01912607188 s3.a) + .oo 0 19sfl Pergmm Press plc
7hmsp1. Ra..A, Vol: 22A. No. 6. pp. 405-417. 1988 Primed in Great Britain.
CODE SHARING, JOINT FARES, AND COMPETITION IN THE REGIONAL AIRLINE INDUSTRY CLINTON V. OSTER, JR. Transportation Research Center, Indiana University, Bloomington, IN 47405, U.S.A.
and DON H.
PICKRELL
Transportation Systems Center, U.S. Department of Transportation, Kendall Square, Cambridge, MA 02142, U.S.A. (Received 20 February 1987; in revbedfom
19 May 1988)
the wake of airline deregulation, virtually all large jet carriers have entered into cooperative marketing agreements that involve shared designator codes and joint fares with commuter airlines. This paper examines the development of these agreements, the role they play in the competitive strategies of major and commuter airhnes, and the implications of these agreements for the structure of the airline industry and for service to travelers to and from small communities. Abdmd-In
lNTRooucnoN:
MARKETWG ALLUNCES
AND
lines involve shared designator codes, all major and most national U.S. airlines now share their codes with one or more regional carriers under such agreemet&l Despite their recent growth, neither joint marketing efforts nor shared designator codes are actually new developments. Commuter carriers and major airlines have long engaged in many of the elements of the current agreements, including coordination of flight schedules to facilitate passenger connections, “joint” fares for passengers making trips that utilize both cooperating carriers’ flights, interlined tickets, adjacent location of commuter and jet carrier gates in hub airports, and automatic transfer of passengers’ baggage. Even code sharing itself was introduced in 1%7 by the Allegheny Commuter System, a series of agreements under which various commuter carriers agreed to take over service to small communities previously provided by Allegheny Airlines (now USAir), with those commuter carriers’ flights identified by Allegheny’s designator code.
AIRLINE CODE SHARING
Cooperative marketing agreements between commuter or regional airlines and large jet-equipped air carriers have become the dominant form of small community airline service within the United States.? Most of these agreements include code sharing, a practice whereby flights operated by a regional carrier are identified by the two-letter designator code of a cooperating jet catrier when listed in the Official Airline Guide (OAG), in computer reservations systems, and on passenger tickets.$ Although not all joint marketing efforts between jet and regional air-
tThis paperuses the terms regional airline and commuter airline interchangeably. These terms refer to airlines that operate fleets of predominately propeller-driven aircraft in scheduled passenger service in short-haul markets (generaUy less than 300 miles in length). Some regional airlines operate a few jet aircraft and some serve a few markets of longer than 300 miles, but even for these carriers shorthaul service using propeller aircraft is the major form of service. At the end of 1985, there were 221 regional air carriers operating in the United States (Regional Airline Association, 1985). $The terms major and national air carriers refer to air carrier groupings established by the Department of Transportation (DOT) for financial and statistical analysis. “Major*’ carriers have annual operating revenues in excess of $1 billion; “national” carriers have annual operating revenues of SlOOmillion or more but less than $1 billion. As of January 1, 1986, the 13 majors are: American, Continental, Delta, Eastern, Flying Tiger, Northwest, Pan American, Piedmont. Reoubiic. Trans World. United. USAir, and Western. Niiteen carriers are classified as nationals: Aircal, Alaska, Aloha, American Trans Air, America West, Arrow, Braniff, Frontier, Hawaiian, Jet America, Midway, New York Air, Ozark, Pacific Southwest, People Express, Southwest, Transamerica, Transtar
(formerly Muse), World, and Zantop. Some of these airlines do not offer scheduled passenger service, concentrating instead on cargo service or passenger charters. As used here, the terms “major airline” and “jet carrier” refer to the airlines in this group that offer scheduled passenger service. OThe practice of sharing designator codes is not limited to agreements between regional and major airlines. One pair of major airhnes shares designator codes with each other, and other major carriers with nationwide routes share their designator codes with air&s that operate jet aircraft over regional route networks. However, this study focuses on code sharing by regional or commuter airlines, and does not discuss these other code-sharing agreements. 405
C. V. OSTER, JR. and D. H. PKX~ELL
406
However, marketing agreements that prominently feature shared airline designator codes did not extend beyond the Allegheny Commuter System until 1978, and code sharing did not spread rapidly throughout the industry until early 1984. The marketing alliances formed since 1984 are intended to provide feed traffic from small communities to a major air carrier at a specific hub airport. They differ from the Allegheny Commuter System in that their focus is not on replacement service. Indeed, the newer arrangements typically involve markets in which the major carrier has never operated its own equipment. Thus, while code sharing between regionals and majors dates back to 1967, the way in which it has been used since 1984 is new. The potential impact of widespread code sharing on the competitive structure of the U.S. airline industry will depend on several issues including: (i) how air travelers respond to any service changes that result from the cooperative marketing efforts that accompany code-sharing agreements (ii) how flight connections under code-sharing agreements are represented in computer reservations systems (iii) whether confusion among air travelers arises from the way in which flights operated under shared designator codes are advertised and (iv) the joint fare policies of airlines that participate in code-sharing agreements. This paper examines competitive strategies of the major carriers following deregulation, and illustrates how marketing alliances that include code sharing represent an extension of these broader airline competitive strategies. It surveys the current scope of activity under code&haring alliances, traces their recent growth, and examines how their widespread use is likely to affect consumer choice in selecting flight itineraries. Finally, the paper uses airline schedules and passenger traffic data to explore the actual effects of code sharing on competition among regional carriers. COMPETITIVE STRA’IECIES OF
THEi MAJOR
AIR CARRIRRS
For major carriers, the development of marketing alliances with commuter carriers is part of a broader postderegulation route strategy to strengthen huband-spoke route networks and to develop new hubs (Meyer and Oster, 1987). Hub-and-spoke operations offer some limited cost savings compared with providing only direct service because of modest scale economies with terminal and baggage handling operations. (Bailey, Graham, and Kaplan, 1985; Morrison and Winston, 1986) Hub-and-spoke systems also allow the use of larger aircraft, because passengers originating from a spoke city and bound for several destinations served from the hub can all be carried aboard the same airplane. Hub-and-spoke
