Journal of Air Transport Management Analysis 2022

Please develop a table of at least 25 articles Citation Research objectives Methodology Findings Conclusion/Future research Find answers to these questions 1- what are the important infrastructure needed in a company and a country to make e-commerce work 2- success factors for e-commerce adoption 3- e-commerce in airlines 4- code sharing 5- competiteve advantage 6- the personalization through e- commerce all those related chapter 2

A comparative performance analysis of airline strategic alliances using
data envelopment analysis
Hokey Min a, * , Seong-Jong Joob
a James R.Good Chair in Global Supply Chain Strategy,Department of Management,BAA 3008C,College of Business Administration,Bowling Green State
University,Bowling Green,OH 43403,USA
b Department of Supply Chain Management,College of Business,Central Washington University-Des Moines,2400 S.240th Street,Des Moines,WA 98198,
USA
a r t i c l e i n f o
Article history:
Received 27 March 2015
Received in revised form
2 December 2015
Accepted 17 December 2015
Available online xxx
Keywords:
Airline strategic alliances
Performance measures
Data envelopment analysis
a b s t r a c t
As “open skies” agreements became more common among different countries and thus began to open u
international routes to further competition, the global airline industry has undergone accelerated
structural changes for the last two decades.These changes include the consolidation and expansion of
airline strategic alliances throughout different regions of the world. Though airline strategic alliances are
generally perceived to be a major driver for enhancing the operating efficiency and the subsequent
competitiveness of participating member airlines, the concrete evidence supporting such a perception is
still lacking in the literature.This paper is one of few attempts to evaluate the comparative efficiency of
the strategic alliances among global airlines and then assess the managerial impact of airline alliances o
the airline's comparative performances.
© 2015 Elsevier Ltd.All rights reserved.
1. Introduction
In the wake of prolonged world-wide recessions and sky-
rocketing oil prices, the airline industry lost $16 billion in 2008 and
$9.9 billion in 2009 (Zacks Equity Research, 2011). Although there is
a growing optimism for the revival of the airline industry with the
recent profit gains, the global airline industry has been hit hard by
rising fuel prices, instable yields,weak traffic volumes,security
hassles,and increased taxation for the last few years.To make it
worse,the competition in the global airline industry gets tougher
after a series of deregulations and “open skies” agreements across
the world that liberalized commercialaviation services and then
opened up international airports and transcontinental routes to full
competition.To survive in this deteriorating market condition,
many international flag carriers chose to consolidate their opera-
tions and created economies of scale through mergers and acqui-
sitions (M&A) due in part to changes in ownership laws and
freedom of the air. M&A of airlines, however, can backfire because it
may limit services to smaller regional routes, increase airfare,
create potentialstrife among integrated workers,raise cost asso-
ciated with increased frequent mileage rewards, and subject
combined airliners to antitrust scrutiny. As illustrated by the recent
mergers of Delta and Northwest in 2008, United and Continental in
2010,and Southwest and Air Tran in 2010,M&A is the continuing
trend of the airline industry.Despite its popularity and potential
benefits,many M&A efforts did not bring fruits to the merged
companies. Defying the conventional wisdom, many M&A attempts
did not go well as they were planned and might undermine the
performances of the merged companies (King et al., 2003).In fact,
the Weekly Corporate Growth Report reported that 70% of the M&A
failed to achieve its anticipated value and 60e80% ofthe M&A
underwent a slow and painful demise (Palmer,2005).
Considering this high risk of M&A failures,airline strategic al-
liances (airline alliances hereafter) including code-sharing,equity
swaps, insurance pooling,and joint governance have become a
popular alternative to M&A.Generally,airline alliances refer to a
distinct form of the market entry mode which provides airlines
with a low-cost means of gaining access to new markets and local
infrastructure such as airports (Doz et al., 1990).One of the most
popular and simplestforms of airline alliances is code sharing
which is a commercial agreement between two airlines (operating
and marketing carriers) that allows an airline (marketing carrier) to
put its two-letter identification code on the flights of another
airline (operating carrier) as they appear in computer reservations
systems (US GeneralService Administration,2011).For example,* Corresponding author.
E-mail addresses: [e-mail] (H.Min), [e-mail] (S.-J.Joo).
Contents lists available at ScienceDirect
Journal of Air Transport Management
j o u r n a lhomepage: w w w . e l s e v i e r . c o m / l o c a t e / j a i r t r a m a n
http://dx.doi.org/10.1016/j.jairtraman.2015.12.003
0969-6997/© 2015 Elsevier Ltd.All rights reserved.
Journal of Air Transport Management 52 (2016) 99e110

Delta Airlines might have an agreementto operate flights for
Korean Airlines on a route to Detroit, Michigan. This flight would be
listed under Korean Airline's identification code (KE) but actually
operated by Delta Airlines. This code sharing agreement allows the
airline to expand its customer bases and service offerings without
additional resources (e.g.,crews),equipment (e.g.,airplanes),and
investment.Also, it helps code sharing partners improve its pas-
senger services through one-stop booking for connecting flights
and concerted service efforts (e.g., shared responsibility for
handling of missing luggage between multiple partnering airlines).
The prevalence ofcode sharing practices,however,raised some
concerns among consumer protectionists. With little guidance and
regulations,code sharing can be often confusing and not trans-
parent to passengers,because passengersoften do not know
exactly which airline will operate their flights. The only exception is
that the U.S. DOT has begun to require airlines to state which airline
is flying a particular route. In addition, code sharing often forces the
passengers to change their planes at different gates in connecting
airports with additionalsecurity checkpoints and thus increases
hassles for confused passengers.Furthermore,code sharing may
increase the chance of monopoly for a certain route and leave no
alternative option for passengers.For instance,all the non-stop
flight services between San Francisco and Toronto are exclusively
operated by Air Canada due to its code sharing agreement with
other potentialcompetitors such as United Airlines through Star
Alliances.Not to mention the aforementioned adverse impacts on
customer services, code sharing complicates airline branding
strategy,service differentiation strategy,pricing strategy,flight
scheduling/routing,baggage handling,and frequent flyer reward
systems.This added complexity can be a potentialsource of in-
efficiency for airlines. Other forms of airline alliances such as equity
swapping and insurance pooling require substantial financial
commitments in time of financial crisis, while joint governance
structures may limit independent decision making opportunities
and thus constraint aligned airlines' operational flexibility. As
illustrated above, there is a growing need to assess the true value of
airline alliances before jumping onto the bandwagon of airline al-
liances. This paper responds to such a need by systematically
measuring and then finding room for improvement in the
comparative (relative) operating efficiencies and service ratings of
airlines which are parts of key airline alliances using data envel-
opment analysis (DEA). This paper also compares the performances
of key airline alliances to those of the non-alliance group for their
competitive strengths and weaknesses,while identifying the po-
tential sources of inefficiency.Based on the DEA and post-hoc
statisticaldata analyses,this paper provides practicalguidelines
for airlines which intend to retool and refine their alliance struc-
tures and practices.
2. Research background and relevant literature
Since deregulation of the U.S. airline industry in 1978 and
liberalization of the European airline industry in 1986 which gave
carriers greater freedom to operate on any routes and fares what-
ever the market would bear, a dramatic restructuring of the global
airlines industry has occurred.This restructuring led to the refor-
mulation of airlines'business strategies that can better cope with
unfettered free competition, elimination of route restrictions,
flexible airfares,and subsidies to “the EssentialAir Service Pro-
gram” ensuring air services to small communities. The increasingly
popular business strategies adopted by the global airline industry
include: the focus on low-cost niche markets; discount pricing; the
development of hub-and-spoke networks; M&A among competi-
tors; and global strategic alliances.Despite the popularity and
benefit potentials of these strategies, it was not clear whether these
strategies actually worked well for airlines as they were intended.
With this in mind, this paper first examined what have been
studied in the past to assess the impacts of some of these strategies
on the airline performances and competitiveness.
2.1.Strategic choices
In line with Porter's research on generic business strategies,
airlines traditionally followed differentiation and (market)seg-
mentation strategies,with little pressure to contain costs (Porter,
1980).This is especially true prior to the enactment of airline
deregulation acts. Thus, cost leadership as a competitive strategy is
still a new but risky concept for airlines, as illustrated by the recent
business failures ofnotable discount carriers such as Skybus.In
addition,chronic industry challenges such as mounting oil prices,
labor strife, high bankruptcy rates, air safety concerns, and
heightened security in the wake of 9/11 put more pressure on air-
lines to find a way to improve operating efficiencies by controlling
costs.While some airlines such as Southwest Airlines have been
able to follow differentiation strategies and cost leadership strate-
gies simultaneously, many airlines continue to struggle with these
strategic tradeoffs.
