Description
The objective of this study is to investigate the technical and scale efficiencies of all airlines across service
type, size and ownership structures operating in India during 2006e2010. The variable returns to scale
(VRS) model of Data Envelopment Analysis (DEA) with two inputs and two outputs is used. For additional
insights, Input Efficiency Profiling (IEP) model of DEA is also used. The key findings are: a large majority
of budget airlines have been found to be efficient; while smaller private sector airlines have been efficient,
both the larger and smaller public sector airlines have also been efficient; the public sector airlines,
although incurring financial losses, are also operating at their most productive scale size; of the two
inputs, there is greater inefficiency with respect to the operating cost input. These findings are consistent
with other studies of airlines that found size, type of service and ownership to impact efficiency.
A DEA study of airlines in India
Ravi Kumar Jain
a, *
, Ramachandran Natarajan
b, 1
a
Symbiosis Institute of Business Management, Symbiosis International University, India
b
College of Business, Tennessee Technological University, USA
a r t i c l e i n f o
Article history:
Received 23 September 2013
Accepted 3 March 2015
Keywords:
Airlines in India
DEA
Input ef?ciency pro?ling
Technical ef?ciency
Scale ef?ciency
a b s t r a c t
The objective of this study is to investigate the technical and scale ef?ciencies of all airlines across service
type, size and ownership structures operating in India during 2006e2010. The variable returns to scale
(VRS) model of Data Envelopment Analysis (DEA) with two inputs and two outputs is used. For additional
insights, Input Ef?ciency Pro?ling (IEP) model of DEA is also used. The key ?ndings are: a large majority
of budget airlines have been found to be ef?cient; while smaller private sector airlines have been ef?-
cient, both the larger and smaller public sector airlines have also been ef?cient; the public sector airlines,
although incurring ?nancial losses, are also operating at their most productive scale size; of the two
inputs, there is greater inef?ciency with respect to the operating cost input. These ?ndings are consistent
with other studies of airlines that found size, type of service and ownership to impact ef?ciency.
© 2015 College of Management, National Cheng Kung University. Production and hosting by Elsevier
Taiwan LLC. All rights reserved.
1. Introduction
The civil aviation industry in India has come a long way since the
Air Corporation Act was repealed in the year 1994 allowing private
airlines to operate in scheduled services category. Several private
operators showed interest and were granted the status of sched-
uled carriers in the year 1995. However, many of those private
airlines soon shut down. Only Jet Airways and Sahara Airlines
survived and continued to offer scheduled services (Table 1). Till
the years 2001e2003, the Indian civil aviation sector was charac-
terized by the domination of government owned national carriers
like Indian Airlines (domestic) and Air India (international),
excessive regulations (e.g., control of aviation fuel) and taxation. All
this resulted in high cost of operating airline services which acted
not only as an entry barrier for the private players but also made air
travel an expensive affair giving it an image of elitist mode of travel.
However, the scenario has changed rapidly over the last decade
and the sector has witnessed a signi?cant growth not only in terms
of the entry of new private players but also in terms of increase in
passenger traf?c (16 percent CAGR between 2001 and 2011 with a
more rapid growth at an average of 19 percent in the latter half of
the decade from 2006 to 2010, Fig. 1). All this can be attributed
arguably to the introduction of structural reforms, entry of new
private airlines especially with different cost structures e.g., low
cost e no frills, airport modernizations, and improvement in ser-
vice standards (ICRA Research Report; March 2012). Also, as the
Indian economy began to grow faster, international passenger
traf?c into and out of India also began to grow (Fig. 1).
In the year 2003, India's ?rst low-cost carrier (LCC), Air Deccan,
entered the market. This was a landmark event for this industry,
which triggered an entry of several other private players. The year
2005 was also a watershed year for the Indian civil aviation sector
as several private low cost and full services carriers (such as
King?sher, IndiGo, SpiceJet, GoAir, and Paramount) commenced
their operations (Table 1). This led to the rapid increase in capacity
as measured by Available Seat Kilometers (ASK). Some of the new
entrants e Air Deccan, Spice Jet, GoAir and IndiGo e were pursuing
a different strategy and competed as low cost carriers (LCC).
The entry of the LCCs has signi?cantlyexpandedthe civil aviation
market by making air travel both affordable and accessible to the
middle class. Low fares offered by LCCs have made air travel very
attractive, prompting travelers to switch to air travel fromroad and
rail travel. In the early days of Air Deccan 40 percent of its passen-
gers were ?rst-time ?yers. SpiceJet, for instance, targeted passen-
gers who were traveling by air-conditioned classes in Indian
Railways. LCCs ushered in a newera of competition among airlines.
For instance, LCCs competitive pricing set off a price war with the
* Corresponding author. Kothur Mandal, Mamidipalli village, Mahabubnagar
District, Hyderabad 509216, India.
E-mail address: [email protected] (R.K. Jain).