systems were thus particularly attractive because the major airlines entered deregulation with a surfeit of aircraft that were too large to be operated over their historical route networks at costs low enough to be competitive with either newly established jet carriers or the former local service airlines. Of perhaps greatest importance, hub-and-spoke networks also served to emphasize the major carriers’ primary strengths in competition with the lowcost new entrant airlines and the former intrastate jet carriers that were permitted by deregulation to enter the industry and expand. The established major carriers entered deregulation with labor contracts featuring high wage rates and restrictive work rules, which raised their operating costs to levels substantially above those of the new entrant and former intrastate airlines (Meyer and Oster, 1984). Establishing large hubs ailowed the major carriers to counter the competitive disadvantages that stemmed from their high cost structures by emphasizing such strengths as frequent flight departures, convenient connections to many destination cities, and sophisticated computerized reservations systems. Marketing alliances between major airlines and commuter carriers can be most easily understood as part of these postderegulation route development strategies. Commuter feed traffic can be important to the profitability of a larger carrier, as once a carrier is committed to making a flight, additional passengers occupying seats that would otherwise have flown empty add little to the cost of the flight but can increase revenue substantially. For example, between 1978 and 1985 American Airlines had an average load factor of 63.6% and an average breakeven load factor of 61.2% (AMR Corporation, 1985). American’s operating income during that period arose from this 2.4% difference, a difference that translates to about 3.5 passengers on a B-727-200 (the backbone of American’s fleet). Thus, a feeder program that added an average of only one passenger per flight (paying the average fare per mile for American passengers) might have increased operating profits from B-727-200 operations over this period by as much as one fourth. It is no surprise, therefore, that regional and major carriers have long coordinated their schedules to facilitate convenient flight connections at hub airports they jointly serve, or that major carriers sought stronger cooperation following deregulation. Commuter carriers have often been able to use gates on the major carrier’s concourse, sometimes at favorable lease rates, to make connections easier for passengers. Commuters have also frequently engaged in joint advertising with major carriers to stress convenient connections, and cooperating airlines have frequently offered joint promotional fares. Similarly, many commuters have had their flights “counted” in the majors’ frequent flyer programs. In the more openly competitive environment featured by deregulation, it was perhaps inevitable that the major carriers would seek more exclusive rela-
Regionalairline i+istry tions with commuters. Not only was the feed traffic potentially profitable to the ma&@rrier,. butdenying feed traffic to a competitor at a hub could confer an important competitive advantage. Indeed, by restricting access to feed traffic at a hub, an existing carrier may be able to raise the barriers to entry into that hub for another carrier. Raising barriers to market entry has become an important feature of postderegulation competitive behavior in the airline industry. For example, hub dominance at cities such as St. Louis and Minneapolis-St. Paul brought about by the TWA-Ozark and NorthwestRepublic mergers has erected major barriers to entry into those cities. In developing exclusive feeder relations with commuter carriers, major airlines originally seemed to prefer a marketing alliance rather than outright financial acquisition of the commuter carrier or starting their own commuter-type operations. Labor considerations may provide one motive for the preference of alliance over ownership, since part of the reason commuters can serve these short-haul, lowdensity markets effectively is their lower wage structure and more flexible work rules. Although some of the major airlines have been able to establish dual wage and work rule structures, most commuter wages are still below the lower or “B-scale” wages of the major airlines. A second reason that major carriers preferred marketing alliances may have been that they afforded considerable influence over the commuter’s operations at virtually no risk to the major carrier, since the agreement could be quickly terminated if any of its elements later proved unsatisfactory. More recently, however, several.of the major airlines have purchased a financial interest in some of their commuter code-sharing partners. While a financial interest carries some risk and may complicate labor relations, it offers greater control over the commuter carrier and can prevent the potential disruption caused by a competing major carrier attempting to lure the commuter away by offering a better marketing alliance agreement. THE GROWIll
OF MARKETING
ALLIANCES
The growth of code-sharing marketing alliance agreements has been so rapid that even compiling an up-to-date inventory of such practices is difficult. While code-sharing agreements were relatively uncommon until 1984-only 16 were apparently in effect at the beginning of that year-there were 52 such agreements by mid-1986. This growth in codesharing agreements coincided with an increase in industry concentration. Between 1984 and 1986, for example, the largest six carriers increased their domestic market share from 65% to nearly 88%. By contrast, between 1978 and 1983 market concentration had steadily declined (Meyer and Oster, 1987). Table 1 summarizes marketing agreements as of June 1986 that entail code sharing between a jet-
407
equipped air carrier and a regional airline serving pn$@f &e jet carrier&&mestic hubs. As the table suggests, code-sharing ‘agreements have become a prominent feature of both the route structures of individual air carriers and the nationwide air service network. Under the 52 separate code-sharing agreements listed in Table 1,16 jet airlines and 46 regional air carriers serve passengers traveling between 39 different major U.S. airports and well over 2000ther points (typically FAA-designated small hub or nonhub airports).t Virtually all of these markets exemplify the “hub-feeding” type of routes to and from small communities commonly served by regional air carriers, although regional carriers also serve a few routes with substantial nonconnecting traffic under these agreements. Table 2 illustrates how rapidly the number of codesharing marketing agreements between regional and major air carriers has grown. The number of such code-sharing alliances remained modest until 1983, with most of the activity accounted for by the eight code-sharing agreements of the Allegheny Commuter System. Although several new code-sharing agreements were inaugurated during 1983 and 1984, by the beginning of 1985 there were still fewer than half the number of such agreements in effect than would be found only 18 months later. Table 3 illustrates the contribution of code-sharing agreements to the post®ulation restructuring of major carriers’ domestic route networks. The seven major carriers listed in the table have entered codesharing agreements at eight primary hubs with regional airlines that serve those airports with frequent flights to and from 115 surrounding points. The 13 participating regional carriers serve nearly 38% as many points, and offer more than 48% as many daily departures as do the major carriers themselves, although each of the points enplanes fewer passengers and the departures are in smaller aircraft. While several major airlines have entered into marketing alliances with two regional carriers serving a single hub, the pattern of each major carrier participating in only one such agreement per hub appears to be more common. These marketing alliances have also become an important feature of regional air carriers’ operations and marketing activities. As Table 4 indicates, by tThe approximately 500 U.S. cities that receive regularly scheduled commercial air service are classified by the Federal Aviation Administration (FAA) as large, medium, small, or nonhuhs on the basis of the relative volumes of commercial air passengers that use them each year. Large hubs are defined as those at which 1% or more of the nation’s commercial air passengers are enplaned (that is, begin or change flights); medium hubs are those enplaning 0.25-0.999% of total annual air passengers; small hubs are those at which O.OS-0.249% of total passengers enplane; and nonbuhs are those enplaning kss than 0.05% of total commercial air passengers. During 1984, 26 U.S. cities were classified as large hubs, 36 as medium hubs, and 59 as small hubs, while the remaining 287 were designated as nonbubs.