Historically,prior studies on strategy formulation in the airline
industry fell into two categories: (1) strategic choices; (2) produc-
tivity measures. The first category includes the studies dealing with
classical strategic management topics such as cost leadership,dif-
ferentiation, deregulation, and market segmentation. For example,
using Porter's generic businessstrategies,Cappel et al. (1996)
theoretically evaluated strategy research as applied to the U.S.
airline industry.At that time, these authors found thatairlines
pursuing a combination strategy of cost leadership and differenti-
ation attained a competitive advantage compared with airlines
adopting a singular strategic approach.
Subsequently,a number of low cost carriers (e.g.,Southwest
Airlines,Jet Blue,and Spirit) gained attention.Cappel et al.(2003)
extended this research stream and examined the airline industry
structure in post deregulation in the European Community (EC) and
post 9/11 in order to determine whether the low cost strategy
would result in superior performance.Their theoreticalquestion
was whether external events (deregulation and 9/11) would have a
temporary or permanent effect on the relationship between
financial performance and generic business strategy choices.
Alamdari and Fagan (2005) also observed that adherence to pure
low-cost strategy could lead to greater profitability than the
adoption of hybrid low-cost and differentiation strategy.
There are additional external factors that might affect the trend
toward the low-cost strategy.Customers who use the internet to
purchase airline tickets find lower fares than customers who use
travel agents. Research indicates the lower fares may be partially a
by-product of a broader and more thorough search (O'Connell and
Williams, 2005; Brunger and Perelli,2009). Other studies have
examined the relationship between the low-cost strategy of new
entrants and changesin airline revenue managementsystems
(Gorin and Belobaba,2004). These authors found thatlow-fare
airline entrants can lead to substantial revenue losses for the
incumbent carriers.However,both incumbents and low fare new
entrants alike benefit substantially from the use of revenue man-
agement systems.A comprehensive review ofrevenue manage-
ment and its development can be found in McGill and Van Ryzin
(1999).
Prince and Simon (2009)argued that much of the previous
research on airline competitive behaviors focused exclusively on
price and only recently researchers have begun to test non-price
forms of competition, e.g., service quality. These researchers
examined the relationship between multi-market contact and
H. Min, S.-J.Joo / Journal of Air Transport Management 52 (2016) 99e110100

service quality.Findings indicate that multimarket contact in-
creases delays and that this effect is greater for contacts on more
concentrated routes. Also concerned with customer service,
Scheraga (2004a) examined the relationship between operational
efficiency and customer service in a globalstudy of thirty-eight
large international airlines.His research categorized areas of cost
savings into: (1) Passenger services such as meals, drinks, and other
services included in the fare,and (2) Cost of sales,such as selling
directly to the customers instead of using travel agents. As
mentioned previously,the Internet has been cited as an external
technological factor affecting the trend toward low-cost strategies
by airline companies (Buhalis, 2004; Brunger and Perelli,2009).
2.2. The impact of airline alliances on airline productivity
In addition to the aforementioned literature focusing on the
airline's strategic choices,there is another stream of research that
addresses airline productivity issues (Schefczyk, 1993; Park, 1997).
This stream of research includes studies measuring the extent of
benefits of airline alliances and then assessing their impacton
airline productivity,market shares,firm value creation,and com-
petiveness. The following provides details of these studies.
As strategic alliance has emerged as a popular business strategy
for many airlines,numerous articles have identified the potential
benefits of forming strategic alliances.For instance,Wan et al.
(2009) identified five potentialbenefits of airline alliances.First,
alliances allow airlines to expand their market bases internation-
ally, while circumventing regulatory and legal barriers (Oum et al.,
1996; Park, 1997; Oum et al.,2001; Morrish and Hamilton,2002).
Second,alliances provide costsaving potentials resulting from
sharing facilities,maintenance costs,and joint marketing (Iatrou
and Alamdari,2005).Third, alliances may lead to traffic increases
for partner airlines,thus load factors and the subsequent revenue
may increase (Hannegan and Mulvey,1995; Park, 1997; Wright
et al., 2010).Fourth, passengers may benefit from more flexible
schedules,shorter traveltimes, improved luggage handling,and
shared frequent flyer programs (Dennis, 2000). Fifth, alliances may
create more effective cooperation,resulting in the elimination of
direct competition (Vowles,2000) or the improvement of firm
value (Wassmer and Meschi, 2011). In a nutshell, these studies are
predicated on the theoretical model that is graphically conceptu-
alized in Fig. 1.
An early study by Kleymann and Seristo (2001) highlighted the
need for trust in an alliance membership and its relationship to
efficiency.Their research also examined whether airline alliances
were efficiency-seeking or market-oriented (either offensively or
defensively).Oum et al.(2004) indicated that alliances involving
high-level cooperation were found to have a stronger positive effect
on both productivity and profitability than alliances involving low-
level cooperation. Jiang et al. (2008) referred to an alliance as inter-
outsourcing and indicate the goal was to improve efficiency for the
carriers and the travelexperience for the customers.The inter-
outsourcing differs from M&A in that the partners choose to
cooperate only so long as it is mutually beneficial to each member.
Wagner et al. (2005) found that airlines involved in alliances
showed higher joint procurement activities than airlines that were
not involved in strategic alliances. Lin (2008a,b) examined the role
of code-sharing in market entry deterrence and its resulting impact
on passenger demand and airfares.On the other hand,Goh and
Uncles (2003) found that airline passengers did not recognize the
benefits of airline alliances such as seamless travel,transferable
priority status,and extended lounge access and thus airline alli-
ances did not give aligned airlines a competitive advantage over
their rivals. Defying the conventional wisdom, Armantier and
Richard (2008) also found that code sharing through airline
alliances did not improve consumer (passenger) welfare in terms of
destination access and flight frequencies.
Similar to the studies cited previously regarding differentiation,
Tiernan et al.(2008) argued that differentiation was the primary
motive for membership in international airline alliances; however
their research did not indicate significantdifferences in service
quality among the alliances they studied. Other studies (e.g.,
Brueckner and Whalen,2000; Wan et al.,2009; Zou et al.,2011)
have measured the impactof airline alliances on airfares. For
example,Brueckner and Whalen (2000) verified that airline alli-
ances resulted in airfare reduction through joint pricing. Later, Wan
et al.(2009) found that the extent of impact of an alliance on air-
fares depended on the ability of an alliance to coordinate fares. Zou
et al.(2011) investigated whether cooperation led to higher fares,
or economics of density led to lower operating costs,then lower
fares. While their results were mixed, they indicated that two of the
major airline alliances appeared to charge significantly higher pri-
ces for through-tickets than the sum of segment fares on comple-
mentary routes.On the contrary,Gayle (2008) found no evidence
that airline alliances facilitated collusive behavior among aligned
airlines.
2.3. The evaluation of airline operating efficiency
Although there is a lack of research examining the consequences
of airline alliances from an operating efficiency standpoint, several
authors have attempted to measure operating efficiencies of air-
lines. For instance, to measure airline operating efficiency, Charnes
et al. (1996) introduced globalefficient production functions in
evaluating the performances ofLatin American airlines in the
presence of uncertainty using a Multiplicative-DEA model.
Following suit, Sengupta (1999)proposed an optimal control
theoretic view of the time path of capital inputs which minimized a
discounted sum of total input costs using a Data Envelopment
Analysis (DEA) model. Adler and Golany (2001) used principal
componentanalysis (PCA)in combination with DEA to analyze
efficient network configurationsin Western European airlines.
Lapre and Scudder (2004) analyzed ten major airlines by separating
them into two groups based on geographic specialists and
geographic generalists.Using DEA and Tobit analysis,Scheraga
(2004b) investigated the structural drivers of airline operating ef-
ficiency in relation to the events of 9/11. More recently, Lin
(2008a,b) reviewed and analyzed previous airline studies in
terms of variables, terminologies, and models used to measure the
performances ofTaiwanese domestic airlines using DEA.Barbot
et al. (2008) evaluated the comparative efficiency of49 airlines
from the different part of the world using both DEA and total factor
productivity.They found that low-cost carriers were generally
more efficient than full-service carriers and labor and fleet utili-
zation tended to affect airline efficiency. Barros and Peypoch (2009)
evaluated the operating efficiency of selected airlines which
belonged to the Association of European Airlines (AEA) from 2000
to 2005 using the two-stage DEA.Similar to the finding of Barbot
et al. (2008),they discovered the influence of economies of scale
on airline operating efficiency.Based on the stochastic frontier
analysis Sj€ogren and S€oderberg (2011) observed that deregulation
improved airline productivity, while state ownership (or decreased
share of private ownership)had no significant effect on airline
productivity.
