the increased interline traffic, linked
frequent flyer programs, and cost savings.\7 The $100 million in new
revenue is equivalent to 5 percent of British Airways' $2.1 billion
in revenues from traffic to and from the United States and 1 percent
of its $8.5 billion in total international revenues. USAir, on the
other hand, earned about $20 million in added revenues from the
alliance in 1994--approximately $8 million from the code-share
traffic and $12 million from the increased interline traffic and the
wet lease arrangement. British Airways representatives stated that
the revenues it was accruing as a result of the alliance represented
a reasonable return on their investment. They also noted that the
alliance was producing benefits in part because a relatively high
proportion of the code-share traffic was premium traffic (first and
business classes), which generally pays higher fares.\8
Finally, according to USAir and British Airways representatives, the
benefits produced from their alliance will increase as the airlines
increase their level of integration and add more U.S. cities to
their code-share network. Although the alliance was implemented in
May 1993, the airlines have been slow to integrate their operations
and marketing because, according to USAir and British Airways
representatives, DOT has on several occasions threatened to
disapprove their code-sharing arrangements. In November 1993, for
example, DOT approved code-sharing between the two airlines for only
a 60-day period and warned that it may disapprove the code-sharing
arrangements at the end of that period.\9 Currently, the alliance has
DOT's approval until March 17, 1995. The uncertainty surrounding
DOT's approval, "one-way" nature of the arrangement, and current
financial distress of USAir--the company had an operating loss of
$491 million in 1994--has resulted in less integration than that of
the Northwest/KLM alliance and thus smaller benefits.
Unlike the Northwest/KLM experience, the results from code-sharing
favor the foreign carrier more, although the difference is
attributable in part to British Airways' equity investment. Gains to
British Airways are largely at the expense of other U.S. airlines.
Limited data and our discussions with British Civil Aviation
Authority officials, including the authority's Head of Air Services
Policy and Industry Affairs, and representatives of British Airways
and several U.S. airlines indicate that much of the
behind-U.S.-gateway traffic (i.e., passengers traveling to and from
U.S. interior cities) now traveling on USAir's flights within the
United States and connecting to British Airways' service used to be
traffic traveling on (1) other U.S. airlines within the United
States that interlined with U.S. or British carriers or (2) the same
U.S. carrier throughout, including on-line service by USAir. For
example:
Interline Traffic. A comparison of British Airways' data for April
through December 1994 with the same 9 months in 1993 show that
the number of passengers traveling on United within the United
States and interlining with British Airways declined by 15
percent; the number of Delta's passengers declined by 12
percent; Northwest's by 9 percent; TWA's by 6 percent;
Continental's by 5 percent; and American's remained virtually
the same. By comparison, the number of passengers traveling on
USAir-British Airways' code-share flights grew from 6,589
(between April and December 1993) to 47,749 (between April and
December 1994)--an increase of about 625 percent.
Representatives from the U.S. airlines listed above told us
that the decline in their interline traffic with British Airways
is now traffic that is flying on the USAir/British Airways
alliance. For example, Delta representatives estimated that the
carrier lost about $25 million in 1994 to the alliance.
Same U.S. Carrier Throughout. (1) One U.S. airline told us that
it lost over $40 million in 1994 because traffic it used to fly
between the United States and London is now taking USAir/British
Airways' code-share flights. This amount represents 11 percent
of that airline's transatlantic operating revenues. (2) Another
U.S. airline's data show that the number of passengers it flies
between eight interior U.S. cities and the United Kingdom,
routes on which it competes with the USAir/British Airways
alliance, declined by about 11 percent between 1992 and 1993,
while its overall traffic between the United States and the
United Kingdom declined by 3 percent.
Similarly, on the basis of its analysis of DOT's data for the first 3
months of 1994, GRA concluded that the alliance was causing a net
negative flow of revenues out of the U.S. airline industry. As
stated earlier, however, data limitations prevent a precise
determination of the losses or the extent to which competing foreign
airlines have been affected.
--------------------
\4 The right to code-share extensively within the United States was
granted to British carriers in March 1991 as part of a revision to
the U.S. bilateral accord with the United Kingdom. In approving the
subsequent USAir/British Airways alliance, the Department of Justice
required USAir to divest itself of its three U.S.-United Kingdom
routes (as the USAir-British Airways agreement proposed).
Subsequently, DOT awarded those routes to American Airlines. USAir
does not code-share on British Airways' flights because (1) the
U.S.-United Kingdom bilateral agreement does not provide for it and
(2) USAir has not requested such authority.
\5 Because of its financial problems, USAir did not pay this dividend
to British Airways for the fourth quarter of 1994.
\6 Under this arrangement, USAir aircraft--painted in British
Airways' livery--and crew operate British Airways' flights between
London and Baltimore, Charlotte, and Pittsburgh.
\7 British Airways' fiscal year is from April to March. In the first
11 months of the alliance (May 1993-March 1994), it produced between
$20 million and $30 million in added revenues for British Airways,
according to airline representatives.
\8 USAir provided us confidential booking data that support British
Airways representatives' statements.
\9 DOT officials emphasized to us that temporary approval is linked
to their efforts to obtain a less-restrictive bilateral agreement
with the United Kingdom. They stated that the 1991 agreement with
the United Kingdom was "unbalanced" and provided too much benefit to
British Airways relative to the opportunities for U.S. airlines to
and beyond London's Heathrow Airport.
UNITED/LUFTHANSA ALLIANCE
BEGINNING TO PRODUCE
BENEFITS FOR BOTH AIRLINES
-------------------------------------------------------- Chapter 2:1.3
In part to counter the success of the Northwest/KLM alliance, United
and Lufthansa in June 1994 implemented a marketing alliance that uses
code-sharing to link both carriers' route networks (in fig. 1.6,
this is the "marketing agreement + code- sharing" model). Under this
arrangement, Lufthansa code-shares on United's flights between
Frankfurt and 25 U.S. interior cities via two of United's
hubs--Chicago O'Hare and Washington Dulles. United code-shares on
Lufthansa flights between Frankfurt and 30 European and Middle
Eastern cities (fig. 2.3). Ultimately, United and Lufthansa plan to
expand the alliance to add more cities and include Thai Airways,
thereby creating a global code-sharing network that spans Europe,
Africa, the Middle East, and the Pacific Rim.
Figure 2.3: United/Lufthansa's
Code-Sharing Network as of Dec.
31, 1994
(See figure in printed
edition.)
Legend: LH--Lufthansa; UA--United.
Note: United also code-shares on Lufthansa's flights between
Frankfurt and Lufthansa's 10 U.S. gateway cities. In addition,
Lufthansa code-shares on United flights between Lufthansa's 10
gateway cities and the 25 U.S. interior cities.
Source: GAO's illustration of United's data.
On the basis of internal data for June through December 1994,
United's Vice President, Resource Planning, stated that the alliance
has increased the airlines' total traffic by about 600 passengers per
day, and he projected that the alliance will increase their traffic
by a total of 219,000 passengers between June 1994 and June 1995. He
and other United representatives emphasized that much of this traffic
represents passengers traveling between the United States and the 30
foreign cities that United previously did not serve. They also
emphasized that the additional passengers per day added by the
alliance has been steadily increasing as the scope of the
code-sharing arrangement expands and the level of integration grows
and could reach 1,000 additional passengers per day by mid-1995.
Although Lufthansa representatives declined to provide data, they
stated that they believed that United's projection was accurate.