C. V. OSTER,J~.andD.H.
408
PICKRELL
Table 1. Code-sharing agreements between jet-equipped carriers and regional airlines (agreements in effect or announced as of June 1.1986) Jet-Equipped
Month/
Affiliated Regional
Connecting
Year n
Alaeka
Air Pat Ryan Air Rering Air ERA Helicopter WB Flying Service Wark Air
3/82 4/03 4/83 g/e3 4/85 10/85
Metro Airline8 Chapparal Airlines
11/M 12/84
AVair
5/85
Sirnnone Airline‘ Air Widweet Command Airway8
4/86 s/es 4/86 4/06
Wing8
6/86
Weet
Airline8
Served NA NA NA
Anchorage Anchorage Noel8 Anchorage Glacier Bay ;z',t;gham King
American
# Of pte
Hub
NA
WA UA NA NA
salron
11 2 2 11 6
Dallae/Pt. Worth Dallae/Pt. Worth San Antonio Waehington, D.C. Charlotte Raleigh-Durham Chicago Naehville Roeton New York Lo‘ Angelo‘ San Francieco
1: 6 NA NA 12 15
Continental
Royal8 Airlines Pioneer Airlineel
s/e3 %/05
Houston Denver
19 9
Delta'
Atlantic Southeaet Afrlinee Rio Airway8 Carair Bueineee Expreee
S/84
Atlanta Hemphim Dallae/Ft. Cincinnati Boeton New York
21 9 7 18 4 6
Eastern
Metro Expreee Atlantie
Expreee
Air Midweet Bar Harbor Airlines
Precision
New
York Air
Airline8
Colgan Airways Gull Air
5/84 s/04 6/86
10/83 4/04 12/85 r/as ll/i35 3/86
4/06
11/H Z/86
Houston Atlanta SanJuan Atlanta Charlotte Kaneae City Boeton New York Philadelphia xiari
12/84
Xeeaba
7/05
Big Sky Pincher
Brothers
Z/86
6 9 WA 4 5 18 NA
Tampa Boeton
E NA NA NA
New York Philadelphia
NA NA
Waehlngton,D.C.
6 NA NA
Boston
niami Northweet
Worth
Winneapolie, St. Paul Billings H8l.M Detroit Cleveland
16 a 3 E
t&ark3
Air Widweet
9/85
st. Lauie
13
Pan -erican'
Raneome
6/86
New York Waehington
5 4
People Expreee
Britt
Piedmont5
Heneon
Airline8
Airway8 Aviation
CCAir6 Jetstream International Britt Airway8 Brockway Air
NA
l/86 11/63 5/85 5/85
Baltimore Charlotte Charlotte
11
11/M 11/85 3/86
Baltimore Deyton Syracuee
2 11 NA
:
‘I-j;*;*:._
Regional airline h~%~~try ~ Table
Jet-Eguippad
Affiliated Regional ISI
R8public
Sisson8
1 (continuid) Wonth/ Yom n
Airlines
Phoenix Air (aka rz$;;;:
4/05
7/85 12/85
TWA
Ra8ort
United
Air Wiscon8in
USAir
Winnoapoli8 St. Paul
Horizon Air Wontair Air Wiacon8in
Chicago Saattlo San Francisco Washington, D.C.
Pocono
7/68
Z/69
New York Saltinoro Philadelphia Pitt8burgh
6/70 3/73
Philadelphia Philadelphia
Airlina8
Airway8
Jarsoy Airway Suburban Airlines* Pennsylvanin Airline8 Chautauqua Airline8 Air Kentucky skyW.ot (aka Wo8tern
lo/73 8/74 lo/80 4/86
Expro88) Souroo:
1 6
WA
24 11 9 3
a 4 6
Pittsburgh Pittsburgh Louisville Lo8 Angel.8 Salt
various
14
7
st. Louis
7/.95 7/85 7/85 6/86
Southan
Worntern
Dstroit Winnoapoli8/ St. Paul Xuaphis
S/S5
Crown
Air
# of pts Served
Connecting Hub
Dako
imsue8;
NA City
NA
w,
Renewal of larlior agreement, which expired with bankruptcy f Continental Airlin.8 in 1983. Code-sharing agrournt with Ran8on Airlin.8 torxinatod with ~cguisition of Ranmono by Pan As affactive Juno 1, 1986. Coda-8haring agroenent with Gull Air terninatod during cabor, 1985. Coda-sharing agrouent with ERpiro Airlines torninated during Way, 1986 with acguimition of Empire by Piadnont. Pan An purchased Ran8ono during April, 1986. Code-sharing agrauont with Tran8 Air torminatod on April 19, 1986; coda-sharing agronent with Enpirs Airway8 torninatad on Way 1, 1986, when Empire was absorb& into Piodnont 8ymten hollowing its acgui8ition by Pisdmont. Wane changed from Sunbird Airlines on January 1, 1986. Code-sharing agraemont with Fischer Brothmr8 Aviation ierninatad during January, 1986. Acquired by USAir but continues to op8rato ind8p8ndontly under coda-sharing agroonnt.