None of these earlier studies,however,examined the inter-
dynamics ofairline alliances and operating efficiencies.In other
words,airline operating efficiencies and their relationship to stra-
tegic competitive advantage have notbeen studied well in the
existing literature. Considering the paucity of prior studies assess-
ing the impact of airline alliances on airline operating efficiency and
H. Min, S.-J.Joo / Journal of Air Transport Management 52 (2016) 99e110 101

the subsequent competitiveness, this paper is intended to create a
new knowledge base regarding the comparative efficiencies of
airline alliances with strategic insights.
3. Research methodology
3.1. Data envelopment analysis framework
DEA is a special application of linear programming based on the
frontier methodology of Farrell (1957). Since Farrell, a major
breakthrough for developing DEA was achieved by Charnes et al.
(1978) and by Banker et al.(1984).DEA is a usefulapproach for
measuring relative efficiency using multiple inputs and outputs
among similar organizations or objects.An entity that is an object
to be measured for efficiency is called a decision-making unit
(DMU). Because DEA can identify relatively efficient DMUs among a
group of given DMUs,it is a promising tool for comparative per-
formance analysis.
To elaborate,DEA can be employed for measuring the compar-
ative efficiency of any entity,which has inputs and outputs and is
homogeneous with peer entities in an analysis.According to a
recent DEA study performed by Haas and Murphy (2003), there is a
remedy for a group of entities that are non-homogeneous (e.g.,
manufacturing plants in the same industry producing different
products,airlines serving different customer bases in different re-
gions). In addition, DEA is applicable to DMUs with categorical and
uncontrollable (or environmental)input data such as air safety
regulations and tarmac rules (Athanassopoulos and Thanassoulis,
1995; Mahajan, 1991).Therefore,DEA can be applied to the wide
variety of DMUs without much restriction.DEA is designed to
identify the best practice DMU without a priori knowledge of which
inputs and outputs are most important in determining an efficiency
measure (i.e.,score) and assessing the extent of inefficiency for all
other DMUs that are not regarded as the best practice DMUs (e.g.,
Charnes etal., 1978).Since DEA provides a relative measure,it
differentiates between inefficient and efficient DMUs relative to
each other. Thus, the best practice (most efficient) DMU is rated as
an efficiency score of one, whereas all other less efficient DMUs are
scored somewhere between zero and one.To summarize,DEA
determines the following (Sherman and Ladino, 1995):
The best practice DMU that uses the least resources to provide
its products or services at or above the performance standard of
other DMUs;
The less efficient DMUs compared to the best practice DMU;
The amount of excess resources used by each of the less efficient
DMUs;
The amount of excess capacity or ability to increase outputs for
less efficient DMUs without requiring added resources.
By denoting E0 as an efficiency score for the base DMU 0,DEA
can be mathematically expressed as:
Maximize E0 ¼
( P R
r¼1 ur0 yr0
)
( P I
i¼1 vi0 xi0
) (1)
subject to:
( P R
r¼1 ur0 yrk
)
( P I
i¼1 vi0 xik
) 1 for all k (2)
ur0 ; vi0 d for all r; i; (3)
where
yrk: the observed quantity of output r generated by unit k ¼ 1, 2,
…, N,
xik: the observed quantity of input i consumed by unit k ¼ 1,2,
…, N,
Fig. 1. A conceptual model theorizing the link between airline alliance and efficiency.
H. Min, S.-J.Joo / Journal of Air Transport Management 52 (2016) 99e110102

ur0: the weight to be computed given to output r by the base unit
0,
vi0: the weight to be computed given to input i by the base unit
0,
d: a very small positive number.
Typically, an ordinary DEA model utilizes a constant returns-to-
scale so that all observed production combinations can be scaled up
or down proportionally (Charnes et al., 1978).On the other hand,
DEA can consider a non-proportionalreturns-to-scale including
increasing or decreasing returns-to-scale (Banker et al., 1984). This
type of DEA can detect and prevent degeneracy that may neglect
alternative optimalsolutions (Fumero,2004).A dual form of the
output-oriented CCR modelwith a real variable f and a non-
negative vectorg can be expressed as:
minf;g f (4)
subject to fx o Xg 0 (5)
Yg y o (6)
g 0 (7)
where xk ¼ input vector for k ¼ 1,…, n
yk ¼ output vectors for k ¼ 1, …,n
X ¼ input matrices
Y ¼ output matrices
Similarly,the dual form of an output-oriented BCC model with
row vector d for input multipliers and row vector m for output
multipliers can be mathematically expressed as:
mind;m;d0 z ¼ dxo d 0 (8)
subject to myo ¼ 1 (9)
dX mY d 0 e 0 (10)
d 0; m 0; (11)
d0 free in sign, where d0 is the scaler related to eg ¼ 1 in the
envelopment model.
In this study, involving the assessment of comparative efficiency
for airline alliances,we use output-oriented system models with
constant and variable returns to scale and the algorithm for ordinal
measuresalong with a Charnes Cooper Rhodes (CCR) model
(Cooper et al., 2007; Charnes et al., 1978). The use of the CCR version
of DEA is justified because it was designed to overcome the limi-
tations imposed by financialdata,to rank airlines'financialper-
formances relative to its competitors,and to assess the overall
comparative operating efficienciesof selected airline alliances
(Thanassoulis,2001). As a post-hoc DEA analysis,we also per-
formed non-parametric rank-sum statistical tests for any discern-
ible group differences between airlines with code sharing and
airlines with no alliance.Instead of a parametric t-test,the non-
parametric rank-sum statistical test was conducted due to a
limited number of the airline alliance groups under evaluation that
would make the normal distribution approximation inappropriate.
3.2. DEA input and output variables
To experiment with the proposed DEA model,we took traffic
and financialdata from the ATW World Airline Report 2010 (Air
Transport World,2010) and service rating information from the
World Airline Star Ranking (SKYTRAX, 2010). In addition, we
gathered other secondary data available from the websites con-
taining airline alliance membership information atthe end of
September,2010.Based on the availability of data which reflected
alliance situations in 2009 due to a time lag between data collection
and publication, we selected eight airlines for SkyTeam, 27 for Star
Alliance, nine for Oneworld, and 15 for non-member airlines, which
was close to the average number of airlines in each alliance. Table 1
shows the number ofairlines in each category.The numbers in
parentheses will be used for identifying respective groups
throughout the study.
Based on the literature review of prior studies on airline effi-
ciency and the norm of the airline industry practices,we include
five continuous variables and one ordinal variable requiring a
specialtreatment for analysis.Two input variables are Operating
Expenses (in thousand U.S.dollars) and Underutilization (in per-
centage),which is computed by subtracting load factors in per-
centage from 100. Generally, any resources used by DMU should be
included as input.To elaborate,operating expenses include many
elements of variable costs,such as fuel,lubricants,airplane parts,
tires, license fees,utilities, taxes and insurance premiums that
comprise another key resource for maintaining airline fleet and
hangar operations.Thus, operating expenses were included as
input. Since underutilization represents the way given physical
resources such as assets are used,we consider underutilization as
input. Herein,in specifying input variables,we exercised caution,
since the endogeneity ofinputs could generate biased efficiency
measures especially with smallsample sizes in the public sector
where politically motivated feedback was reflected in the DEA
estimation. Through simulation experiments,Orme and Smith
(1996) observed thatinefficient DMUs using low levels of the
endogenousinput were likely to make greater efficiency im-
provements than equally inefficient DMUs using higher levels of
the input and thus could lead to inflated efficiency scores. A similar
observation was made by Cordero et al. (2015) who discovered that
the high positive endogeneity levelcould impair DEA estimates.