Representatives of both airlines stated that the traffic generated by
the alliance is exceeding their expectations. They said that most of
this traffic was being diverted from other airlines that serve those
markets, but they noted that increasing competition between the
alliance and the KLM/Northwest and USAir/British Airways alliances
was likely generating some new traffic. Likewise, representatives
from other U.S. and foreign airlines stated that they were losing
traffic and revenues to the United/Lufthansa alliance, but none had
data on or provided an estimate of these losses, in part because the
alliance was only recently implemented. Nevertheless, United
representatives emphasized that the impact of this alliance will be
less than the Northwest/KLM alliance because United and Lufthansa are
prevented by U.S. antitrust laws from achieving a level of
integration comparable to that of Northwest and KLM. For example,
they are prevented from jointly setting fares. However, DOT
officials emphasized that the United/Lufthansa alliance differs from
the Northwest/KLM alliance in that United and Lufthansa are
significant competitors in several city-pair markets served by the
alliance. As a result, they noted that competition may be reduced if
they were able to integrate operations with the protection of
antitrust immunity. Justice Department officials noted that
significant competition issues would be raised that did not exist in
the Northwest/KLM case.
REGIONAL ALLIANCES ALSO PRODUCE
BENEFITS FOR AIRLINES,
DEPENDING ON THE LEVEL OF
INTEGRATION ACHIEVED
---------------------------------------------------------- Chapter 2:2
Several regional alliances, which connect a limited number of routes
to and from a specific region, have generated modest traffic gains
for the carriers involved. Successful alliances have been
characterized by a high level of integration. For example, over the
last 2 years, United and Ansett Australia have worked closely to
develop and market their alliance. Both United and Ansett
representatives told us that the number of code-share passengers far
exceeded their expectations. United representatives estimated that
approximately 120 passengers a day (or 43,800 passengers per year)
are traveling on United between Sydney and the United States that are
also connecting to Ansett flights between Sydney and eight interior
Australian cities. Before the alliance, United did not serve these
cities. Through code-sharing with Ansett, United can market service
to these cities through the CRSs.
United representatives also stated that United was obtaining about
$14 million in revenue from the alliance. Although less than 1
percent of United's $2.6 billion in transpacific passenger revenues
and $4.2 billion in total international passenger revenues, this
revenue is important, according to United representatives, given the
thin profit margins in the airline industry. For example, the $14
million provided by the Ansett alliance contributed to United's $521
million overall operating profit in 1994--only the airline's second
operating profit in 5 years.
The increased ridership that United has experienced in its flights
across the Pacific as a result of the alliance has come largely at
the expense of its two main competitors on routes between Australia
and the United States--Northwest and Qantas-- according to
representatives from United, Ansett, Northwest, and Qantas. United's
Vice President, Resource Planning, noted, though, that some of the
traffic gains were being stimulated by increased competition between
the alliance, Northwest, and Qantas. However, we were unable to
obtain data to determine (1) Northwest's and Qantas' traffic and
revenue losses because of the alliance and (2) the extent to which
traffic had been stimulated by increased competition--and presumably
lower fares--as a result of the alliance.
United has also entered into a regional alliance with British Midland
that has generated similar benefits, increasing the number of
passengers riding on United across the Atlantic by approximately
30,000. Data provided by British Midland support this estimate.
From January 1991 through March 1992, British Midland carried an
average of 151 passengers per month on an interline basis with United
between London's Heathrow Airport and five cities in northern Europe
(United flies the passengers between the United States and Heathrow).
Since beginning a code-sharing arrangement with United in April 1992,
British Midland has flown an average of 2,072 United passengers per
month between Heathrow and the five cities (or about 25,000
passengers per year). (Fig. 2.4 demonstrates this increase for one
of the five cities.)
Figure 2.4: Comparison of the
Number of Passengers Traveling
on Connecting United/British
Midland's Flights Between the
United States and Glasgow
Before and After Code-Share
Alliance
(See figure in printed
edition.)
Source: GAO's analysis of British Midland's data.
According to United and British Midland representatives, these gains
have been the result of their joint efforts to market the alliance.
They noted that the gains have come largely at the expense of one of
United's main competitors across the Atlantic and British Midland's
main intraeuropean competitor--British Airways. In addition, British
Midland's Industry Affairs Manager stated that a small portion of the
alliance's gains were resulting from the increasing competition
between the alliance and British Airways. He stated that he believed
that the alliance's success has caused the competing airline to
respond and that this has led to better service and lower fares,
thereby stimulating some new traffic. Again, however, we were unable
to obtain data from the airlines to precisely determine the extent of
the losses for competing airlines as a result of the alliance or the
extent to which any new traffic had been generated by the alliance.
In contrast to the regional alliances described above, two regional
alliances--Northwest/Ansett Australia and TWA/Gulf Air--failed to
produce such results and were terminated. Before aligning with
United, for example, Ansett Australia had a regional alliance with
Northwest. However, the level of coordination and integration
between the carriers was far less than in the United/Ansett Australia
alliance, according to Ansett representatives. As a result, the
alliance produced only a handful of passengers each month.
BENEFITS DERIVED BY AIRLINES
FROM POINT-SPECIFIC ALLIANCES
VARY
---------------------------------------------------------- Chapter 2:3
The majority of the 61 alliances to date involve arrangements that
are more limited than strategic or regional alliances. These
arrangements entail code-sharing in a small number of city-pair
markets. Oftentimes, these point-specific alliances involve
blocked-space agreements in which one carrier purchases a block of
seats on another carrier's flights and sells them independently.
Many of these arrangements have failed because the airlines involved
compete against each other rather than effectively integrating their
operations. Nevertheless, although they do not produce the same
magnitude of benefits for airlines as strategic and regional
alliances do, these alliances can be profitable if the partners
effectively integrate their operations and marketing.
As of December 31, 1994, DOT had approved 50 point-specific
alliances. Roughly one-third of these alliances have been terminated
by the airlines involved because they failed to produce the traffic
and revenues expected. For example, in 1992 American and Cathay
Pacific terminated their blocked-space arrangement through which
American purchased and resold seats on Cathay Pacific's flights
between Los Angeles and Hong Kong. According to American
representatives, the airline entered this agreement because it
believed it would be too costly to fly the route itself. However,
they stated that although the number of passengers generated by the
arrangement met their expectations, they were unable to make a profit
on the route because they had to charge very low fares to attract
passengers.\10 Although consumers benefit from such reduced fares,
Cathay had reduced the fares on its seats so low, according to
American representatives, that American had to lower its fares to the
point that it could not make a profit.
Several point-specific, blocked-space alliances are producing
benefits for partners. For example, as of December 31, 1994, Delta
had blocked-space agreements with nine airlines around the globe in
which it either purchases seats on their flights to specific cities
or the foreign carrier purchases seats on Delta's flights (see app.
I for a listing of Delta's nine partners). According to Delta
representatives, it is too costly to serve many of these points
directly. In other cases, foreign carriers cannot afford to provide
direct service.
Several of Delta's arrangements have been very successful for the
carrier, in contrast to the airline's overall international results.
The success of its alliances is occurring primarily because Delta has
worked closely with each foreign partner to integrate operations and
jointly market their arrangement. In its arrangement with Swissair,
for example, Delta flight attendants are present on Swissair aircraft
for flights between New York and Zurich. In addition, according to
Delta's Director of Interline Marketing and Manager of International
Route Development, the airline has strict quality assurance
procedures to which it and the foreign partner agree to adhere.
Although declining to provide a specific estimate of revenue gains,
Delta representatives emphasized that the revenue produced by these
alliances is especially important given that the airline lost $338
million on international operations in 1994.\11
American Airlines and South African Airways have also developed a
successful blocked-space arrangement. Between the alliance's
implementation in November 1992 and September 1994, American sold
over 16,600 seats on South African Airways' flights between New York
and Johannesburg (an average of over 700 per month). American
representatives emphasized that because the carrier has worked
closely with South African Airways to develop their alliance and
prevent situations similar to their experience with Cathay Pacific,
the arrangement has been profitable for both airlines. Although
declining to give an exact dollar figure, American representatives
emphasized that the revenue produced by the alliance was important.
However, they noted that the revenues were less than 1 percent of the
airline's $1.4 billion in transatlantic passenger revenues and $3.5
billion in total international passenger revenues in 1994.\12
Such arrangements can have negative impacts on other U.S. airlines.