P
mid-1986 16 of the 20 largest regional air carriers operating within the continental United States (these 20 carriers together account for nearly three-quarters of all regional airline passengers in the contiguous 48 states) participated in code-sharing agreements with major airlines. By the end of 1986, virtually all of the largest 50 regional carriers operating in the contiguous 48 states either had marketing alliance agreements or were partially owned by jet carriers. THE EFFECl-8 OF MARRETING ALLIANCES CONSUMER
ON
CHOICR
The jet carrier’s principal objective in entering a marketing alliance is to increase the number of travelers to and from small communities that use the jet carrier’s flights for the long-haul portion of their trip.
Many of the aspects of marketing alliances are designed to make it more convenient for a passenger to connect between the jet carrier and its commuter alliance partner. Features such as close schedule coordination and adjacent gate space at the connecting airport certainly add to consumer convenience, although these practices predate the recent growth of exclusive marketing alliances. The key element that distinguishes the current marketing alliances from the previous, less formal practices in the industry is code sharing. While many aspects of marketing alliances clearly benefit the consumer, the effects of code sharing itself on the consumer are less obviously beneficial and will depend on three interrelated factors. The first is the treatment of flight connections under code-sharing agreements in the most widely used computer re-
410
C. V. OSER, JR. and D. H. PICKRELL Table 2. Growth in number of code-sharingagreements between jet-equippedcarriers andregionaiairlincs # of Alliancom
# of Participating
X of Najor
Carrior8:
Date l/l/83
10
4
10
7
l/l/84
16
6
16
9
l/1/85
23
9
22
17
6/l/85
30
11
28
22
l/1/86
41
14
36
33
6/l/86
52
16
46
39
Sources:
-, various airline
i88uomt
and
per*oniel.
varioue i88ueer w, comereations
telephone
sexvations systems. Second is the extent to which
consumers understand the nature of service provided under code-sharing relationships, and third is the practice by some major airlines of making joint connecting fares available only to their code-sharing partners. Computer reservation systems
Although there is some variation in the precise manner in which different computer reservations systems (CRS) select flights for display on travel agents’ computer terminals, most work along the same general lines. In response to a travel agent’s request for an origin and destination, departure date, and desired departure (or arrival) time, the
with
computer reservation system selects and displays a list of flight alternatives on the agent’s computer terminal screen. ‘Epically three connecting flight alternatives can be displayed on the computer screen at one time. Each group of flight alternatives displayed together is known as a “screen.” Flight alternatives are ranked on the list according to some combination of elapsed travel time and how closely the time of the flight matches the departure time requested by the traveler. The sequence in which flights are presented on the computer screen appears to be important in the traveler’s choice of flights, since between 70% and 90% of computer reservations system bookings appear to be made from the first “screen” of flights displayed, while half are made from the first line displayed (Delta
Table 3. Service under code-sharingagreementswith regional airlines at selected major air carriers’route hubs -Air Major
Primary
American
DallasPt. Worth
Continental
Denver
Delta Ea8tern
Atlanta
Northwe8t
Average Daily
Average # point8
Daily
X points
268
68
2
77
13
91
28
1
22
9
Rouston
92
28
1
40
10
Atlanta
328
61
1
102
21
300
68
2
66
13
Uinneapolia Et. Paul
99
27
1
37
16
Piedmont
Charlotte Baltimore
140 53
44 26
2 2
31 55
11 11
Republic
Detroit
91
40
1
52
14
Total, 7 Major
Carriere
1,192
390
13
482
115
Source:
Nwombar 1, 1985, and EM
411
Regional airline industry Table 4. Partieip@ion in de-sharing agreements among the u) @rpt regional airlines operat&g+ih the contirMta1 United States (tank& by 1985 enplanements) und*r Code-Sh8ring
Major Hub Air Wisconsin1
8.08
chic.go
United
Washington, D.C. (Dullem)
United
Rapire Airlinw2
6.05
Pan Anrican Nov York (JPR) Nov York (La Guardia) PiOdBOnt
Britt Airways3
5.82
Chicago Dayton Indianapolis
Retro Airlinea
5.28
Atl8nta San Juan Dallas-Pt. worth
Atlantic aouthmast Airlines
4.55
Atlanta Reaphia
Dolta Dm1ta
Horizon
Airlines
4.3s
soatt1m
United
Hanson Aviation
4.08
Raltimora Charlotto
Piedmont
PM’
3.6%
Air Nidwemt
3.5t
Sinon
3.42
Detroit Chicago
Republic Anrica
3.2t
Salt Lake City
Lom Anqmlom
Wo8t.m WoStorn Dolta
Airline’
skyLwt
LI.stom
Ozark Anricmn
2.9*
Cincinnati
Ransom Airlines
2.0s
Now York Washington,
2.St
Aspon
NolIO
Now York niami Boston T-pa
Coruir
Rrwkway
Piodmoe
D.C.