Although DEA should always yield feasible improvement targets in
the absence of measurement error, the achievement of such targets
might be more difficult for DMUs employing low levels of re-
sources.Thus,the efficiency measures in the presence of endoge-
neity may create so-called sparsity bias when comparing units
using different levels of resource inputs.Also, Ruggiero (1999)
warned that the use of multiple non-discretionary inputs
involving exogenous variables could lead to distorted measure-
ments in the DEA model. So, we excluded those inputs from
consideration and conducted a two-stage DEA analysis to mitigate
the potential measurement error caused by the endogeneity. On the
output side, the overall performance of an airline can be measured
by operating revenue that best reflects the operational efficiency of
the airline. Other well-known financial ratios such as profit margin
and return-on investment were not considered relevant, because a
less profitable airline may be more efficient in utilizing its
personnel and equipment than the more profitable airline. In
addition to operating revenue (in thousand U.S. dollars), two
continuous output variables that we chose are Passengers(in
number of passengers) in thousand and RPKs (revenue passenger
kilometers) in million U.S. dollar-kilometers.RPK is a popular
measure of sales volume of passenger traffic in the airline industry.
The ordinal variable is Service Rating that measures service levels of
airlines using a scale one (lowest) to five (highest). Table 2 exhibits
descriptive statistics of these variables. Analysis will be conducted
in two stages.First, after including five continuous variables,we
computed and compared efficiency scores of airlines by groups that
H. Min, S.-J.Joo / Journal of Air Transport Management 52 (2016) 99e110 103

they belonged to. Second, we computed efficiency scores of airlines
by service ratings.
4. Results and discussions
We measured efficiency scores for four groups in two stages.
First,we assessed technical efficiency (TE) and pure technical effi-
ciency (PTE) using an output-oriented,constant returns-to-scale
system model (SYS-O-C) and output oriented,variable returns-to-
scale system model (SYS-O-V),respectively.It is noted that these
efficiency scores were computed based on the comparison of all
DMUs (airlines) across groups simultaneously so that we could use
the same reference points in a single sample.The rationale being
that the comparison of all DMUs across groups would better
represent the true efficiency scores of airlines than the comparison
of DMUs within a group (i.e.,a particular alliance or non-alliance
group) given that DEA efficiency scores are meant be relative per-
formance measures in lieu of absolute performance measures.By
the same token, we also computed the efficiency scores of airlines
by service ratings. Since service ratings were ordinal measures, we
followed the steps for measuring the efficiency ofDMUs with
ordinal variables (Cooper et al.,2007).
TE shows overall or managerial efficiency,which is the multi-
plication of PTE and scale efficiency (SE).PTE reveals pure mana-
gerial efficiency that is based on endogenous managerialfactors
such as the level of passenger satisfaction, advertising efforts, and a
load factor.SE reflects exogenous factors such as operating envi-
ronments or market conditions.Two out of eight (25.0%) are 100
percent efficient in SkyTeam, four out of 27 are 100 percent efficient
in Star Alliance (14.8%),one out of nine (11.1%) is 100 percent effi-
cient in Oneworld,and seven out of15 (46.7%) are 100 percent
efficient among the airlines not affiliated with any alliance.This
result suggests thatairlines that belong to an alliance do not
necessarily achieve higher efficiency than the ones withoutan
alliance agreement.
PTE will show whether or not the airlines are comparatively
efficient for their internal operations.The number of PTE efficient
airlines for each alliance is three for SkyTeam (37.5%),11 for Star
Alliance (40.7%),three for Oneworld (33.3%),and eight for no alli-
ance (57.1%).Again,the airlines in an alliance are not better than
those without an alliance in terms of the percentage of PTE efficient
airlines as shown in Table 3.
Using TE and PTE scores in Table 3 and the relationship between
TE (TE ¼ PTE SE) and PTE, we can identify sources of inefficiency
such as pure managerial (PTE) and/or environmental (SE) factors.
The airlines such as ShanghaiAirlines, Aerlingus,and Pakistan
Airlines show that SE scores close to one,and,therefore,their TE
inefficiency is due to internalmanagerialelements (PTE).Mean-
while, Adria, Croatia Airlines, and Malev Hungarian are 100 percent
efficient on pure managerial aspects,and their inefficiency is
caused by different operating environments such as a market focus
on unpopular regional destinations and a lack of promotional op-
portunities resulting from the mostly government owned and
influenced local media. To further identify the sources of in-
efficiency,we also looked at TE scores of airline groups.Mean TE
scores are 0.9336 for SkyTeam, 0.9004 for Star Alliance, 0.9046 for
Oneworld, and 0.9431 for the airlines not affiliated with an alliance.
For the Mean TE scores,the airlines without an alliance demon-
strate the highest score, and the airlines in Star Alliance exhibit the
lowest score.For mean PTE scores,the airlines in Sky Team show
the highest (0.9476),and the airlines in Star Alliance maintain the
lowest (0.9398).
Surprisingly,the airlines without an alliance have performed
better than the airlines with an alliance for most cases.To gener-
alize this finding, we conducted post-hoc statistical analysis.
Because the efficiency scores were measured comparatively,we
chose a non-parametric rank sum test. To accommodate four
groups,we employed a KruskaleWallis test for group differences.
Table 4 exhibits the results of the test.For both TE and PTE,there
was no statistically significant difference among the groups.Thus,
we argue that strategic alliances are neutralto competitive ad-
vantages in this study.In fact,Porter (1990) pointed out that the
effects of alliances could be temporary and seldom provided
competitive advantagesto their participants. Our finding is
congruent with Porter's observation.
Due to the convexity assumption of DEA,which states that the
fundamental set of possible inputeoutput mixes is convex, it is not
allowed to include categorical variables directly in basic DEA
models.Instead,we have to use the special DEA models that can
handle the categorical variables.Service ratings for airlines,which
range one (worst service) to five (best service),are ordinalmea-
sures and controllable by decision makers.In our data set, the
service ratings vary from two to five or four different categories.
Conversely,there is no airline with service rating one in the data
set. To incorporate the service ratings in our efficiency analysis, we
follow the algorithm suggested by Cooper et al. (2007). Fig. 2 shows
the detailed steps of the algorithm used in our analysis.
Table 5 displays the efficiency analysis that is computed using
the same variables summarized in Table 3 and the service ratings as
an additional variable.The first column displays the names of air-
lines with service ratings (numericalvalues after airlines).The
second column includes technical efficiency scores (TE) computed
with CCR output-oriented models.From the third columns,refer-
ence columns revealreference airlines in the firstcolumn, and
Table 1
Airline alliance membership.
Membership SkyTeam (1) Star Alliance (2) Oneworld (3) Non-member (4)
Number of airlines 8 27 9 15
Table 2
Descriptive statistics.
Underutilization (%) Operating expenses (in $1 000) Passengers (in 1000) RPKs (in $ millions) Operating revenue (in $1 000) Service rating
Max 38.60 31,535,817.00 161,047.00 304,025.00 31,924,069.00 5.00
Min 16.10 250,000.00 1144.00 1193.00 232,521.00 2.00
Mean 25.42 6,296,117.34 24,540.41 52,489.66 6,260,800.32 3.42
SD 5.96 7,123,182.26 27,835.90 57,638.93 6,973,153.93 0.65
Type Input Input Output Output Output Output
SD: Standard Deviation.
H. Min, S.-J.Joo / Journal of Air Transport Management 52 (2016) 99e110104

lambda columns show weights for the reference airlines. That is to
say, if an airline in the first column is inefficient, one of the airlines
in reference columns,which has the highest lambda value will be
the reference airline. For example, there are two reference airlines
for Alitalia whose service rating is 3.Between the two reference
airlines, LAN Airlines has higher lambda value than AirTran Airways
does.Since AirTran's rating is the same as that of Alitalia,Alitalia
does not have to improve its current service rating. Instead, Alitalia
should improve other variables to be efficient.On the other hand,
the airlines such as Air Canada,Continental,United,US Airways,
and Pakistan Airlines need to improve their service ratings from
three to four because their reference airlines with the highest
lambda values have higher service ratings such as four.
Since we are interested in the efficiency difference among alli-
ances,we investigate the difference with service ratings using a
non-parametric statistic, a KruskaleWallis rank sum test. It is noted
that the KruskaleWallis rank sum test for post-hoc DEA analysis is
common (e.g.,Asmild et al., 2013; Botti et al., 2009; Gonclaves,
2013). The KruskaleWallis test statistics that we used are:
Q ¼ 12=NðN þ 1Þ
P k
i¼1R2
i =ni 3ðN þ 1Þ, where N (totalnumber of
observations) ¼ n1 þ n 2 þ…,þ n k and Ri ¼ rank sums (Gibbons,
1993).Although the probability of the test statistic is close to 0.1
that is a liberal criterion for the interpretation of significance,the
statistic on rank differences among the groups ofairlines is not
significant as shown in Table 6.Thus, we can conclude that the
airlines belonging to an alliance fail to provide differentiated ser-
vices compared to the ones not affiliated with an alliance.Overall,
regardless ofthe potential benefits of airline alliances through
collaboration such as jointmarketing,coordinated flightsched-
uling, combined frequent-flyer programs,airport facility sharing,
joint maintenance and ground support (Li, 2000; Oum et al., 2004),
we did not find any competitive advantage attributed to strategic
alliances.