For example, a blocked-space arrangement between a U.S. airline and
a smaller country's flag carrier can force other U.S. airlines to
exit the market between the United States and that country.
According to TWA representatives, for example, that airline recently
exited the U.S.-Switzerland market because it could not compete with
daily nonstop service from New York to both Geneva and Zurich by the
alliance of Delta and Swissair.\13 However, DOT and State Department
officials emphasized that other factors contributed to TWA's exit in
this case. Nevertheless, to date there have been few such
occurrences of alliances forcing U.S. airlines out of markets.
However, such occurrences, according to several U.S. airline
representatives we interviewed, are increasingly possible with the
increase of code- sharing and would tend to have a negative impact on
the U.S. airline industry to the extent that long-haul flights by
U.S. airlines are replaced by foreign carrier operations.
Point-specific arrangements that involve only code-sharing and do not
involve blocked-space agreements or any type of integration and
promotion produce minimal benefits for the carriers involved and have
little impact on other carriers. For example, under its arrangement
with Midwest Express, Virgin Atlantic lists as its own, Midwest
Express flights between Boston and Milwaukee so as to link Milwaukee
passengers with Virgin's flights between Boston and London. Outside
of code-sharing on this one route, the two carriers have little
integration of operations, according to Virgin's Director, Strategy
and Route Planning. Between December 1992 and June 1993, 203
code-share passengers traveled between Milwaukee and London under
this arrangement (29 per month).
--------------------
\10 Between July 1990 and March 1992, 101,243 passengers traveled on
American's block of seats (an average of about 4,800 per month).
\11 Delta had an overall operating loss of $217 million in 1994.
\12 American recorded an overall operating profit of $1 billion in
1994.
\13 TWA recorded an overall operating loss of $137.4 million in 1994.
ALLIANCES PROVIDE BENEFITS FOR
CONSUMERS, BUT INSUFFICIENT
DATA EXIST TO DETERMINE EFFECT
ON FARES
---------------------------------------------------------- Chapter 2:4
Alliances between U.S. and foreign airlines produce several benefits
for consumers. For example, close schedule coordination between
partners often produces shorter layover times between connections.
In addition, airlines can provide one-stop check-in for passengers
even though they are connecting to a flight by another airline (the
alliance partner). Consumers' choices are also often enhanced. For
example, a passenger who wants to fly from Indianapolis to Lyon,
France, now has three competing options with minimal layover times
between flights. The passenger could fly
Indianapolis-Pittsburgh-London-Lyon on USAir/British Airways.
Alternatively, the passenger could fly Indianapolis-Detroit-
Amsterdam-Lyon on Northwest/KLM. Finally, the passenger could fly
Indianapolis-Washington, D.C.-Frankfurt-Lyon on United/Lufthansa.
Without the code-sharing alliances, the passenger would have to
interline on several different carriers, with less convenient layover
times.
Code-sharing alliances also increase international service for
customers in many U.S. cities. This service occurs when airlines
join together to serve markets that otherwise would not receive such
service. Continental's recent alliance with Alitalia will benefit
consumers in Houston, for example, because Alitalia will provide
nonstop service between Houston and Rome starting in the summer of
1995. Currently, no airlines provide such service. Before the
alliance, Continental did not serve the market because it did not
have the right to do so under the U.S.-Italian bilateral accord.
Likewise, Alitalia did not serve the market because it believed that
it could not make a profit on the route, even though it had the right
to fly the route. Because it can now obtain traffic feed from
Continental's domestic service, Alitalia will enter the market and
provide nonstop service between Houston and Rome. Consumers in other
cities have similarly benefited, or soon will benefit, from increased
service due to alliances, including:
Atlanta. As a result of Delta's blocked-space agreement with
Brazilian carrier Varig, Atlanta receives daily nonstop service
by Varig to and from Rio de Janeiro and direct, same-plane
(one-stop) service to and from Sao Paulo. Before the alliance,
no U.S. or foreign airline provided such service.
Cincinnati. As a result of Delta's blocked-space agreement with
Swissair, Cincinnati receives nonstop service by Delta to and
from Zurich. Before the alliance, no U.S. or foreign airline
provided such service.
Memphis. As a result of Northwest's alliance with KLM, Memphis
will receive, starting in the summer of 1995, nonstop service to
Europe for the first time. The city will also receive
substantially increased operations as it becomes a key hub for
the alliance in which service between over 80 U.S. interior
cities and 30 European and Middle Eastern cities will be linked
via Memphis in the United States and Amsterdam in Europe.
Washington, D.C. As a result of Delta's arrangement with Austrian
Airways, Washington, D.C. (Dulles Airport), will receive,
starting in the spring of 1995, direct, same-plane (one-stop)
service by Austrian to and from Vienna. Before the alliance, no
carrier provided such direct service between these two cities.
Whether consumers are currently paying higher or lower fares because
of code-sharing alliances is unknown, however, because DOT's traffic
data, which contain fare information, do not identify which
passengers are traveling on code-share flights or contain information
on the fares charged by foreign carriers (ch. 3 discusses these
limitations in detail). In the long run, consumers could pay lower
fares, according to many U.S. and foreign airline representatives,
as (1) airlines in alliances integrate further and achieve cost
efficiencies that could be passed on to the consumer and (2)
competition increases among alliances and between alliances and other
airlines. According to other airline representatives, the trend
toward strategic alliances could produce only a few "mega-carriers"
that will dominate the international marketplace, reduce competition,
and result in higher fares.
CONCLUSIONS
---------------------------------------------------------- Chapter 2:5
Alliances can be an effective strategy for airlines to increase their
traffic and revenues. Although these gains are often relatively
small compared to such measures as a carrier's overall international
operating revenues, they represent key sources of new traffic and
revenue for participating airlines in an industry characterized by
razor-thin profit margins. The magnitude of these gains depends on
the geographic scope of the code-sharing arrangement and the level of
integration achieved by allied airlines. Most gains come at the
expense of competing U.S. and foreign airlines; however, it is
likely that at least some are generated by traffic stimulation caused
by increased competition among alliances and between alliances and
other airlines in the short term. Nevertheless, insufficient data
exist to determine whether consumers are paying higher or lower fares
as a result of alliances and whether alliances will reduce or
increase competition in the long term and thereby lead to higher or
lower fares.
Finally, whether or not the U.S. airline industry gains as a result
of an alliance depends on the specifics of each deal. The
experiences of Northwest and United, for example, indicate that U.S.
partners can prosper greatly from such alliances. The experience of
USAir, on the other hand, is more complex because of British Airways'
equity investment. This alliance's code-sharing arrangement is
having negative consequences for many other U.S. airlines, largely
to the benefit of British Airways. Nevertheless, it is important to
consider that effect in the context of British Airways' investment in
USAir--an investment that was of critical importance to the viability
of the financially struggling U.S. airline.
KEY ISSUES CONCERNING ALLIANCES
REMAIN UNRESOLVED
============================================================ Chapter 3
Although DOT's recent international policy statement strongly
supports the creation of code-sharing alliances between U.S. and
foreign airlines, several major issues still need to be addressed.
First, as the statement points out, the agency must monitor the
effects of alliances on competition and the health of the U.S.
aviation industry. However, insufficient data are reported to DOT to
allow the agency to track these issues. Second, the agency has not
determined, in the light of the perceived benefits accruing to
Northwest and KLM as a result of immunity, whether antitrust immunity
should be potentially available for other alliances in markets that
allow for significantly increased access for U.S. airlines.
Finally, although DOT has proposed new rules to ensure that consumers
are notified as to which airline will actually be operating a
code-share flight before being booked on that flight, neither the
agency's current regulations nor its proposed rules limit the number
of times the same flight can be listed on travel agents' CRSs. The
listing of the same code-share flight option several times consumes
valuable screen space and crowds out competing listings of other
carriers' flights.