Pan Anrican Pan Am8rican
Syracumo Binghamton
Piadmont Piod8ont
2.6t
Donvor
Nono
Rusinosm Rxpre&
2.3s
Philadelphia Daltimorm
Nono
Ponnmylvania Airlines
2.28
Pittmburgh
USAir
Rzza;~~in
2.0s
Dmvar
Non0
Royala Airlines
1.98
Houeon
Nff Orluns
Continent81 Rona
Philadelphia
USAir
Air~ay8~
Suburban Airlinen
sources:
1.85
Regional Airlinu lumociatfon, m -, R8y 1, 19.61 s various issuemr and published air carrier
lchmdulu:
Includem lnplanements for Rimsim8ippi Valley Airlines, which mm aaquirmd by Air Wimconmin during 19911. ; Acquired by Pihont Airlinmm and urqmd into lattar'm lymtu SffMtiva Nay 1, 1966. Aoquired-by.Boplo Bxprosm Airlinem duri 1996. Cad+mharina aanamnt with Amsrioan Air1Y nu for Shnm
1
in joint marketing agroamnt vitb united domm not inalude ,&a-sharing. Code-sharing agreannt with tilta to taka lffwt during Juno 996. 3 Tox88 Air Corporation announced plan to lcqUir& Rocky Rountain Ahvaym on May 2, 1996. Participates
tirlinmm, but lgromnt
412
C. V. OSTER,JR. and
Air Lines, 1986). Of course, these percentages partly reflect the close match between CRS systems’ and travelers’ criteria for ranking flight alternatives. However, they also partly reflect agents’ and travelers’ natural tendency to stop seeking information or additional alternatives when doing so produces little or no apparent improvement. Recognizing this tendency, moving the positioning of a flight upward or downward in the ranking of alternatives that a CRS constructs can have a decided effect on the probability that is presented to the traveler and, if presented, accepted as the choice. In determining these rankings, travelers are assumed to have an ordered preference for nonstop flights, single-plane service, online connections (those between tlights operated by the sume airline), and interline connections (those requiring a traveler to transfer between flights operated by different carriers). In an attempt to incorporate these preferences in the ranking procedures, one of the major airline computer reservation systems adds arbitrary but successively larger elapsed time penalties to flights requiring online transfers and interline connections. The other major computer reservation system simply ranks and presents all online connecting flights and connections between flights of code-sharing partners before displaying any flights that require interline connections. These two systems are used by 57% of the nation’s CRS-equipped travel agents and together they accounted for three-quarters of the 1986 passenger bookings by CR&quipped agents. The preference given nonstop and single-plane nights over connections is not an important concern in assessing code-sharing, since the passenger itineraries over which code-sharing may have an infiuence all entail connections. The critical issue in evaluating the treatment of code-shared flights is the relative desirability of “true online,” “code-shared online,” and interline connections, particularly where code-shared online connections involve flights operated by different carriers but listed under a single designator code and thus treated in the computer as tme online connections. The basic questions are why travelers value online connections over interline connections, and, more important, to what extent code-shared connections share the desirable attributes of traditional online connections. Air travelers presumably prefer online connections to interline connections because of both ease of connection and perceptions about likely treatment in the event of a missed flight. An online connection usually involves a shorter walk than an interline connection because of the proximity of a single carrier’s gates to one another, a factor that takes on added importance if the timing of the connection is close or if the inbound flight is late. Many travelers may also believe that a carrier is more likely to hold an outbound flight for connecting passengers from its own flights than from the flights of other carriers. Baggage handling procedures for online connections may reduce the likelihood that luggage will be mis-
D. H. PICKRELL placed, and air travelers may prefer the clearer responsibility of a single carrier in the event of lost luggage or a missed connection. Yet incorporating these preferences into the rankings of flight alternatives in a computer reservations system presents two difficulties. First, with the spread of code sharing, newly redesignated online (i.e. “code-shared” online) connections may not necessarily offer these desired attributes. For example, some online connections between code-sharing carriers involve longer walking distances than some interline connections with non-code-sharing regional carriers serving the same airport. For example, prior to the Northwest-Republic merger, the gates of Republic Airlines’ code-sharing partner at Detroit, Simmons Airlines, are not even on the same concourse as Republic’s gates, while Delta and its code-sharing partner (Ransome Airlines) occupy separate terminals at Washington, D.C.3 National Airport. Second, the features offered by online connections (direct baggage transfers, shorter distances between boarding gates, and through rather than interline fares) probably make them preferable to interline transfers only if both connecting flights are approximately equally convenient to the traveler. Even among travelers willing to accept a delay or alter their schedules to take an online connection, there is likely to be wide variation in how large a delay each would tolerate in exchange for the presumably greater convenience afforded by an online transfer, and no simple arbitrary time penalty can reflect such variation. Moreover, some travelers might be less willing to accept a given delay for a code-shared online transfer than for a true online connection. Thus, it is particularly difficult to estimate average time penalties that accurately reflect travelers’ preferences for use in ranking interline flight connections in computer reservations systems. In fact, such penalties may be redundant because any added ease of online connections due to gate proximity is already reflected in shorter “minimum legal connecting times,” which are calculated to include the time needed to walk between gate areas and are already incorporated in computerized procedures for selecting connecting opportunities. The potential problem arises when an interline connection more closely matches the traveler’s needs in terms of elapsed travel time and proximity to the desired departure time, but is relegated to a lower status than code-shared online connections in CRS displays either arbitrarily or as a result of the time penalty the interline connection is assessed. In determining the sequence in which flight alternatives are displayed to the travel agent, the most widely used airline CRS penalizes interline connections the equivalent of 90 minutes of elapsed travel time while the other widely used system displays interline connections only after displaying all online and all codesharing connections. In other words, a connection between an independent regional airline and a major jet carrier must be more than 90 minutes faster in