To check and see if there is any competitive advantage of airline
alliances over time,we calculated DEA efficiency scores ofthree
alliance groups and a non-alliance group for a five-year span during
2006 through 2010, which included pre-recessionary, recessionary,
and post-recessionary periods. Table 7 and Fig. 2 show patterns of
airline efficiency scores for a five year period. Overall, all the groups
registered the lowest DEA efficiency score in 2008 soon after the
great recession started in December of 2007 and then fully recov-
ered in 2010 after the recession ended in June of 2009 (see Fig. 3). It
is intriguing to note that an efficiency gap between alliance groups
and a non-alliance group shrank in the peak of great recession in
Table 3
Efficiency scores by groups (in 2010).
SkyTeam Star Alliance Oneworld No alliance
Airlines TE PTE Airlines TE PTE Airlines TE PTE Airlines TE PTE
Aeroflot 1.0000 1.0000 Adria 0.8305 1.0000 American Airlines 0.9587 0.9604 AerLingus 0.8502 0.8505
Air Europa 0.9176 0.9270 Aegean Airlines 0.9379 1.0000 British Airways 0.9435 0.9465 AirTran Airways 1.0000 1.0000
Air France KLM 0.9531 1.0000 Air Canada 0.9233 1.0000 Cathay Pacific 0.9847 1.0000 Alaska Airlines 0.9692 0.9909
Alitalia 0.8164 0.8189 Air China 1.0000 1.0000 Finnair Group 0.8266 0.8555 Emirates 1.0000 1.0000
China Southern 0.9791 0.9814 Air New Zealand 0.9083 0.9278 Iberia Group 0.8239 0.8391 Hainan Airlines 0.9116 0.9125
Delta Air Lines 1.0000 1.0000 ANA 0.8838 0.9288 JAL 0.8368 0.9004 Hawaiian Airlines 1.0000 1.0000
Kenya Airways 0.8973 0.9307 Asiana Airlines 0.8303 0.8374 LAN Airlines 1.0000 1.0000 JetBlue 1.0000 1.0000
Korean Air 0.9056 0.9230 Austrian Airline 0.7708 0.7741 Malev Hungarian 0.7975 1.0000 Malaysia Airlines 0.8470 0.8543
Blue1 0.8175 1.0000 Quantas Group 0.9695 0.9920 Norwegian Air 0.9940 1.0000
bmi 0.8160 1.0000 Pakistan Airlines 0.8691 0.8698
Brussels Airlines 0.8110 0.8353 Philippine Airlines 0.7804 0.7812
Continental 0.9877 1.0000 Southwest 1.0000 1.0000
Croatia Airlines 0.8093 1.0000 Transareo 1.0000 1.0000
Egypt Air 0.9121 0.9270 Westjet 1.0000 1.0000
LOT Polish Airlines 0.7919 0.8151 Virgin Atlantic 0.9245 0.9346
Lufthansa 1.0000 1.0000
SAS 0.8356 0.8357
Shanghai Airlines 0.8404 0.8406
Singapore Airlines 0.9221 0.9221
South African Airways 0.9493 0.9618
Swiss 0.9239 0.9468
TAM 0.9088 0.9100
TAP Portugal 0.9088 0.9134
Thai Airways Int'l 0.9979 0.9994
Turkish Airlines 1.0000 1.0000
United Airlines 0.9944 1.0000
US Airways 1.0000 1.0000
Mean 0.9336 0.9476 Mean 0.9004 0.9398 Mean 0.9046 0.9438 Mean 0.9431 0.9462
Table 4
KruskaleWallis test for group differences.
Alliance N Mean rank Chi-square Degree of freedom Significance
SYSOC (TE) SkyTeam 8 27.50 4.768 3 0.190
Star Alliance 27 33.93
Oneworld 9 32.94
No Alliance 15 22.50
SYSOV (PTE) SkyTeam 8 29.94 0.558 3 0.906
Star Alliance 27 30.69
Oneworld 9 32.22
No Alliance 15 27.47
H. Min, S.-J.Joo / Journal of Air Transport Management 52 (2016) 99e110 105

2008, while such a gap seems to be larger during a pre-recessionary
period. This result implies that an airline alliance is not considered
a recession-proof strategy. On the other hand, with an exception of
the Star Alliance group,both Oneworld and Sky Team alliance
groups tended to perform better than a no-alliance group
throughout the five-year evaluation period (see Table 7 and Fig. 3).
This result suggests that an airline alliance itself will not necessarily
enhance an airline's competitive advantage,but its make-up (e.g.,
partnership characteristics)may influence airline efficiency.For
instance, alliances such as Oneworld and Sky Team were comprised
of a relatively smallnumber of partners and thus allowed such
alliance groups to form more focused relationships among
partnering airlines, whereas Star Alliances comprised of more than
20 partnering airlines might have diluted the quality of their re-
lationships and thus did not reap the full benefits of airline
alliances.
To further examine the potential influence of airline alliances on
airline performances and competitiveness, we analyzed the airline
efficienciesbefore and after joining alliances. For comparative
purposes,we considered three high-profile airlines as the sample.
These airlines are: American, Delta, and United-Continental.
Although other airlines are considered for the sample experi-
ments,only those three airlines provided data compatible to each
other during the entire evaluation period (see Table 8). To elaborate,
Fig. 2. The steps for evaluating the airline efficiency with respect to service ratings.
H. Min, S.-J.Joo / Journal of Air Transport Management 52 (2016) 99e110106

we chose input and output variables from a return-on-assets
perspective (Joo et al.,2011).Because net incomes reflected the
airlines' non-operating activities and were frequently manipulated
using different accounting practices,we used operating revenues
Table 5
Efficiency analysis by service ratings in 2010.