DOT'S POLICY STATEMENT
EMPHASIZES THE NEED TO MONITOR
ALLIANCES' EFFECTS, BUT AGENCY
LACKS NECESSARY DATA TO DO SO
---------------------------------------------------------- Chapter 3:1
In its November 1994 U.S. International Aviation Policy Statement,
DOT reiterated its support for code-sharing alliances between U.S.
and foreign airlines. Previously, the agency had approved and
reapproved nearly all code-sharing arrangements, although in several
cases DOT did not act until the Justice Department had completed an
informal review of the agreements' competitive effects. DOT's policy
statement acknowledged that some alliances could have negative
effects on competition in the long term. To monitor such effects,
DOT created an economic analysis unit in late 1994. The new unit
will be hindered in its ability to fulfill its mission, however,
because (1) U.S. airlines report traffic data to DOT that are not
sufficiently detailed to analyze code-sharing and (2) foreign
airlines involved in alliances with U.S. airlines are not required
to report data on their code-share traffic even though that traffic
is traveling to and from the United States.
DOT'S POLICY STATEMENT
RECONFIRMS AGENCY'S SUPPORT
FOR CODE-SHARING ALLIANCES
-------------------------------------------------------- Chapter 3:1.1
In November 1994, DOT reaffirmed the agency's support for
international code-sharing alliances. In releasing the agency's
policy statement, the Secretary of Transportation emphasized:
"We believe that enhanced airline competition and the trends of
privatization, marketing alliances, code-shares and cross-border
investments that fuel globalization are here to stay--and that these
developments offer great benefits for all nations. For our part, the
United States will support these trends."
The policy statement also emphasized that such alliances benefit both
U.S. airlines and consumers. DOT said that alliances can give the
airlines access to more markets, and thus they can gain traffic.
Moreover, according to DOT, the number of service options available
to consumers effectively increases. This increase occurs because
U.S. airlines can advertise "on-line" service to many more overseas
destinations and provide more convenient, "seamless" connections via
their foreign partners than under interline agreements. Likewise,
the agency asserted that competition among airlines will increase as
alliances compete for international passengers, thereby resulting in
lower fares and increased quality of service.
Prior to its policy statement, DOT expressed similar support for
code-sharing alliances. Between December 1987--when the agency first
required that U.S. airlines submit their code-sharing arrangements
with foreign airlines for approval--and December 31, 1994, DOT had
approved 61 alliances involving nearly 150 different code-sharing
arrangements. In general, DOT has approved and reapproved
code-sharing arrangements with little analysis. Only in one case has
DOT rejected a code-sharing arrangement. In 1991, the agency
required United and British Airways to end their point- specific
arrangement between Seattle and London as a condition of the
agreement with the United Kingdom that allowed United to replace Pan
Am as one of two U.S. carriers allowed to serve Heathrow Airport.\1
DOT also delayed approval of Delta's code-sharing arrangement with
Virgin Atlantic. The carriers applied for approval in April 1994 but
did not receive it until February 1995, because DOT was dissatisfied
with the lack of progress in liberalizing the current restrictive
bilateral accord with the United Kingdom.
--------------------
\1 In this case, DOT held that the code-sharing alliance could
potentially reduce competition on this route because United and
British Airways would be the only airlines serving it.
POLICY STATEMENT
ACKNOWLEDGES THAT ALLIANCES
COULD HAVE LONG-TERM
NEGATIVE CONSEQUENCES
-------------------------------------------------------- Chapter 3:1.2
DOT's policy statement does note that because of the greater traffic
access gained by alliance partners, alliances between U.S. and
foreign airlines may have negative impacts on competition in the
international marketplace in the future. Discussing the potential
long-run effects on competition, DOT states:
"Although we expect the expansion of cooperative arrangements
(alliances between U.S. and foreign airlines) to be largely
beneficial, there may be some negative effects. The greater traffic
access of participants may give them considerable competitive muscle,
and we may need to watch for harmful effects on competition."
Similarly, several U.S. airline representatives warned that
strategic alliances may lead to an international marketplace
dominated by only a handful of "mega-carriers" that are not
effectively competing with each other or are preventing other U.S.
airlines that are not strategically allied from entering foreign
markets. They noted that such situations would result in consumers
paying higher fares. Likewise, they cautioned that the potential
also exists for alliances to negatively affect the health of the U.S.
airline industry over the long run. Such negative effects could
occur if (1) foreign airlines take traffic and revenues away from
U.S. airlines with little corresponding benefit for the U.S.
partner in the alliance or (2) foreign countries whose national
airlines are already in strategic or regional alliances fail to
increase access for other U.S. airlines to and beyond their markets.
As the number of alliances continues to increase, according to nearly
every U.S. and foreign government official and airline
representative we interviewed, DOT and foreign governments will need
to monitor alliances for such negative long-term impacts.
ACKNOWLEDGING THAT IT HAS
CONDUCTED INSUFFICIENT
ANALYSIS IN THE PAST, DOT
HAS CREATED GROUP TO MONITOR
LONG-TERM ISSUES
-------------------------------------------------------- Chapter 3:1.3
In conjunction with the release of the policy statement, the
Secretary of Transportation announced that the agency would establish
an economic analysis unit "to focus solely on long-term strategy and
analysis of the international airline sector." The Secretary
acknowledged previously during testimony before the Subcommittee on
Aviation, House Committee on Public Works and Transportation, in May
1994 that DOT had not conducted sufficient analysis of such issues as
the impacts of code-sharing before key bilateral negotiations. He
noted that as a result, there is concern that DOT has granted foreign
airlines increased access to the U.S. market without obtaining
equivalent opportunities for U.S. airlines in foreign markets.
In his testimony, the Secretary acknowledged that the agency did not
conduct the analyses necessary to estimate the value of code-sharing
to British Airways before concluding the 1991 accord with the United
Kingdom. As a result of its desire to bolster cash-strapped Pan Am
and TWA by replacing them with United and American as the U.S.
airlines allowed to operate at London's Heathrow Airport (as well as
United's and American's strong desire to serve that airport), DOT
agreed in March 1991 to allow British Airways extensive access to
interior U.S. cities via code-sharing. At the same time, the
agreement continued to limit to two the number of U.S. airlines that
could serve Heathrow Airport and maintained tight restrictions on the
ability of those airlines to carry local traffic between London and
destinations beyond Heathrow. Many U.S. airline representatives
believe that the United Kingdom no longer has any incentive to open
its highly restricted market to U.S. airlines because British
Airways has already secured significant access to the U.S. market
through code-sharing.
In November 1994, DOT created the new economic unit--the Office of
Aviation and International Economics--and allocated it five staff.
According to DOT officials, the office will allow DOT to (1) take a
more strategic and long-term approach to bilateral negotiations and
international aviation policy-making and (2) monitor the impacts of
marketing alliances between U.S. and foreign airlines on competition
and on the health of the U.S. airline industry over the long run.
NEW ECONOMIC ANALYSIS UNIT
WILL BE HANDICAPPED BY
SEVERAL DATA LIMITATIONS
-------------------------------------------------------- Chapter 3:1.4
To fulfill its mission, DOT's Office of Aviation and International
Economics will need complete and accurate data on passengers
traveling on code-share flights to and from the United States,
whether the airline flying the route is a U.S. or foreign carrier.
Without such information, the office cannot effectively track an
alliance's impact on traffic flows and fares. Because of three key
limitations, DOT's current traffic data, which are reported quarterly
by U.S. airlines from a 10-percent sample of their tickets, do not
provide the complete and accurate information needed.