413 Regionalairline industry ment. A traveler booking a flight on a major carrier terms of elapsed travel time to be ranked by one computer reservations system *m,a connection ixpeaa; a clertain lev&&,+omfort to be provided by between another regional airline and its code-sharing the sea&g density, noise d vibration levels, cabin partner, and can never be ranked above any such configuration, and pressurization that are typical of large jet aircraft. Air travelers have also come to code-sharing connection in the other major CRS. Such penalties make it difficult for travelers to expect on-board food and beverage service, room to learn of flight opportunities involving independent work in flight, and space for carry-on luggage. Of regionals even if those opportunities more closely course, any potential for deception can be minimixed match travelers’ needs. It is difficult to estimate how if travelers are informed when service will actually often a traveler who selects a code-shared online be operated by a regional air carrier, at least for connection remains unaware of a better interline those who are familiar with the levels of on-board connection, but because the typical commuter flight comfort provided by the smaller aircraft these caraverages little more than one hour in duration, it is riers operate. (This is the rationale underlying the essentially impossible for an independent operator September 1985 disclosure rule issued by the U.S. to make up this difference on all but the very longest DOT, which is discussed in more detail below.) With flights operated by regional carriers under flights (and in one major CRS not even this would cause it to be ranked above even the slowest code- code-sharing agreements, the degree to which exsharing connection). Thus, almost everywhere a pectations about comfort are met depends strongly faster interline connection is available, travelers will on the specific aircraft used. For example, many of the larger regional aircmft (those seating 40 or more) remain unaware of it. As an illustration, the Appendix shows four mar- offer levels of passenger comfort that compare fakets where interline connections offering shorter or vorably to those provided by narrow-body jets in equal elapsed travel times appeared after the first terms of noise, seating density, and on-board amenscreen, which was filled by online connections re- ities. However, in some of the smaller commuter quiring longer travel or waiting times. Consider the aircraft (those seating 19 or fewer passengers), seatchoices presented to a passenger wishing to go from ing is less spacious, noise and vibration levels may Cedar Rapids, Iowa, to New York’s LaGuardia air- be higher, and on-board amenities requiring a flight port and requesting to depart Cedar Rapids at about attendant are notably absent. In the new generation 8:00 A.M.As the Appendix indicates, a flight oper- of aircraft seating 30 to 40 passengers, which are just ated by Britt Airways leaves at 8:05 A.M. with a 50- now being placed in service, comfort levels fall beminute connection in Chicago to a nonstop Ameri- tween these extremes, although they are probably can Airlines flight to New York-a flight combi- much closer to those found in the larger turboprop nation that almost exactly matches the passenger’s aircraft. During 1984, aircraft seating 40 or more request. However, this flight combination was not passengers accounted for 29% of passenger-miles among the first three presented to the travel agent ‘carried by regional airlines, while 44% were carried (i.e. on the first screen), but rather could be found on aircraft seating 19 or fewer passengers (Regional only if the travel agent requested the second screen. Airline Association, 1984). Such differences in comfort, however, may not be The first screen presented three “code-shared online” connections, all of which were substantially less a critical issue. First, most regional airline flights are convenient but were presented first because of the typically short, often involving less than an hour in preference given to these flights over interline con- the aircraft, and comfort levels and on-board amennections. While a traveler can of course ask the agent ities are less important to most passengers on short to examine the full range of choices, few travelers flights than on longer flights. Second, the markets are likely to be aware of the penalties or their po- in which these code-sharing alliances operate typitential effect; in fact, surprisingly few travel agents cally do not have sufficient traffic to support jet serseem to be aware of the time penalties and their vice. Thus, code sharing is generally not misleading effects on flight rankings. the consumer into flying aboard propeller aircraft when jet service is available. A deception issue may arise, however, if a traveler is induced to choose a Consumer deception There are two ways in which consumers might be different routing or travel mode based on false exdeceived by code-sharing arrangements. The first, pectations of jet service. Travel agents should be able as implied by the discussion above, could occur if a to help passengers avoid such problems, however, traveler selects a code-shared online connection because codes for aircraft types are included in the when a more advantageous interline connection was screen displays provided by computerized reservaavailable but remained unknown to the traveler be- tion systems as well as in published schedules. A consumer may also expect the safety record of cause of the interline penalty or ranking protocols imposed by computer reservations systems. Second, a jet carrier when a flight operated by a regional the use of shared designator codes in published and carrier under a major airline’s designator code is computerized schedules may lead some travelers to booked. The regional industry as a whole has not believe that flights operated by regional airlines will amassed as favorable a safety record as the major instead be operated by a major carrier in jet equip carrier portion of the airline industry, although the
414
C. V. OSTER,JR. and
largest (and most experienced) regional carriers as a group have a safety record virtually the same as the jet-equipped carriers (Meyer and Oster. 1984). Thus, a passenger boarding an airplane operated by a large, experienced regional airline has no more reason to be concerned about safety than when boarding a large jet aircraft operated by a major airline. And, as Table 4 illustrated previously, the larger, more experienced regional airlines have to date been by far the most active participants in codesharing alliances among the regional carriers. Possible concern might arise if a passenger were induced by its participation in a code-sharing agreement to choose a small, inexperienced regional carrier over a more experienced independent regional airline, although there are still many individual safe operators among the smaller regional carriers. On balance, any impact of code-sharing alliances on safety performance cannot yet be assessed. Regional carriers participating in the oldest of these arrangements, the Allegheny Commuter System, have recorded excellent safety performance, but these carriers were typically experienced when they joined the Allegheny System. If widespread codesharing results in the operation of more standardized flight equipment, greater financial stability, and better-trained personnel, for example, then safety might improve. However, if code sharing stimulates very rapid growth of relatively small, inexperienced regional airlines, then there might be cause for concern about safety. The experience with code sharing is simply too limited at this point to predict with confidence the impacts of any of these possible developments on safety. In any event, it is important to keep in mind that air travel on a regional airline is typically much safer than travel by auto, and that it is auto travel rather than travel via a major airline operating jet aircraft that represents the principal transportation alternative in most markets served under code-sharing agreements. In response to concerns about possible consumer deception, the Department of Transportation issued a final rule on September 17, 1985, requiring major air carriers to provide notice of code-sharing arrangements by (1) identifying in published schedules (including the OAG) those flights for which the carrier providing the service is different from the airline code displayed for the flight, (2) directly informing travelers in any oral communications (such as in telephone calls to airline reservations agents) whenever a flight will be operated by an airline different from the carrier whose code is used, and (3) providing frequent periodic notice in advertising media of the existence of code-sharing relationships, including the identities of carriers actually providing the service. The notice required by this rule is likely to reduce the confusion that has thus far surrounded codesharing alliances. However, the rule relies on voluntary efforts of travel agents to inform consumers of code sharing. Thus, even with the DOT rule, some consumers may not understand the nature of a codeshared flight until just prior to boarding.