Airlines TE Reference Lambda Reference Lambda Reference Lambda Reference Lambda
Transareo 2 1.0000 Transareo 2 1.0000
Aeroflot 3 1.0000 Aeroflot 3 1.0000
Air Europa 3 0.9086 Aeroflot 3 0.2396 LAN Airlines 3 0.1807 AirTran Airways 3 0.2085
Alitalia 3 0.8117 LAN Airlines 3 1.1810 AirTran Airways 3 0.3572
China Southern Airlines 3 0.9690 Air China 3 0.3052 Southwest 3 0.4893 Emirates 4 0.0339 JetBlue 4 0.2529
Delta Air Lines 3 1.0000 Delta Air Lines 3 1.0000
Kenya Airways 3 0.8973 Aeroflot 3 0.3718
Adria 3 0.8240 LAN Airlines 3 0.0682 AirTran Airways 3 0.0141
Aegean Airlines 3 0.9284 LAN Airlines 3 0.1515 AirTran Airways 3 0.1738
Air Canada 3 0.8969 Delta Air Lines 3 0.0122 Lufthansa 4 0.0229 Emirates 4 0.7840
Air China 3 1.0000 Air China 3 1.0000
Blue1 3 0.8053 LAN Airlines 3 0.0587 AirTran Airways 3 0.0380
bmi 3 0.8292 AirTran Airways 3 0.3733 JetBlue 4 0.0262
Brussels Airlines 3 0.8090 Aeroflot 3 0.1703 LAN Airlines 3 0.2787
Continental 3 0.9446 Delta Air Lines 3 0.1826 Lufthansa 4 0.0098 Emirates 4 0.6664
Croatia Airlines 3 0.8011 LAN Airlines 3 0.0509 AirTran Airways 3 0.0584
Egypt Air 3 0.8993 Aeroflot 3 0.5181 LAN Airlines 3 0.0796 AirTran Airways 3 0.0919
LOT Polish Airlines 3 0.7913 Aeroflot 3 0.0378 LAN Airlines 3 0.2775
SAS 3 0.8332 Air China 3 0.7794 LAN Airlines 3 0.4466
Shanghai Airlines 3 0.8321 LAN Airlines 3 0.2497 AirTran Airways 3 0.2921
TAM 3 0.9051 Air China 3 0.3425 LAN Airlines 3 0.8596 AirTran Airways 3 0.2128
TAP Portugal 3 0.9051 Aeroflot 3 0.9815 LAN Airlines 3 0.0677
United Airlines 3 0.9664 Delta Air Lines 3 0.3164 Lufthansa 4 0.0665 Emirates 4 0.4987
US Airways 3 0.9641 Delta Air Lines 3 0.0904 Southwest 3 0.2464 Emirates 4 0.4867
American Airlines 3 0.9407 Delta Air Lines 3 0.4924 Lufthansa 4 0.0863 Emirates 4 0.3887
Iberia Group 3 0.8165 Air China 3 0.6362 Emirates 4 0.2484
LAN Airlines 3 1.0000 LAN Airlines 3 1.0000
Malev Hungarian 3 0.7889 LAN Airlines 3 0.1006 AirTran Airways 3 0.1058
AerLingus 3 0.8375 LAN Airlines 3 0.3963 AirTran Airways 3 0.2621
AirTran Airways 3 1.0000 AirTran Airways 3 1.0000
Alaska Airlines Group 3 0.9600 Aeroflot 3 0.0565 Air China 3 0.1071 LAN Airlines 3 0.6765 AirTran Airways 3 0.0428
Hawaiian Airlines 3 0.9920 Aeroflot 3 0.1725 LAN Airlines 3 0.0130 AirTran Airways 3 0.2793
Norwegian Air Shuttle 3 0.9895 LAN Airlines 3 0.0983 AirTran Airways 3 0.3897
Pakistan Airlines 3 0.8724 Aeroflot 3 0.0595 JetBlue 4 0.3441
Philippine Airlines 3 0.7711 Aeroflot 3 0.3567 AirTran Airways 3 0.1178 JetBlue 4 0.2517
Southwest 3 1.0000 Southwest 3 1.0000
Westjet 3 0.9885 Aeroflot 3 0.1536 LAN Airlines 3 0.2359 AirTran Airways 3 0.3843
Air France KLM 4 1.0000 Air France KLM 4 1.0000
Korean Air 4 0.9233 Turkish Airlines 4 0.6846 Emirates 4 0.4693
Air New Zealand 4 0.9147 Turkish Airlines 4 0.6167 JetBlue 4 0.1174
ANA 4 0.9279 Air France KLM 4 0.1817 Emirates 4 0.5143 JetBlue 4 0.9325
Austrian Airline Group 4 0.7851 Turkish Airlines 4 0.8068
Lufthansa 4 1.0000 Lufthansa 4 1.0000
South African Airways 4 0.9669 Turkish Airlines 4 0.6036
Swiss 4 0.9371 Turkish Airlines 4 0.5664 Emirates 4 0.1560
Thai Airways Int'l 4 0.9921 Turkish Airlines 4 0.4361 Emirates 4 0.0747 JetBlue 4 0.6243
Turkish Airlines 4 1.0000 Turkish Airlines 4 1.0000
British Airways 4 0.9409 Air France KLM 4 0.1592 Emirates 4 0.6434 JetBlue 4 0.2136
Finnair Group 4 0.8419 Turkish Airlines 4 0.6903
JAL 4 0.9109 Air France KLM 4 0.3028 Emirates 4 0.0675 JetBlue 4 1.2995
Quantas Group 4 0.9481 Lufthansa 4 0.0938 Emirates 4 0.7903
Emirates 4 1.0000 Emirates 4 1.0000
Hainan Airlines 4 1.0000 Hainan Airlines 4 1.0000
JetBlue 4 1.0000 JetBlue 4 1.0000
Virgin Atlantic 4 0.9219 Turkish Airlines 4 0.1685 Emirates 4 0.1328 JetBlue 4 0.4970
Asiana Airlines 5 1.0000 Asiana Airlines 5 1.0000
Singapore Airlines 5 0.9494 Cathay Pacific 5 1.1077
Cathay Pacific 5 1.0000 Cathay Pacific 5 1.0000
Malaysia Airlines 5 1.0000 Malaysia Airlines 5 1.0000
Table 6
KruskaleWallis test results for group differences in service ratings.
Alliance N Mean rank Chi-square Degree of freedom Significance
Technical efficiency SkyTeam 8 26.56 6.236 3 0.101
Star Alliance 27 34.78
Oneworld 9 32.56
No Alliance 15 21.70
H. Min, S.-J.Joo / Journal of Air Transport Management 52 (2016) 99e110 107

for an output. Three input variables that we employed were current
assets, net property and equipment (owned and leased), and other
assets.The other assets for airlines encompassed route and slot
acquisition costs,airport operating rights,and gate lease rights in
addition to traditionalother assets found in the balance sheet of
non-airline firms. Using these input and output variables, we con-
ducted a span of analysis and then tested to see if there existed any
differences in the operating efficiencies ofthree chosen airlines
before and after joining alliances. Given two groups (i.e., before and
after alliances) of observations for each airline,we applied a bilat-
eral DEA model (a slack-based measure of efficiency with a constant
return to scale or an SBM-C model) to analyze such differences (Joo
et al., 2010). This model is known to be a comprehensive measure of
efficiency.After computing DEA efficiency scores,we ranked the
three chosen airlines in the order of their relative operating effi-
ciencies during the span of 22e26 years as summarized in Table 9.
Based on these rankings,we performed KruskaleWallis rank sum
tests to examine whether or not airline alliances significantly
influenced airline efficiencies.As Table 10 indicates,the perfor-
mances of all three airlines got deteriorated after joining alliances.
Although this result is surprising in that airline alliances did not
improve their performances, it is consistent with the earlier result
that we obtained from the comparison of alliance vs.non-alliance
groups.That is to say,airline alliance alone did not seem to make
any significant difference in airline performances despite its benefit
potentials.Instead, we speculated that a multitude of factors
including unexpected supply chain disruptions (e.g., the September
11 incident), economic fluctuations (e.g., worldwide financial
crisis),and internalbusiness strategy (e.g.,global market expan-
sion) might have influenced airline performances.
5. Concluding remarks
As a fast and flexible way to access complementary resources
and skills that reside in other companies,a growing number of
firms formed a strategic alliance (Dyer et al.,2001).In this regard,
the airline industry is no exception as evidenced by increased code
sharing practices among major globalairlines. Its popularity is
based on the underlying premise that strategic alliance helps
Table 7
Efficiency scores of groups for a five year period.
N 2006 2007 2008 2009 2010 Mean
Economic
condition
Pre-recession Recession Post-recession
Sky Team 7a 0.8676 0.8658 0.7760 0.8374 0.8849 0.8463
Star Alliance 23a 0.7806 0.8313 0.7730 0.7714 0.8141 0.7941
Oneworld 8a 0.8551 0.8751 0.8106 0.8034 0.8684 0.8425
No Alliance 15 0.7697 0.8084 0.7504 0.7757 0.8262 0.7861
Mean 0.8142 0.8452 0.7775 0.7970 0.8484
a Note: Due to the unavailability ofcomparable data for a five-year span,we
excluded Alitalia (Sky Team),bmi (Star Alliance),Brussels (Star Alliance),Conti-
nental (Star Alliance), Shanghai (Star Alliance), and Iberia (Oneworld) from the DEA
analysis.
Fig. 3. DEA efficiency trends for a five year span.
Table 8
Span of analysis.
Airlines Year joined an alliance Before After Alliance
American 1999 1986e1998 (13 years) 2000e2012 (13 years) Oneworld
Delta 2000 1988e1999 (12 years) 2001e2012 (12 years) Sky Team
United-Continental 1997 1986e1996 (11 years) 1998e2008 (11 years) Star Alliance
Table 9
Efficiency rankings before and after joining alliances.
Year American Delta United-Continental
2012 4 15 e
2011 9 16 e
2010 13 18 e
2009 20 19 e
2008 12 22 14
2007 19 23 18
2006 21 13 20
2005 22 14 15
2004 23 17 16
2003 24 24 21
2002 25 21 19
2001 26 20 22
2000 18 Joined 17
1999 Joined 7 12
1998 16 12 13
1997 17 6 Joined
1996 15 11 5
1995 3 10 6
1994 5 9 9
1993 2 8 10
1992 6 3 11
1991 10 5 7
1990 11 1 3
1989 1 2 2
1988 7 4 4
1987 8 e 1
1986 14 e 8
Table 10
The results of KruskaleWallis rank sum tests.