First, in reporting data from their ticket sample, U.S. airlines are
not required to identify the traffic that traveled on code-share
flights. As a result, DOT cannot readily isolate code-share
passengers and analyze trends in ridership and fares. Second, DOT's
reporting requirements are sufficiently vague that they result in
some airlines' misreporting which airline actually operated a
code-share flight. Rather than reporting which airline partner
actually operated a flight, some airlines simply report what is
printed on the ticket, which may be the code of the carrier that
marketed the service and not the carrier that actually operated the
flight. As a result, DOT's data base includes information indicating
travel on a given carrier that could not have taken place. GRA
found, for example, that the data show a number of passengers as
traveling on a KLM aircraft from Boston through Amsterdam to Athens,
even though KLM does not fly from Boston to Amsterdam. What actually
occurred was that the passengers flew on a Northwest aircraft for the
Boston-Amsterdam leg and then switched to KLM for the
Amsterdam-Athens leg. In its analysis of DOT's traffic data, GRA
concluded that "if DOT wants to monitor individual code-sharing
arrangements, it should place additional emphasis on accurate
reporting." According to airline representatives we interviewed, the
reporting requirements are too vague to ensure that airlines report
the operating carrier.
Third, DOT's data do not include information on many foreign
code-share flights operated as part of an alliance with a U.S.
airline because the agency does not require foreign airlines to
report data on a sample of their tickets, as U.S. airlines are
required to report. Currently, foreign airlines are required to
report to DOT only data on their overall traffic between their
gateway cities (e.g., New York-London). According to DOT analysts,
such data are of limited use in analyzing the effect of alliances
because they are too general. For example, they do not identify
code-share traffic or provide information on fares. Because foreign
airlines are not required to report data from a sample of their
tickets involving travel to or from the United States, DOT's traffic
data provide information only from tickets sampled by U.S. airlines.
As a result, the agency has data only on trips that at some point
involve a U.S. carrier. For example, DOT does not collect detailed
traffic data from tickets for flights originating in Detroit and
traveling on KLM aircraft to Amsterdam. Even though the flights are
Northwest/KLM code-share flights, tickets are not sampled because no
U.S. carrier is involved in the actual transporting of passengers.
KLM representatives estimated that KLM transported about 150,000
code-share passengers in 1994 in which no Northwest aircraft was
involved. Thus, DOT does not have key data, including the fare
charged, on this traffic.
Because foreign airlines are not required to report such data, DOT
also does not have information on many blocked-space alliances
because oftentimes only the foreign airlines' airplanes are used.
DOT analysts told us that they refer to this limitation as their
"foreign blind spot" and acknowledged that it prevents them from
completely (1) analyzing shifts in traffic from U.S. to foreign
carriers caused by code-sharing or (2) determining the extent to
which code-sharing benefits foreign airlines. In discussing this
limitation, GRA stated that "it is strongly suggested that DOT
consider the possibility of obtaining ticketing information from
foreign carriers...." GRA also emphasized that
"if DOT wants to continue to monitor the effects of international
code sharing on airlines and consumers, it should consider expanding
the reporting requirements for code-sharing operations, particularly
those of foreign carriers."
In part to address these limitations and begin studying the effects
of code-sharing alliances, DOT in 1994 required the three U.S.
airlines involved in strategic alliances--Northwest, United, and
USAir--to file special reports on their code-share traffic. As of
December 31, 1994, only two of the airlines--Northwest and
USAir--were filing the special reports. United representatives
stated that they will start reporting such data in early 1995.
However, according to representatives from all three airlines, it is
unfair to impose a reporting requirement on them that is not imposed
on the rest of the industry. In addition, the utility of these
special reports is limited because they do not provide the agency
with detailed data, such as the fares charged, on KLM's or
Lufthansa's traffic in cases in which only they fly the routes.
DOT HAS NOT EXAMINED THE ROLE
OF ANTITRUST IMMUNITY IN
BILATERAL TALKS IN LIGHT OF THE
NORTHWEST/KLM EXPERIENCE
---------------------------------------------------------- Chapter 3:2
Many competing airlines and foreign government officials stated that
they believe antitrust immunity has provided the Northwest/KLM
alliance with a significant advantage over the other two strategic
alliances and international carriers not strategically allied. DOT
granted immunity to the alliance in conjunction with the 1992 "open
skies" accord with the Netherlands in the hope that other countries
would follow the Netherlands' lead of agreeing to eliminate all
bilateral restrictions. As of March 10, 1995, only five smaller
countries had, and most major aviation trading partners have rebuffed
U.S. efforts to obtain open skies.\2 In light of the success of the
Northwest/KLM alliance, however, many U.S. and foreign airline
representatives and foreign government officials suggested that DOT
reexamine its policy. Many noted that the alliance's increasingly
apparent success may present DOT with a new opportunity to entice
foreign governments to liberalize their accords. Others held that
the anticompetitive effects of immunity, such as price fixing,
outweighed any benefits that could accrue from reduced bilateral
restrictions. DOT's policy statement is silent on this issue, and
DOT officials have not determined, in light of the Northwest/KLM
experience, whether antitrust immunity should be available for other
alliances in markets that allow for significantly increased access
for U.S. airlines.
--------------------
\2 In February 1995, DOT also signed a liberalized accord with
Canada. Because of several limitations on U.S. airlines--such as
phased-in access for U.S. airlines to Montreal, Toronto, and
Vancouver--it is technically not an "open skies" accord, according to
DOT officials.
MANY BELIEVE THAT ANTITRUST
IMMUNITY GIVES NORTHWEST AND
KLM AN ADVANTAGE
-------------------------------------------------------- Chapter 3:2.1
In November 1992, DOT approved the application of Northwest and KLM
for antitrust immunity, although DOT found that the antitrust laws
would not bar the carriers from integrating their operations as
planned because they were not significant competitors on most routes
that the alliance would serve. This action was closely linked to the
September 1992 "open skies" bilateral agreement between the United
States and the Netherlands that removed all restrictions on air
travel between the two countries. Furthermore, the accord
contemplated the antitrust immunity that Northwest and KLM sought.
The accord states that the United States and the Netherlands agree
"(a) to give sympathetic consideration, in the context of the Open
Skies agreement, to the concept of commercial cooperation and
integration of commercial operations between airlines of the United
States and the Netherlands through commercial agreements or
arrangements, provided that such agreements or arrangements are in
conformity with the applicable antitrust and competition laws; and
(b) to provide fair and expeditious consideration to any such
agreements or arrangements filed for approval and antitrust
immunity."
In approving the Northwest/KLM application for antitrust immunity,
DOT emphasized that the grant of such immunity was consistent with
the open skies accord.\3 DOT also implied a favorable treatment of
future applications by other U.S. and foreign airlines in exchange
for liberal aviation accords, noting that
"we would expect that our willingness to take such action [granting
antitrust immunity] might well encourage other countries to seek
similar liberal aviation arrangements with the United States . . .
so that comparable opportunities may become available to other U.S.
carriers."
In general, however, the move to such liberal aviation accords has
not occurred. In addition to the recently signed liberalized accord
with Canada, five smaller countries--Austria, Belgium, Iceland,
Luxembourg, and Switzerland--have agreed to open skies accords with
the United States since the open skies accord was signed with the
Netherlands in 1992 (as of March 10, 1995). The United States still
has restrictive agreements with governments representing major
aviation markets, such as the United Kingdom, and no agreement at all
with France and Thailand.\4
--------------------
\3 In granting antitrust immunity, DOT directed Northwest and KLM to
submit their arrangement for reexamination after 5 years.
\4 In 1978, the United States signed a relatively liberal accord with
Germany. However, in 1994, the two nations agreed to a more
restrictive accord that sets frequency and capacity growth
restrictions on U.S. airlines over the next 4 years, at which time
liberal provisions come back in force. In addition, the accord
commits both countries to seeking an "open skies" accord that would
apply at the end of this period.