D. H. ~CKRELL
Joint fares and code sharing
Prior to deregulation, the Civil Aeronautics Board (CAB) required that all trunk and local service carriers establish joint fares for all connecting routes generating at least 50 origin and destination passengers per year. Under the Airline Deregulation Act, regional carriers were eligible to join this mandatory joint fare program if they wished, and virtually all chose to do so (Meyer and Oster, 1984). The mandatory joint fare program expired with CAB ratemaking authority during 1983, leaving regional airlines to negotiate joint fares individually with the major carriers as many had done prior to 1978. At first, the principal change from the mandatory joint fare program to the negotiated arrangements was a division of revenues more in line with the ratio of local fares (a so-called rate-prorate basis), although there was variation among the agreements. A more recent feature of joint fare arrangements in code-sharing alliance markets has been the withdrawal by major airlines of their joint fare agreements from competing or independent regional carriers. By offering a joint fare to one regional airline in a market but not a second, a major carrier can effectively make the price faced by passengers for travel on one regional airline significantly higher than the price of using another. For example, a passenger traveling from Flint, Michigan, to Boston in September 1985 could connect to the same Republic jet flight in Detroit from either a Republic Express flight (operated by Simmons) for $200 or from a Allegheny Commuter flight (operated by Fischer Brothers using the same type of equipment) for $285-42.5% more. Such resulting price differentials can induce a shift of some traflIc to the favored regional airline. Indeed, regional carriers from whom joint fares have been withdrawn have reported dramatic shifts of traffic to their competitors who remain able to offer joint fares. Typically, because of advance bookings these shifts are not seen immediately upon withdrawal of the joint fare but phase in over a four- to six-week period. A regional carrier from whom a joint fare was withdrawn could, of course, try to compete by simply cutting its fare to absorb the price differential created by its competitor’s continued use of joint fares with a major airline. However, such a cut might drive the regional carrier’s fare below its costs, resulting in financial loss. CODE SHARING
@dD COMPETITION AMONG
REGIONAL AIIUJNES The effects of marketing alliances on regional airline competition are illustrated by Tables 5 and 6. Table 5 summarizes service changes in 22 feeder markets involving nine connecting hubs in which there was competition between a code-sharing alliance partner and an independent regional in July 1985. These markets were found by searching through the Oj$cial Airline Guide for such commuter competition in markets involving each of the
415
Regional airline industry Table 5. !%xviccchangesin code-sharingalliance airline markets competitionbetween alliancepqqjq$%vand indepe?dents (,q'pel@yaircraft departq@ .:' Alliance Partners Total July
(22 markets):
Sewice 1965
October
1986
Percent
Change
Independent July
July
Change
Withdraw
Percent
change
102
18
120 -32.2
16
30
46
46
18
64
+ 187.5
-40.0
+39.1
73
56
131
56
0
56
(17
1985 1966
177
(5 markets):
Remains
October
66
-79.6
October 1966
Indopmdmt
09
+14.6
1985
Percent
Total Departurea
Independent8
markotm):
-100.0
-23.3
Source:
, Volame
-57.3
11, Nuaber
19 and Volrue
major connecting hubs in the continental United throughout the period (the middle portion of the States in July 1985. While there may be a few such table) and the more common situation of 17 markets markets not represented in the table, the 22 markets where the independent withdrew from the market are intended to constitute the universe of such mar- (the bottom of the table). Where the independent kets in the United States at that time. As the top remained, the total level of service was higher in part of the table indicates, between July 1985 and October 1986than in July 1985 although, as the table October 1986, the alliance partners increased de- ‘indicates, the independents* share of departures fell. partures by 14.6% while the independents reduced Where the independent withdrew from competition, departures by 79.6%. As a result of these changes, the total level of service fell, as would be expected. these markets had over 32% fewer departures in More significantly and perhaps more ominously, the October 1986 than in July 1985. alliance partner’s level of service was lower after the These markets reflect two situations: 5 markets independent withdrew than it had been when there where the independent remained in the market was competition. In other words, when the indeTable 6. Servicechangesin code-sharingalliance airline marketscornpctitionbetween alliance partners (weekday aircraft departures) Total Service to the Nub Continued July
Competition
1985 1996
Portent
Change
One Partner
Source:
170
October
July
(15 urket8):
161
Remainm
+6.5
(14 urkete):
1995
October
1966
Percent
Change
139 66 -52.5
Volume
11, N-r
19 and Volume
pendent left the market, not only did the marketing alliance partner not increase departures to compensate for the loss of the independent’s flights, but the alliance partner reduced its flights when no longer faced with competition. The role of code sharing or even the role of marketing alliances in these competitive battles cannot be easily quantified, yet it is suggestive that in this sample of markets, none were found where an independent drove an alliance partner from the market. Even more suggestive is the experience of Britt Airways in the Columbia, Missouti to St. Louis market. Between January 1985 and July 1985 Britt shared the market with Resort Air, another regional carrier, and an occasional jet flight by Ozark (for which passenger data are not available). During this period, Britt carried an average of 78% of the regional airline passengers with Resort Air carrying the remaining 22%. In August 1985, Resort Air became a code-sharing partner with TWA and B&t’s share fell from 76% in July to 43% in September. In October 1985, Oxark withdrew its jet flights and substituted code-shared service by its alliance partner Air Midwest with the result that by November B&t’s share had fallen to 11% while Resort Air and Air Midwest had shares of 38% and 59%, respectively. Britt was a substantially larger regional than either Resort Air or Air Midwest, yet without a code-sharing alliance and access to joint fares it was apparently unable to compete effectively. The competitive advantage a marketing alliance appears to give an alliance partner over an independent points to a drastically diminished role for independents, and raises the question of the results of competition between two alliance partners. Table’6 summarizes service in a sample of 29 markets where two alliance partners competed in July 1985. The top portion of the table describes service in the 15 markets where there was still competition in October 1986. These markets had involved hubs intensely contested by two major carriers, such as Atlanta (Eastern and Delta), Chicago (American and United), and Dallas (American and Delta). As the table indicates, service was slightly higher in October 1986 in these markets. The bottom of the table summarizes the 14 markets where only one alliance partner remained in October 1986. These were typically markets involving hubs where one jet carrier had come to dominate the hub and the other had reduced its presence, such as Houston (Continental over Eastern), Charlotte (Piedmont over Eastern), and Memphis (Northwest over Delta). These markets where only one regional remained had a 52.5% decrease in departures between July 1985, when there was still competition, and October 1986. THE FUTUREOF