American Delta United-Continental
Rank sum (before joining alliances)115 78 66
Rank sum (after joining alliances) 236 222 187
Test statistic 3.103 4.157 3.973
p-value 0.001 0.003 0.000
H. Min, S.-J.Joo / Journal of Air Transport Management 52 (2016) 99e110108

achieve competitive advantages.To verify such a premise,this pa-
per examined whether the code sharing practices as part of stra-
tegic alliances among airlines could be translated into the alliance
participants'competitive advantage.In doing so,we attempted to
assess the impact of code sharing practices on alliance participants'
comparative operating efficiency given that the airline's compara-
tive operating efficiency would lead to its competitive advantage.
Defying the conventional wisdom,we found no significant differ-
ences in airline performances between airlines with strategic alli-
ances and airlines without alliances.Also, it should be noted that
airline performances before and after joining alliances did not show
any signs of improvements.
That is to say,airline alliances did not necessarily improve the
participating airline's comparative operating efficiency despite its
cost saving potentials due to shared resources and customer bases.
Perhaps, a cost reduction and favorable brand recognition attached
to code sharing practices might have been offset by the lack of
differentiated services among code sharing airlines.The rationale
may be that airline alliances are likely to further induce similarity
rather than differentiation due to overlapped geographical market
coverage and standardized services through cooperative service
agreements among alliance partners. This finding is congruent with
the Porter (1990)’s observation indicating that strategic alliances
would not guarantee managerial success.The important implica-
tion for airline executives emanating from this finding is that
strategic alliance alone cannot improve its profit margins and ser-
vice quality unless the airline enhances its internal operating effi-
ciency (e.g.,better asset utilization, higher labor productivity).
Another lesson learned from this finding is that the airline industry
is service-driven and thus economies ofscale created by airline
alliances and the subsequentcost saving opportunities are not
sufficient enough to enhance the airline's competitiveness.Espe-
cially, airline alliances primarily seeking marketexpansion and
brand recognition without creating organizationallearning op-
portunities and acquiring rare or not-easily imitable resources may
not help the airline gain distinctive competitive advantages over its
rivals in the saturated airline market.From a practical standpoint,
ways to differentiate a particular alliance from its competing alli-
ances may include: redemption offrequent flier mile credits for
each alliance partner's flights; no limitation on a number of stops in
a certain geographicalarea (e.g.,Europe); no high-season or fuel
surcharge; and broader route networks covering all five continents.
To summarize, one of the most important practical guidelines that
we should keep in mind is that alliance efforts have to geared to-
ward the improvement of quality of alliances in terms of alliance
partners'service compatibility,the extent of route networks,and
the generosity ofreciprocalloyalty programs (e.g.,frequent flier
miles), rather than simply jumping on the alliance bandwagon.
Another intriguing finding of this study is the influence of alli-
ance partnership characteristics (e.g.,quality of partnership)on
airline efficiency. In particular, we discovered that a smaller alliance
group such as Oneworld and Sky Team tended to outperform a
larger alliance group such as Star Alliance.The potential explana-
tion for this tendency is the fact that the larger the alliance,the
more time-consuming the assimilation process willbe. Thus,the
integration of different airlines (especially new participants) into
the existing alliance structure and rules will be more challenging
and consequently can undermine alliance benefits.Based on this
fining, the caution should be exercised to join a larger alliance
group lacking organizational and technical compatibility.
To the best of our knowledge, this study is one of few attempts to
assess the impact of airline alliances on the global airline's
competitiveness using DEA and its post-hoc analyses and then
investigate whetherthe airline's participation in code sharing
practices can be developed into its viable strategic weapon.In
addition,this paper is one of the first to incorporate both quanti-
tative (financial) and qualitative (service quality ratings) attributes
into airline performance metrics using the specialalgorithm and
factored those attributes into the comparative evaluation of airline
operating efficiency. Furthermore, we assessed the effect of airline
alliances on their participants'performances based on time-series
data (a length of alliance history).Despite our noveleffort, this
study is far from being perfect due in part to its reliance on the
limited time frame and surrogate measures extracted from finan-
cial data.Another limitation ofthis study includes the potential
presence of an unobserved bias due to the limited variables in the
DEA analysis. To overcome some of the shortcomings of this study,
future research efforts can be geared toward assessment ofthe
impact of code sharing practices on airline service quality and
airfare pricing from both the airline and its customers'(airline
passengers') perspectives. Also, its impact may be assessed in terms
of organizational culture,human resource practices,and branding
which are difficult to quantify.
References
Adler, N., Golany,B., 2001.Evaluation of deregulated airline networks using data
envelopment analysis combined with principalcomponent analysis with an
application to Western Europe.Eur.J. Oper.Res. 132,260e273.
Alamdari, F., Fagan, S., 2005. Impact of the adherence to the original low-cost model
on the profitability of low-cost airlines.Transp.Rev.25 (3),377e392.
Athanassopoulos,A.D.,Thananssoulis,E.,1995.Separating market efficiency from
profitability and its implications for planning.J. Oper.Res.Soc.46 (1),20e34.
Air Transport World,2010.ATW World Airline Report 2010.http://atwonline.com/
sites/atwonline.com/files/misc/ATW%20World%20Airline%20Report%202010_0.
pdf. retrieved on September 10,2010.
Armantier,O.,Richard,O., 2008.Domestic airline alliances and consumer welfare.
RAND J.Econ.39 (3),875e904.
Asmild, M., Hougaard, J.L., Kronborg, D., 2013. Do efficiency scores depend on input
mix? A statistical test and empirical illustration. Ann. Oper. Res. 211 (1), 37e48.
Banker, R.D., Charnes, A., Cooper, W.W., 1984. Some models for estimating technical
and scale inefficiencies in data envelopment analysis.Manag. Sci. 30 (9),
1078e1092.
Barbot,C., Costa,A., Sochirca,E., 2008. Airlines performance in the new market
context: a comparative productivity and efficiency analysis.J. Air Transp.
Manag. 14 (5),270e274.
Barros,C.P.,Peypoch,N., 2009. An evaluation of European airlines'operational
performance.Int. J. Prod.Econ. 122 (2),525e533.
Botti, L., Briec,W., Cliquet,G., 2009.Plural forms versus franchise and company-
owned systems: a DEA approach ofhotel chain performance.Omega 37 (3),
566e578.
Brueckner,J.K., Whalen, W.T.,2000. The price effects of internationalairline alli-
ances.J. Law Econ.43 (2),503e546.
Brunger, W.G., Perelli, S., 2009. The impact of the internet on airline fares: customer
perspectiveson the transitions to internet distribution. J. Revenue Pricing
Manag.8 (2/3), 187e199.
Buhalis, D., 2004.eAirlines: strategic and tactical use of ICTs in the airline industry.
Inf. Manag.41 (7),805e825.
Cappel,S.D.,Pearson,T.R.,Romero,E.J.,2003.Strategic group performance in the
commercial airline industry.J. Manag.Res.3 (2), 53e60.
Cappel,S.D.,Tucci,J.E.,Wyld, D.C., 1996.Does a low cost strategy equal success for
commercialairlines: a conceptualquestion.In: Proceedings ofAcademy of
Strategic and Organization Leadership.
Charnes,A., Cooper,W.W., Rhodes,E.,1978.Measuring the efficiency of decision
making units.Eur.J. Oper.Res.2 (6), 429e444.
Charnes,A., Gallegos,A., Li, H., 1996.Robustly efficient parametric frontiers via
multiplicative DEA for domestic and internationaloperationsof the Latin
American airline industry.Eur.J. Oper.Res.88, 525e536.
Cooper, W.W., Seiford, L.M., Tone, K., 2007. Data Envelopment Analysis: a
Comprehensive Textwith Models, Applications,References and DEA-solver
Software,second ed.Springer,New York.
Cordero,J.M., Santin,D., Sicilia, E., 2015.Testing the accuracy ofDEA estimates
under endogeneity through a Monte Carlo simulation. Eur. J. Oper. Res. 244 (2),
511e518.
Dennis, N., 2000. Scheduling issues and network strategies for international airline
alliances.J. Air Transp.Manag.6 (2),75e85.
Doz, Y., Prahalad,C.K.,Hamel,G.,1990.In: Bartlett,C., Doz, Y., Hedlund,G. (Eds.),
Control, Change, and Flexibility: the Dilemma of Transnational Collaboration in
Managing the Global Firm.Routledge,London,United Kingdom.
Dyer, J.H., Kale, P., Singh, H., 2001. How to make strategic alliances work. MIT Sloan
Manag.Rev.42 (4),36e43.
Farrell,M.J., 1957.The measurement of productive efficiency.J. R. Stat.Soc.Ser.A
(III), 253e290.
H. Min, S.-J.Joo / Journal of Air Transport Management 52 (2016) 99e110 109

Fumero,F., 2004. Multiple solutions identification in data envelopment analysis.