DIFFERING VIEWS EXPRESSED
CONCERNING DOT'S USE OF
ANTITRUST IMMUNITY
-------------------------------------------------------- Chapter 3:2.2
Numerous representatives of U.S. and foreign airlines and foreign
government officials expressed (1) concern about the competitive
impacts of allowing only one alliance to have antitrust immunity and
(2) interest in obtaining such immunity for their particular
alliance. United representatives, for example, noted that the level
of integration their airline can achieve with Lufthansa is limited by
antitrust laws, thus ensuring that the Northwest/KLM alliance will
outcompete them. They noted, for example, that Northwest and KLM
have an advantage in attracting lucrative corporate accounts in that
they are able to make joint presentations to corporations concerning
fare discounts on international travel throughout the world. Thus,
U.S. corporations whose employees regularly travel to both Europe
and the Pacific Rim and foreign corporations whose employees
regularly travel internationally have a strong incentive to fly on
the Northwest/KLM network rather than the United/Lufthansa network.
Likewise, officials from several European and Pacific Rim nations
stated that it was unfair for DOT to give only one alliance antitrust
immunity.
In light of such sentiments, many we interviewed noted that the
increasingly apparent success of the Northwest/KLM alliance presented
DOT with a new "carrot" in its efforts to obtain open skies with
other nations. Nevertheless, others objected to such an approach,
stating that U.S. antitrust laws are designed to protect consumers
and prevent anticompetitive behavior; therefore, they continued, it
does not make sense to condone such anticompetitive behavior as price
fixing in the hopes of increasing competition.
DOT HAS NOT EXAMINED
ADVANTAGES AND DISADVANTAGES
OF CONSIDERING ANTITRUST
IMMUNITY FOR OTHER ALLIANCES
-------------------------------------------------------- Chapter 3:2.3
In defining the international aviation policy of the United States,
DOT's statement does not address issues of antitrust immunity. DOT
officials stated that the approach of exchanging antitrust immunity
for open skies was one that was employed by the previous
administration and that it is not necessarily the approach of the
current administration. DOT's Acting Assistant Secretary for
Aviation and International Affairs, for example, noted that although
antitrust immunity could be a powerful incentive for
governments--which are often seeking to benefit one national flag
carrier--to eliminate their restrictions on U.S. airlines, many
factors must be considered. Such factors, he noted, include a
government's subsidy of that airline or the anticompetitive effects
of immunity on routes where the two carriers are major competitors.
DOT officials stated that they have not examined, in light of the
Northwest/KLM experience, the advantages and disadvantages of
granting antitrust immunity in exchange for open skies. Although the
agency is currently actively pursuing open skies accords with nine
smaller European nations, the proposed "model agreement" does not
discuss antitrust immunity. In addition, DOT officials told us they
have not determined whether they would grant antitrust immunity to an
alliance in exchange for open skies with any of these nations.
Finally, DOT officials stated that they have not examined whether
Northwest and KLM should continue to be the only alliance that has
such immunity. They noted that the grant of immunity conferred on
that alliance extends until 1997, at which time it will be reviewed
and either renewed or terminated.
TRIPLE LISTINGS OF THE SAME
FLIGHT ON CRS DISPLAYS LIMIT
COMPETITION AND TRAVEL AGENTS'
EFFICIENCY
---------------------------------------------------------- Chapter 3:3
Because code-sharing involves two carriers placing their individual
designator codes on the same flight, a code-share flight is listed
twice in CRSs.\5 The number of listings for the same flight can
increase to three when connections are involved. When a flight is
listed several times, other flights that could be listed on the first
CRS display screen are "crowded out." Travel agents overwhelmingly
tend to book customers on flights listed on the first screen. As a
result, listings of connecting code-share flights several times limit
competition and reduce consumers' choices. In addition, according to
ASTA representatives and member travel agencies, they reduce the
efficiency of travel agents who take time to review flight listings
on lower CRS screens and make it harder for those agents to provide
customers with accurate information on which airline is actually
operating a code-share flight. To address this problem, the European
Union (EU) issued regulations in 1993 limiting the display of
code-share flights in European CRSs to a maximum of two. DOT's
rules, however, do not limit the number of times a flight can be
listed.
--------------------
\5 U.S. travel agents, who book approximately 80 percent of all
flights in the United States, generally use one of four CRSs: (1)
Sabre, which is owned by American Airlines' parent corporation; (2)
Apollo, which is owned by a partnership consisting of United, USAir,
British Airways, KLM, and other foreign airlines; (3) Worldspan,
which is owned by Delta, Northwest, TWA, and some Asian airlines; and
(4) System One, which is owned by an affiliate of Continental.
TRIPLE LISTINGS OF THE SAME
FLIGHT ARE PREVALENT IN CRS
DISPLAYS
-------------------------------------------------------- Chapter 3:3.1
CRSs consider consumer preferences in listing flight options. For
example, several CRSs offer a display that ranks flights in the
following order: nonstop flights, direct flights (one or more stops
on the same aircraft), and connections. Connecting flights are often
listed in terms of elapsed time between departure and arrival. In
reviewing flight listings in Sabre, Apollo, Worldspan, and System
One, we found that each CRS listed code-share flights three times
when connections were involved. For example, in our examination of
flight listings for 17 major U.S.-European city-pair markets listed
on the first two screens of Apollo and Worldspan, we often found the
same code-share flight listed three times, occurring in 38 percent of
the cases reviewed in Worldspan and 47 percent in Apollo.\6 Triple
listings occur because both carriers in an alliance list flight
segments under their own code and because CRSs also display a third
listing in which the connection is shown as an interline connection
in which the airlines that are actually operating the flights are
listed.
--------------------
\6 We reviewed flight listings for round trips in each market (thus,
34 flights for the 17 city-pair markets) for judgmentally selected
departure times. Although the flight listings were not drawn from a
statistical sample, they were requested for flights between major
U.S. cities and Europe.
TRIPLE LISTINGS OF THE SAME
FLIGHT LIMIT COMPETITION AND
DECREASE TRAVEL AGENTS'
EFFICIENCY
-------------------------------------------------------- Chapter 3:3.2
Triple listings of the same code-share flight can limit competition.
Travel agents overwhelmingly tend to book flights that are listed on
the first CRS screen. Industry studies have shown that as often as
90 percent of the time, travel agents book flights listed on the
first CRS screen. For example, a System One study of 5 days of
bookings on its system found that 93 percent were made from the first
screen. Likewise, in 1992 DOT concluded that because of time
constraints, travel agents are more likely to book a flight that
appears on the first screen. Triple listings of the same flight on
the first screen can prevent competing flight options from being
listed on that screen. Those competing options are "crowded out" and
pushed to lower, less-employed screens.
We reviewed the first screen for the 17 international city-pairs on
the Worldspan and Apollo systems and found that 19 percent of them
contained three listings of the same flight (i.e, one flight listed
three times on the first screen). In some cases, we found competing
flight options, which were pushed to a lower screen, that had fares
and/or elapsed times from departure to arrival that were equivalent
to those of the code-share flight. As shown in figure 3.1, for
example, Lufthansa flight 2423 from Berlin to Frankfurt, which
connects with United flight 941 from Frankfurt to Chicago, is listed
three times on the first screen. It is listed three different ways:
LH 2423-LH 6430 (screen one, lines 1 and 2);
UA 3647-UA 941 (screen one, lines 3 and 4); and
LH 2423-UA 941 (screen one, lines 5 and 6).
Because the same flight connection is listed three times and consumes
six of the eight lines on the first CRS screen, a competing flight
option (Lufthansa 2628--American 157 interline service) with the same
fare and an equivalent elapsed time as the code-share flight has been
pushed to the second screen. As a result, competition can be reduced
because a travel agent who habitually books flights from the first
screen would not provide consumers with information on this competing
flight option.
Figure 3.1: Crowding Out of
Flight Option From the First
CRS Screen as a Result of Three
Listings of the Same Code-Share
Flight Option
(See figure in printed
edition.)
Note: Request was for travel from Berlin to Chicago departing around
noon on Saturday, December 10, 1994.