H. PICKRELL lowing deregulation and established code-sharing alliances with regional airlines to provide feed traffic for these hubs from nearby small communities. Code-sharing alliances appear to have helped the major catriers develop and strengthen their hub-andspoke route systems and raise barriers to entry into their hubs. While close cooperation between commuters and major carriers has long been a feature of the airline industry, two elements of the more recent agreements seem to break with past practices without any clear benefit to travelers. The first of these is the computer reservations system treatment of connections between flights operated by code-sharing carriers as online connections. The preference in the reservations systems given “code-shared” online connections over interline connections can make it difficult for consumers to learn of more convenient interline flight alternatives (measured in terms of total elapsed travel time or proximity of scheduled to desired departure or arrival time). In combination with the withdrawal of joint fares from non-code-sharing partners, the second new element of recent agreements, code sharing makes it difficult for an independent regional carrier to compete with the alliance partner of a major carrier. As the data reported in Table 5 indicate, the typical experience when an independent competes with an alliance partner is for the independent to withdraw from the market. Many independent regional carriers apparently feel that they have been left only with the choice ‘of either becoming aligned with a major carrier or suffering serious losses of passenger traffic and market shares to their affiliated competitors. Although code sharing is an understandable competitive strategy, the added benefits it offers to consumers beyond those already provided by cooperative arrangements that existed before the recent growth of code-sharing practices are unclear. Close coordination of schedules, leasing adjacent gate space in hub airports, and joint marketing efforts have long been features of cooperation between regional and major carriers, and these features continue to yield substantial benefits to consumers. However, these benefits would remain the same whether they stemmed from the informal cooperative marketing arrangements that have historically characterized relations between commuters and major airlines, or from more formal code-sharing agreements. Code-sharing agreements may also be accompanied cost savings to regional carriers if the majors provide services such as cost and revenue accounting, employee training, bulk purchases of fuel and even aircraft, and other services where there may be some scale economies. Through these arrangements, the major carriers may also have the incentive to extend their promotional fares and advertising campaigns to smaller communities, thereby making it easier for small community residents to learn of service and fare opportunities. The jet car-
C. V. OSTER, JR. and D.
416
REGIONAL
AIRLINE
COMPETITION
Virtually all large jet airlines have developed clearly defined hub-and-spoke route networks fol-
Regional airline industry tiers may also exert influence
on thecommuters to improve their customer service. ,De$pite these p?ptential gains from close affiliatio??%de shar&does not appear to offer separate benefits to travelers, but instead, in combination with selective withdrawal of joint fares, may eventually reduce competition in the regional airline industry with an accompanying reduction in travel alternatives for travelers to and from small communities. REPERENCES
AMR Corporation, Texas.
(19&?) 1985 Annual Report. Dallas,
417
Bailey E. E., Graham D. R. and Kaplan D. P. (1985) Dere&ating rhe Airlines. The MIT Press, Cambridge, :i”&&&&h&etts. ‘:’ .* ‘i, Delta Au Lines (1986) Third-Party Compfaint, Before rhe UnitedStatesDepartmentof Tmnsportation,Washington,
D.C., Docket 44094.
Meyer J. R. and Oster C. V. (1984) Deregulation and rhe New Airfine Entrepreneurs. The MIT Press, Cambridge, Massachusetts. Meyer J. R. and Oster C. V. (1987) Deregulation and rhe Future of Inter&y Passenger Travel. The MIT Press, Cambridge, Massachusetts. Morrison S. A. and Winston C. (19%) The Economic Effects of Airline Deregzdarion.The Brook& Institution, Washington, D.C. Regional Airline Association (19&1) 1984 Annual Report. Washington, D.C.
APPENDIX
Examples (All
of Computer Reservation Display of interline Connections cases were taken from American Airlines’ SABRE System)
Flint, Michigan to Requested Departure
Miami Time:
7:oo
An
Date Requested: September 26, 1985 Nearest Interline Connection to Requested Departure Time: 7:OOA Leave: Flint AL 1420 Arrive: Detroit 7:25A 1l:OOA Leave: Detroit 8:20A RC 416 Arrive: Wiami Displayed: Second Screen despite a tie with a Simmons to Republic marketing alliance connection. A connection to an Eastern flight leaving Detroit eight minutes later was not displayed at all on the computer.
AL - Fischer Bros
(AlleghenyCommuter)
RC - Republic Airlines
Lansing, Wichigan to Orlando Requested Departure Time: 6:00 AW Date Requested: September 26, 1985 Nearest Interline Connection to Requested Departure Time: Leave: Lansing 6:15A AL 1411 Arrive: Detroit 6:45A Uave : Detroit 8:OOA DL 271 Arrive: Orlando 1l:lOA Displayed: Third Screen, only when the agent forced the computer to show connections over Detroit. No better flights were ohown on either of the earlier screens. AL - Fischer Bros (AlleghenyCommuter) DL - Delta Airlines Cedar Rapids/Iowa City, Iowa to New York (La Guardia) Requested Departure Time: 8:00 AR Date Reguested: October 2, Nearest Interline Connection to Regue8ted Departure Time: Leave: Cedar Rapids 8:05A RU 700 Arrive: Chicago 9:lOA Leave: Chicago 10:OOA AA 302 Arrive: New York 12:52P Displayed: Second Screen behind three online connections with departure times of 9:2lA, 7:OOA, and 6:51A RU- Britt Airways AA- American Airlines
South Bend, Indiana to Dallas/FortWorth Reguosted Departure Time: 7:ooAn Date Requested: October 2, Nearest Interline Connection to Requested Departure Time: Leave: South Bend 7:35A RU 272 Arrive: Chicago 8:lOA Leave: Chicago 9:05A AA 511 Arrive: Dalla8 1l:OSA Displayed: Second Screen behind three online connections with departure times of 6:OOA, 6:3lA, and 0:3OA. RU - Britt Airways AA - American Airlines