Peer review under responsibility of College of Management, National Cheng
Kung University.
1
Tel.: þ1 931 372 3001.
HOSTED BY
Contents lists available at ScienceDirect
Asia Paci?c Management Review
j ournal homepage: www. el sevi er. com/ l ocat e/ apmrvhttp://dx.doi.org/10.1016/j.apmrv.2015.03.004
1029-3132/© 2015 College of Management, National Cheng Kung University. Production and hosting by Elsevier Taiwan LLC. All rights reserved.
Asia Paci?c Management Review 20 (2015) 285e292
incumbent Full Service Carriers (FSCs) such as Indian Airlines, Jet
Airways, and Air Sahara. This compelled the FSCs to discount their
fares by as much as 60e70 percent in some routes to match
the prices of LCCs. While the aggressive pricing strategy of LCCs has
both deepened and widened demand which has percolated to non-
metrohttp://www.domain-b.com/Aerospace/recommend.aspx
towns and Tier-II cities, this has hurt not only the pro?tability of
LCCs but also the revenues, margins, and market shares of FSCs. For
instance, Jet Airways, which controlled about 50 percent of the
domestic market in 2003, saw its share (including that of its
acquisition Jet Lite) drop to about one-third by 2007.
According to Kaul (2007) of Centre for Asia Paci?c Aviation, “…
the aggressive expansion of the LCC segment comes at a cost to the
whole sector. India's airlines are expected to post a combined loss
of approximately USD$500 million in the current ?nancial year
ending 31-Mar-07,”
Aggressive expansion of capacity and inability to control costs
were other factors which contributed to the mounting losses. These
conditions can lead only to two types of outcomes for the air-
linesdeither some of them go bust in a market shake-out or they
merge/get acquired by other airlines or business groups. Whereas
in the 1990s, many private carriers went bust, this time around the
industry has witnessed a wave of consolidation. Year 2007 became
a landmark year in the industry when major consolidation took
place (Table 1).
In 2008, there was a steep fall in the domestic air travel due to
the slowdown in the Indian economy, the H1N1 ?u scare, and the
terrorist attack in Mumbai in November 2008 (Fig. 1). There was
excess capacity all around and the airlines responded by developing
plans to lay off employees and by offering deeply discounted fares
to stimulate demand. Rival airlines Jet Airways and King?sher
formed a strategic alliance for code sharing and cutting costs. The
trend was to shift more capacity to LCC operations. The Indian
economy slowed in 2008e09 but there was no recession and as the
economy picked up in the second half of 2009, the demand for air
travel made a comeback (Fig. 1). Interestingly, the “pure” LCCs like
SpiceJet and IndiGo made pro?ts while carriers like King?sher, Jet
Airways, Air India and Indian Airlines operated both FSC and LCC
services, experienced huge losses. The anticipated cost reduction
due to synergies from the mergers could not be realized. In 2012,
the slowing economy affected the overall demand for air travel and
the domestic demand was down by 4.9 percent compared to 2011.
The airlines were also adjusting their capacities downwards to
more realistic levels (IATA, 2013).
Despite these setbacks and massive losses, the long term pros-
pects for the industry appear to be quite good. The reason why
foreign airlines such as Air Asia (Asia's largest lowcost airline wants
to take 49 percent stake in a joint venture with the Tata Group) and
Abu Dhabi-based Etihad (which announced to buy a stake in Jet
Airways in early 2013) ?nd the Indian market attractive is the huge
potential for growth in air travel (Kazmin, 2013). India accounts for
only about 2 percent of global air traf?c. Only about 4 percent of
India's population of over 1.1 billion people had ever been on a
?ight. Industry estimates suggest that the total passenger traf?c
will growfrom143 million in 2010e11 to 290e300 million by 2020
making India the third largest civil aviation market in the world. To
meet this demand, a ?eet of over 1000 aircrafts will be required
with around 350e400 operational airports across the country (The
Association of Private Airport Operators, 2013). This makes Indian
civil aviation industry very attractive. Low cost and budget airlines
are better positioned to dominate this market which is highly price
sensitive.
The above backdrop raises the research question whether the
developments described have brought about any signi?cant change
Fig. 1. Domestic and international passenger traf?c growth.
Source: Centre for Asia Paci?c aviation (2010) preparing for long term growth of Indian
aviation, New Delhi. Arushi and Stefan Drews (2011).
Table 1
Evolution of the Indian civil aviation industry.
Year Major milestones
1953 Nine Airlines existed including Indian Airlines & Air India
1953 Nationalization of all private airlines through Air Corporations Act;
1986 Private players permitted to operate as air taxi operators
1994 Air Corporation Act repealed; Private players allowed to operate scheduled services
1995 Jet, Sahara, Modiluft, Damania, East West granted scheduled carrier status
1997 4 out of 6 operators shut down; Jet & Sahara continue
2001 Aviation Turbine Fuel (ATF) prices decontrolled
2003 Air Deccan starts operations as India's ?rst LCC
2005 King?sher, SpiceJet, IndiGo, GoAir, Paramount start operations
2007 Industry consolidates; Jet acquired Sahara; King?sher acquired Air Deccan
2010 SpiceJet starts international operations
2011 IndiGo starts international operations, King?sher exits LCC segment. Air India and Indian
airlines merger completed.