Source: GAO's illustration of the Worldspan display.
Triple listings of the same flight option also reduce the efficiency
of travel agents who attempt to identify all options for their
customers. ASTA's Assistant Director, Industry Affairs, emphasized
that several listings of the same flight create more work for travel
agents, who must toggle back and forth between screens to determine
which flight options are new and which are merely repeated listings
of the same flight. Travel agency executives told us that their
travel agents' productivity has decreased because agents have to work
harder to provide the same level of service. The problems they
characterized included the waste of valuable computer screen space
and confusion caused by several CRS listings. They noted that such
listings make it harder for agents to provide customers with accurate
information on which airline is actually operating a code-share
flight. To help alleviate this confusion, travel agent managers at
one travel agency we contacted are conducting monthly staff meetings
in part to discuss with their agents the status of code-share
alliances.
DOT HAS NOT TAKEN ACTION TO
LIMIT THE NUMBER OF TIMES
THE SAME FLIGHT CAN BE
LISTED IN CRSS
-------------------------------------------------------- Chapter 3:3.3
Although DOT proposed regulations in August 1994 aimed at ensuring
that consumers are notified of which airline is the actual operator
before taking a code-share flight, neither the agency's current
regulations nor its proposed rules limit the number of times a
code-share flight may be listed. In its 1992 revision of its
regulations governing CRSs, DOT rejected proposals to impose such
limits. The agency acknowledged that listing a flight several times
may affect the display position of competing flights and make the
display less useful for travel agents, but it noted that individual
CRS vendors are not prohibited from limiting the number of listings
as long as the service is listed at least once under each
participant's code. DOT emphasized that such listings allow each
participant in a code-share alliance to establish its own market
presence.
Most airline representatives we interviewed stated that the double
listing of code-share flights allows an airline to establish a market
presence and preserves the consumer benefits resulting from
code-sharing. Many, however, characterized the listing of the same
flight more than twice as unnecessary and excessive. Northwest's
Vice President, International and Regulatory Affairs, and Vice
President, Government Affairs, for example, stated that Northwest
would support a DOT rule that prohibited more than two listings as
long as that rule preserved code-share partners' ability to each list
a given flight once (double listing). United representatives,
however, cautioned that an unqualified ban on more than two listings
may adversely affect future three-way alliances, such as the possible
United/Lufthansa/Thai Airways alliance, in which all three partners
would seek to list a given flight as their own.
Concerned about the potential for confusion and consumer deception
that may result from triple listings of the same flight, however, the
EU included such a ban in its October 1993 revision of its CRS rules.
These rules limit to two the number of times a code-share flight can
be listed (i.e., once under each partner's code). According to the
Principle Administrator of the European Community's Directorate
General of Transport, the EU acted because of the negative impact of
numerous listings on competition, consumers, and travel agents.
American Airlines and TWA, supported by ASTA, have petitioned DOT not
only to follow the EU's lead but to go farther. In June 1994,
American and TWA filed petitions with DOT asking the agency to issue
regulations that would prohibit the double listing of flights.
Representatives from several other U.S. airlines strongly disagreed
with the petition. Although generally agreeing that more than two
listings of the same code-share flight should be eliminated, these
representatives stated that they considered such proposals as "one
flight, one listing" to be draconian actions that would seriously
undercut one of the rationales behind code-sharing; that is, each
airline partner is able to market the flight as its own product.
According to these representatives, effective marketing requires
appropriate CRS "shelf space."
CONCLUSIONS
---------------------------------------------------------- Chapter 3:4
DOT's policy statement and recent rulemaking proposal to ensure that
consumers are adequately notified before traveling on a code-share
flight represent important steps forward in defining U.S.
international aviation objectives and protecting the flying public.
However, several major issues remain unresolved. First, without
complete and accurate data, DOT cannot adequately monitor the
competitive effects of alliances. Although the agency already
collects data from U.S. airlines based on a sampling of their
tickets, the data do not identify which passengers have taken
code-share flights or, in some cases, which airline actually operated
a code-share flight. In addition, because it does not impose similar
reporting requirements on foreign airlines, DOT lacks key data on
thousands of passengers traveling to and from the United States on
foreign airlines that are flying under code-share arrangements with
U.S. airlines.
Second, the question of whether DOT should, in light of the
Northwest/KLM experience, grant antitrust immunity to other alliances
in markets that allow for significantly increased access for U.S.
airlines has yet to be examined. Because foreign governments as well
as other U.S. and foreign airlines are just now discovering the
success of the Northwest/KLM alliance and believe that much of its
impact is due to antitrust immunity, DOT has a new opportunity to
entice foreign governments to liberalize their accords with the
United States. Without a thorough examination of the Northwest/KLM
experience and a comparison of the benefits of open skies with the
potentially anticompetitive effects of immunity, however, DOT cannot
determine if the use of antitrust immunity as a carrot in other
bilateral negotiations is appropriate or whether Northwest and KLM
should continue to enjoy the protection of antitrust immunity.
Finally, the listing of the same flight option several times in CRSs
limits competition. However, airlines enter code-share alliances
precisely to market another airline's flight as their own, thereby
necessitating two listings. Recognizing these factors, the EU has
limited to two the number of times a code-shared flight can be
listed. Outside of the concern expressed about the potential effect
on possible three-way alliances, we found that general agreement
exists in the airline industry that more than two listings should be
prohibited. However, no such agreement exists on whether to ban the
double listing of flights, and we do not believe that sufficient
evidence exists to justify limiting to one the number of times a
flight can be listed.
RECOMMENDATIONS
---------------------------------------------------------- Chapter 3:5
We recommend that the Secretary of Transportation (1) require that
U.S. airlines, as part of their regular reporting of traffic data to
DOT, identify passengers that traveled on code-share flights and that
they take steps to ensure that they report which airline actually
operated those flights; (2) require, either by regulation or by
conditioning the approval of code-sharing alliances, that foreign
airlines involved in code-sharing alliances with U.S. airlines
report data on their code-share traffic to DOT; (3) direct the
agency's new economic unit to analyze DOT's existing data and the
data obtained above to determine if U.S. consumers and the aviation
industry have been significantly affected in a negative way before
reapproving all strategic code-sharing alliances and any other
alliance that the Secretary deems significant; (4) examine, in light
of the Northwest/KLM experience, whether antitrust immunity should be
potentially available for other alliances in markets that allow for
significantly increased access for U.S. airlines; and (5) prohibit
more than two listings of the same code-share flight in CRSs. In
limiting the number of CRS listings of the same flight option to two,
the Secretary may wish to examine whether an exception should be
granted for alliances with three partners so that each partner may
list a given flight as its own.
AGENCY COMMENTS
---------------------------------------------------------- Chapter 3:6
We discussed a draft of this report with senior DOT and State
Department officials, including DOT's Acting Assistant Secretary for
Aviation and International Affairs and State's Director, Office of
Aviation Policy and Programs. They emphasized that such alliances
produce benefits for partners and consumers. Likewise, they said
that they believe that alliances have increased competition as
alliances compete with each other and nonallied airlines. They noted
that this increased competition has likely led to lower fares and
better service for consumers and stimulated new traffic. However,
they stated that sufficient data do not exist to demonstrate this
possibility or to determine the effects that alliances have had on
fares or will have in the long term.
DOT officials noted that most carriers' views reflected in our report
were consistent with views expressed directly to the Department.
They stated that alliances should be viewed in the context of the
global market forces that are reshaping the industry. They noted
that like other marketing and service innovations, cooperative
arrangements that include code-sharing now have taken root among the
world's major airlines. DOT officials emphasized that the large
number of passengers flying on code-sharing flights is, in their
view, empirical proof of the value of these services to U.S.
consumers. They stated that those benefits will not be available to
flag carriers and citizens whose governments attempt to prohibit or
discourage code-sharing. As a result, they stressed, the challenge
for governments is to be vigilant as to potential harm without
stifling innovation that could be beneficial to consumers.
