Case 4

An evaluation of European airlines’
operational performance
Outline
 Abstract
 Motivation and Objective
 Research on airline efficiency
 Methodology
 Data and results
 Discussion
 Conclusion
Abstract
 This paper uses data envelopment analysis (DEA) to evaluate
the operational performance of a sample of AEA—
Association of European Airlines from 2000 to 2005,
combining operational and financial variables.
 In this paper an innovative DEA two-stage procedure
proposed by Simar and Wilson
[2007.Estimationandinferencein two stage, semi-parametric
models of productive efficiency. Journal of Econometrics
136, 31–64] is used.
Abstract
 In the first stage a DEA model isused to rank the airlines by
their overall efficiency.
 In the second stage a bootstrapped truncated regression is
used to evaluate the drivers of efficiency.
 The regressions test the roles played, respectively, by
population and network alliance in the efficiency of the
airlines. The implications of this research for managerial
purposes are then drawn.
Motivation and Objective
 An intensification of the battle for European passengers between
the national carriers and the low-cost airlines.
 To investigate the policies adopted by the AEA airlines to respond
to the new competitive environment.
 To analyze the role played by national markets aspects in
explaining differences revealed in the efficiency rankings.
Members of the Association of European
Airlines and characteristics
※ Star alliance, One word alliance, Sky team alliance
are some kind of airline strategic alliance
Research on airline efficiency
 The DEA permits the use of
multiple inputs and outputs
and does not impose any
functional form on the data.
 Many papers adopt both the
DEA and econometric frontier
approaches simultaneously.
Author
Cost model
Cavesetal.,1981,1984;
Windle, 1991;
Baltagi et al.,1995;
Oum and You,1998;
Liu and Lynk,1999.
Total factor
productivity approach
Bauer,1990;
Oum and You,1995;
Barbot et al.,2008.
Stochastic econometric
frontier model
Cornwell et al.,1990;
Good et al.,1993;
Sickles,1985;
Sickles et al.,1986;
Captain and Sickles,1997;
Coelli et al.,1999;
Inglada et al.,2006.
DEA model
Distexhe and Perelman,1994;
Good et al.,1995;
Adler and Golany,2001;
Fethi et al.,2001;
Scheraga,2004;
Greer,2008;
Bhadra,2008.
Methodology
 T characterized as the technology set, defined:
T   x, y   RN  RM : x  RN can produce y  RM  .
N inputs denoted with x
M outputs denoted with y
z means weight of factor
Methodology
 Farrell/Debreu-type output-oriented TE measure:


TE  x j , y j   max  :  x j , y j   T .
 In practice, T is unobserved and so were place it with its DEA-
estimate, :
Tˆ
n


N
M
k
 x, y   R  R :  zk ym  ym , m  1,..., M , 


k 1
ˆ
T  n
.
 z x k  x , i  1,..., N , z  0, k  1,..., n 
k i
i
k
 

k 1
Methodology
 In this paper, we assume constant returns to scale to gain more
discriminatory power in comparison between DMUs and then
analyze the returns-to-scale component in the second stage.
Methodology
 Regression analysis
TE j  a  Z j   j , j  1,..., n.
 Replaced
TEitsj DEA estimate
by
.
TEˆ j
TEˆ j  a  Z j   j , j  1,..., n.
where
 j ~ N  0,  2  such that  j  1  a  Z j , j  1,..., n.
Data and results
 Used data on European airline companies in the years from
2000 to 2005 (29 airline companies in 6 years = 174
observations), obtained in the AEA year book available in
(http://www.aea.be/).
Data and results
 Output
 RPK-operational revenue by passenger kilometer
 EBIT-earning before interest and taxes
 Input
 Number of employees
 Operational cost
 Number of planes
 Number of DMU is greater than three times the number of inputs
plus outputs.
 Using an output-orientation to determine whether an airline is
capable of producing the same level of output with less input.
Data and results
Data and results
 Trend is a yearly trend.
 Population is the country population of origin obtained in the
Eurostat statistics.
 Low cost are the low cost companies that are member of AEA.
 National airlines measure the influence on efficiency of being a
long-established, national flagship carrier.
Data and results-First stage
Data and results-First stage
There are significant differences in efficiency among the airlines
analyzed.
2. Almost all European airlines operated at a high level of pure
technical efficiency in the period.
3. All CRS(CCR) technically efficient airlines are also technically
efficient in VRS(BBC), signifying that the dominant source of
efficiency is scale.
4. According to the SE, almost all European airlines authorities are
efficient, while a small number are not.
1.
Data and results-Second stage
Two-stage DEA as suggested by Coelli et al.(1998).
Truncated bootstrapped regression by Simar and Wilson (2007).


i ,t  0  1Trendi ,t   2Trendi2,t  3 LogPopulationi ,t
  4 Low cos ti ,t  5 StarAlliancei ,t  6OneWorldi ,t
 represent
 8efficient
NationalaAirlines
thei ,tCCR
score of the airline
7 SkyTeam
i ,t  ii ,in
t period t.
 i ,t
Data and results-Second stage


The truncated model 2 drops the statistically insignificant
variable NationalAirlines.
Model 3 drops the variable SkyTeam, with the positive t-statistics
which are statically significant for all parameters.
Discussion
The efficiency is increase over the period, according to the trend.
And it increases at a decreasing rate.
2. The population contributes to the efficiency of the airline.
3. Low costs companies promote the efficiency of the European
airlines.
4. All alliances contribute to the technical efficiency, although the
SkyTeam is insignificant in the second-stage.
1.
Discussion



The companies with poor performance should adjust their
management process based on pure TE.
The variation in efficiency score may caused from the existence
of strategic groups and their differences in resources.
The manager of inefficient companies should…
Adopt a benchmark management procedure.
 Upgrade the quality of management practices.
 Adopt human resources policies.
 Pursue market-oriented strategies.

Conclusion


Use DEA-CCR model to determine relative efficient and the
bootstrapped truncated regression model explains the efficiency
drivers.
There is a growth trend in the. The demographic dimension of
the airline’s home country is important, representing economies
of scale and membership of a alliances is also important.