Cent.Eur.J. Oper.Res. 12 (3),307e322.
Gayle,P.G.,2008. An empirical analysis of the competitive effects ofthe Delta/
Continental/Northwest code-share alliance.J. Law Econ.51 (4),743e766.
Gibbons,J.D.,1993.Nonparametric Statistics: an Introduction.Quantitative Appli-
cations in the Social Sciences.Sage Publications,No. 90, Newbury,CA.
Goh, K., Uncles,M., 2003. The benefits of airline global alliances:an empirical
assessment of the perceptions of business travelers. Transp. Res.A Policy Pract.
37 (6),479e497.
Goncalves,O.,2013.Efficiency and productivity of French ski resorts.Tour.Manag.
36, 650e657.
Gorin, T., Belobaba,P., 2004. Revenue managementperformance in a low-fare
airline environment: insights from the passenger origin-destination simu-
lator.J. Revenue Pricing Manag.3 (3), 215e236.
Haas,D.A., Murphy, F.H.,2003. Compensating for non-homogeneity in decision-
making units in data envelopment analysis.Eur.J. Oper.Res. 144 (3),530e544.
Hannegan,T.F.,Mulvey, F.P.,1995. Internationalairline alliances:an analysis of
code-sharing's impact on airlines and consumers.J. Air Transp.Manag.2 (2),
131e137.
Iatrou,K., Alamdari,F.,2005.The empirical analysis of the impact of alliances on
airline operations.J. Air Transp.Manag. 11 (3), 127e134.
Jiang, B., Talluri, S., Calantone,R., 2008. Determinants of interoutsourcing:an
analytical approach.Decis.Sci.39 (1),65e84.
Joo, S.J., Min, H., Kwon, I.W.G., Kwon, H., 2010. Comparative efficiencies of specialty
coffee retailers from the perspectives of socially responsible global sourcing. Int.
J. Logist.Manag.21 (3),490e509.
Joo, S.J., Nixon, D., Stoeberl,P.A., 2011.Benchmarking with data envelopment
analysis: a return on asset perspective.Benchmarking Int.J. 18 (4),529e542.
King, D.R., Dalton, D.R., Daily, K.M., Covin, J.G., 2003. Meta-analysesof post-
acquisition performance: indications of unidentified moderators. Strateg.
Manag.J. 25 (2), 187e200.
Kleymann,B., Seristo,H., 2001.Levels of airline alliance membership: balancing
risks and benefits.J. Air Transp.Manag.7, 303e310.
Lapre, M.A., Scudder, G.D., 2004. Performance improvement paths in the U.S. airline
industry: linking trade-offs to asset frontiers. Prod. Oper. Manag. 13 (2),
123e134.
Li, M.Z.F.,2000. Distinct features of lasting and non-lasting airline alliances.J. Air
Transp.Manag.6, 65e73.
Lin, E.T.J.,2008a.Route-based performance evaluation of Taiwanese domestic air-
lines using data envelopment analysis: a comment. Transp. Res. E 44, 894e899.
Lin, M.H., 2008b. Airline alliances and entry deterrence. Transp. Res. E 44, 637e652.
Mahajan,J., 1991.A data envelopmentanalytic modelfor assessing the relative
efficiency of the selling function.Eur.J. Oper.Res.53 (2), 189e205.
McGill, J.I., Van Ryzin, G.J., 1999.Revenue management:research overview and
prospects.Transp.Sci.33 (2),233e256.
Morrish, S.C.,Hamilton,R.T.,2002. Airline alliancesdwho benefits? J.Air Transp.
Manag.8 (6), 401e407.
O'Connell,J.F.,Williams, G.,2005.Passengers'perceptions of low cost airlines and
full service carriers: a case study involving Ryanair,Aer Lingus,Air Asia and
Malaysia Airlines.J. Air Transp.Manag. 11 (4),259e272.
Orme,C.,Smith,P., 1996.The potential for endogeneity bias in data envelopment
analysis.J. Oper.Res.Soc.47 (1),73e83.
Oum, T.H., Park, J.-H., Kim, K., Yu, C., 2004. The effect of horizontal alliances on firm
productivity and profitability: evidence from the global airline industry.J. Air
Transp.Manag.57,844e853.
Oum, T.H., Park, J.H., Zhang, A., 1996. The effects of airline code sharing agreements
on firm conduct and international air fares. J. Transp.Econ. Policy 30 (2),
187e202.
Oum,T.H.,Yu, C.,Zhang,A., 2001.Global airline alliances: international regulatory
issues.J. Air Transp.Manag.7 (1),57e62.
Palmer,B., 2005. Why 70% of the Mergers and Acquisitions Fail,and Why this
doesn't have to Happen to You.http://interlinkbusiness.com/artmergers.html.
retrieved on January 3,2012.
Park, J.H.,1997.The effects of airline alliances on markets and economic welfare.
Transp.Res.E Logist.Transp.Rev.33 (3), 181e195.
Porter,M.E., 1980.Competitive Strategy.The Free Press,New York.
Porter, M.E., 1990. The Competitive Advantage of Nations. The Free Press, New York.
Prince,J.T.,Simon,D.H., 2009.Multimarket contact and service quality: evidence
from on-time performance in the U.S.airline industry.Acad.Manag.J. 52 (2),
336e354.
Ruggiero,J., 1999.Efficiency estimation and error decomposition in the stochastic
frontier model: a Monte Carlo analysis.Eur.J. Oper.Res. 115 (3),555e563.
Schefczyk, M., 1993. Operational performance of airlines: an extension of traditional
measurement paradigms.Strateg.Manag.J. 14 (4),301e317.
Scheraga,C.A., 2004a. The relationship between operational efficiency and
customer service:a global study of thirty-eight large internationalairlines.
Transp.J. 43 (3),48e58.
Scheraga, C.A.,2004b. Operational efficiency versus financial mobility in the global
airline industry: a data envelopmentand Tobit analysis.Transp.Res. A 38,
383e404.
Sengupta,J.K., 1999.A dynamic efficiency model using data envelopment analysis.
Int. J. Prod.Econ.62, 209e218.
Sherman,H.D.,Ladino,G.,1995.Managing bank productivity using data envelop-
ment analysis.Interfaces 25 (2),60e73.
Sj€ogren,S., S€oderberg,M., 2011.Productivity of airline carriers and its relation to
deregulation, privatisation and membership in strategic alliances. Transp. Res. E
Logist.Transp.Rev.47 (2),228e237.
SKYTRAX, 2010. World Airline Star Ranking. http://www.airlinequality.com/
StarRanking/summary.htm.retrieved on September 15,2010.
Thanassoulis,E., 2001.Introduction to the Theory and Application of the Data
EnvelopmentAnalysis: a Foundation Textwith Integrated Software.Kluwer
Academic Publishers,Norwell, MA.
Tiernan, S., Rhoades,D., Waguespack,B., 2008. Airline alliance service quality
performancedan analysis of US and EU member alliances. J. Air Transp. Manag.
14,99e102.
US GeneralService Administration,2011.Code Share Factsheet.http://www.gsa.
gov/portal/content/103887.retrieved on January 27,2012.
Vowles, T.M., 2000. The geographic effects of US airline alliances. J. Transp. Geogr. 8
(4), 277e285.
Wagner,C.M., Huber,B., Sweeney,E., Smyth,A., 2005.B2B e-marketplaces in the
airline industry: process drivers and performance indicators.Int. J. Logist.Res.
Appl. 8 (4),283e297.
Wan, X., Zou, L., Dresner, M., 2009. Assessing the price effects of airline alliances on
parallel routes.Transp.Res.E 45,627e641.
Wassmer,U., Meschi,P.X.,2011.The effect of code-sharing alliance formations and
terminations on firm value: the role of co-specialization and scope extension.
J. Air Transp.Manag. 17 (5),305e308.
Wright, C.P., Groenevelt, H., Shumsky, R.A., 2010. Dynamic revenue management in
airline alliances.Transp.Sci.44 (1), 15e37.
Zacks Equity Research,2011.Airline Industry Outlook e January 2011.http://www.
zacks.com/stock/news/46361/AirlineþIndustryþOutlook.retrieved on January
31,2012.
Zou, L., Oum, T.H.,Yu, C., 2011.Assessing the price effects ofairline alliances on
complementary routes.Transp.Res.E 47,315e332.
H. Min, S.-J.Joo / Journal of Air Transport Management 52 (2016) 99e110110