DOT officials also concurred that additional data are needed to allow
them to better track alliances' long-term impacts on competition.
They stated that the special reports that Northwest and USAir have
begun to provide--and that United will soon provide--will enable the
agency to begin building a fundamental information base early in the
history of these alliances, as they proceed more deliberately with
respect to general reporting requirements.
They said, however, that our recommendations would improve and expand
the agency's existing data and allow the new economic unit to more
effectively analyze strategic and other major alliances. However,
they noted that in some cases code-sharing rights are exchanged in
bilateral agreements and that because of resource constraints, it
would not be practicable for the unit to analyze smaller,
noncontroversial alliances before DOT reapproves them. On the basis
of their comments, we revised our proposed recommendation to call for
the new unit to determine if U.S. consumers and aviation industry
have been significantly affected in a negative way before reapproving
"all strategic code-sharing alliances and any other alliance that the
Secretary deems significant" rather than calling for such an analysis
on "all alliances" prior to reapproving them.
DOT officials agreed that the agency has not determined whether
immunity should be potentially available for other alliances in
markets that allow for significantly increased access for U.S.
airlines. They declined further comment on antitrust issues, stating
that the issue was very sensitive and that the agency was currently
in negotiations with several countries. Finally, DOT officials noted
that American and TWA had petitioned DOT to pass rules limiting the
number of times a flight can be listed in CRSs. They stated that DOT
is currently analyzing the petitions and therefore declined to
comment on our recommendation.
CODE-SHARING ALLIANCES BETWEEN
U.S. AND FOREIGN AIRLINES
APPROVED BY DOT, AS OF DEC. 31,
1994
=========================================================== Appendix I
Foreign airline Year
U.S. airline partner(s) approved
------------------------ ------------------------ --------
Air L.A. Aeromexico 1993
America West Aeromexico 1992
American Airlines Air New Zealand*** 1991
Airbremen GmBH*** 1990
British Midland 1993
Cathay Pacific*** 1990
China Airways 1994
Gulf Air 1994
Lufthansa*** 1991
Malev Hungarian*** 1989
Qantas 1990
South African Airways 1992
Transwede Airways 1994
Carnival Iberia 1993
Linea Aerea Nacional 1992
Chile
Challenge Air Cargo Lufthansa 1992
Continental Airlines AirBC 1994
Alitalia 1994
Air Nova 1994
Air Ontario*** 1993
Ansett New Zealand*** 1992
Scandinavian Airlines 1991
Systems
Delta Air Lines Aeroflot 1991
Aeromexico 1994
Austrian Airlines 1994
Malev Hungarian 1991
Sabena 1993
Singapore Airlines 1992
Swissair 1993
Transportes Aeroes 1994
Portugueses 1994
Varig
Hawaiian Airlines Japan Air Lines*** 1992
Midwest Express Virgin Atlantic 1992
Northwest Airlines Air UK Limited 1994
Ansett Australia*** 1992
Asiana 1994
KLM 1991
Pan Am Ardia Airways*** 1990
Malev Hungarian*** 1988
TWA China Airlines*** 1990
Gulf Air*** 1988
Malev Hungarian*** 1989
Philippine Airlines 1991
United Airlines ALM Antillean Airlines 1993
Ansett Australia 1992
Ansett New Zealand 1993
British Airways*** 1987
British Midland 1992
Cayman Airways 1994
Emirites Air 1993
Lufthansa 1994
National Airlines 1994
Chile, S.A. 1993
Transbrasil 1994
Transportes Aeromar
USAir Alitalia 1991
All Nippon Airways 1992
British Airways 1993
Cayman Airways 1992
Compania Mexicana de 1994
Aviacion 1991
LADECO*** 1994
Qantas
------------------------------------------------------------
Notes:
1. "Year approved" represents the year in which DOT approved the
first code-share arrangement of an alliance. Alliances often entail
subsequent DOT approvals of arrangements to code-share more flights
to additional cities.
2. *** denotes that alliance has been terminated by the carriers
involved.
Source: DOT.
ALLIANCES BETWEEN U.S. AND
FOREIGN AIRLINES BY TYPE, AS OF
DEC. 31, 1994
========================================================== Appendix II
STRATEGIC ALLIANCES
------------------------------------------------------ Appendix II:0.1
1. Northwest Airlines/KLM
2. United Airlines/Lufthansa
3. USAir/British Airways
REGIONAL ALLIANCES
------------------------------------------------------ Appendix II:0.2
4. American Airlines/British Midland
5. American Airlines/Gulf Air
6. Continental Airlines/Alitalia
7. United Airlines/Ansett Australia
8. United Airlines/British Midland
9. United Airlines/National Airlines Chile, S.A.
10. Northwest Airlines/Ansett Australia***
11. TWA/Gulf Air***
POINT-SPECIFIC CODE-SHARES
------------------------------------------------------ Appendix II:0.3
12. Air LA/Aeromexico
13. American Airlines/China Airways
14. American Airlines/Qantas
15. American Airlines/South African Airways
16. American Airlines/Transwede Airways
17. American Airlines/Airbremen GmBH***
18. American Airlines/Air New Zealand***
19. American Airlines/Cathay Pacific***
20. American Airlines/Lufthansa***
21. American Airlines/Malev Hungarian***
22. America West/Aeromexico
23. Carnival/Iberia
24. Carnival/Linea Aerea Nacional Chile
25. Challenge Air Cargo/Lufthansa
26. Continental Airlines/AirBC
27. Continental Airlines/Air Nova
28. Continental Airlines/Air Ontario
29. Continental Airlines/Scandinavian Airlines Systems
30. Continental Airlines/Ansett New Zealand***
31. Delta Air Lines/Aeroflot
32. Delta Air Lines/Aeromexico
33. Delta Air Lines/Austrian Airlines
34. Delta Air Lines/Malev Hungarian
35. Delta Air Lines/Sabena
36. Delta Air Lines/Singapore Airlines
37. Delta Air Lines/Swissair
38. Delta Air Lines/Transportes Aeroes Portugueses
39. Delta Air Lines/Varig
40. Hawaiian Airlines/Japan Air Lines***
41. Midwest Express/Virgin Atlantic
42. Northwest Airlines/Air UK Limited
43. Northwest Airlines/Asiana
44. Pan American/Ardia Airways***
45. Pan American/Malev Hungarian***
46. TWA/China Airlines***
47. TWA/Malev Hungarian
48. TWA/Philippine Airlines
49. United Airlines/ALM Antillean Airlines
50. United Airlines/Ansett New Zealand
51. United Airlines/Cayman Airways
52. United Airlines/Emirites Air
53. United Airlines/Transbrasil
54. United Airlines/Transportes Aeromar
55. United Airlines/British Airways***
56. USAir/Alitalia
57. USAir/All Nippon Airways
58. USAir/Cayman Airways
59. USAir/Compania Mexicana de Aviacion
60. USAir/LADECO***
61. USAir/Qantas
Note: *** denotes that alliance has been terminated by the carriers
involved.
Source: GAO's analysis of DOT's and airlines' data.
MAJOR CONTRIBUTORS TO THIS REPORT
========================================================= Appendix III
RESOURCES, COMMUNITY, AND
ECONOMIC DEVELOPMENT DIVISION
------------------------------------------------------- Appendix III:1
Francis P. Mulvey
Marnie S. Shaul
Timothy F. Hannegan
Howard F. Veal
Deena M. DeVane
OFFICE OF THE GENERAL COUNSEL
------------------------------------------------------- Appendix III:2
Michael G. Burros
EUROPEAN OFFICE
------------------------------------------------------- Appendix III:3
David G. Artadi
FAR EAST OFFICE
------------------------------------------------------- Appendix III:4
James L. Morrison
Robert E. Sanchez
Conor B. O'Brien