2012 Government allows direct ATF imports, FDI proposal for allowing foreign carriers
to pick up to 49 percent stake under consideration. King?sher goes bankrupt.
2014 Falling international crude prices lead to sharp cuts in the price of aviation turbine fuel (ATF)
or jet fuel. Debt-ridden SpiceJet cancels ?ights, cuts third of its ?eet and seeks help from the government.
Source: ICRA research.
R.K. Jain, R. Natarajan / Asia Paci?c Management Review 20 (2015) 285e292 286
in productive ef?ciency of the airlines. This study investigates the
technical and scale ef?ciencies of all airlines across cost, size and
ownership structures e including budget and full-service, private
and state owned, large and small e operating in the country of-
fering scheduled services on domestic and international routes. The
scale at which they operate is identi?ed. While there are several
airline ef?ciency studies conducted globally, none to our knowl-
edge have attempted to study the Indian airlines industry. The
present study, by addressing this gap, contributes to the existing
body of literature in this area.
The rest of the paper is organized as follows. In Section 2, the
existing literature is reviewed. The research methodology is dis-
cussed in Section 3. The sample and the data sets are presented in
Section 4. Results are presented and discussed in Section 5. Con-
clusions follow in Section 6.
2. Literature review
The existing literature in this area has covered major airlines in
Asia, Europe and the USA across different cost structures (full ser-
vice carriers, budget airlines and low cost carriers); size (large and
small), and service type (domestic and international). Some of the
important studies are cited below.
Schefczyk (1993), used DEA technique, apparently for the ?rst
time, to analyze and compare operational performance of 15 in-
ternational airline using non-?nancial data such as Available Ton
Kilometer, Revenue Passenger Kilometer etc. The study demon-
strated that DEA can be a very useful tool to assess the technical
ef?ciency of international airlines which otherwise was dif?cult to
do using ?nancial data because: (1) most airlines lease a substantial
fraction of their ?eet; and (2) different accounting and taxation
rules in various countries result in different impacts of leased assets
on pro?t and balance-sheet information of airlines.
Good, Roller, and Sickles (1995) examined the performance of
the eight largest European and the eight largest American airlines
for a ten year period between 1976 and 1986 using two methods e
one a parametric using statistical estimation and another non-
parametric (DEA) using linear programming. The authors
observed discrepancy in the productive ef?ciency of European
airlines even under the conditions of deregulations and liberaliza-
tion of the airline industry. The study suggested that if European
airlines were as productively ef?cient as their American counter-
parts then they could save approximately $4 billion per year (in
1986 dollars).
Fethi, Jackson and Jones (2001) studied ef?ciency across a panel
of 17 European airlines in the 1990s during the early phase of
liberalization using stochastic DEA constructs (which develops
production frontiers that incorporate both inef?ciency and sto-
chastic error). They use the Land, Lovell, and Thore (1993) model
incorporating information on the covariance structure of inputs
and outputs. Stochastic DEA was chosen to achieve a closer envel-
opment of the mean performance of the companies in the sample
and reduce, if not eliminate, the effect of extreme outliers. This
model computes the relative ef?ciencies after allowing for sto-
chastic error. They concluded that the airlines that were ef?cient in
1995 resembled those that were ef?cient in 1993 but not those in
1991. Interestingly, they also found that larger size airline com-
panies were ef?cient thus implying that size of an airline does
contribute to their degree of ef?ciency.
Scheraga (2004) investigated the structural drivers of opera-
tional ef?ciency as well as the ?nancial position of airlines in a
industry ridden by ?nancial crises post-September 11, 2001. The
study analyzed 38 airlines from North America, Europe, Asia and
the Middle East and across different cost structures (low cost and
budget airlines included for the ?rst time in such a study) to
investigate whether relative operational ef?ciency implied superior
?nancial performance. The author used DEA technique to derive
ef?ciency scores for individual airlines followed by Tobit analysis to
investigate the underlying structural drivers of ef?ciency. The study
concluded that the traditional framework developed in the litera-
ture still provided reasonable explanatory power for realized
relative operational ef?ciency. The author also observed that rela-
tive operational ef?ciency did not inherently imply superior
?nancial performance and airlines that had chosen relatively ef?-
cient operational strategies found themselves vulnerable in terms
of ?nancial performance and thus suffered the consequences in the
post-September 11 environment.
Zhu (2011) measured performance ef?ciency of 21 airlines in the
US during 2007e2008 using the centralized ef?ciency model, a
two-stage process used to optimize performance simultaneously,
instead of a standard DEA technique thus generating ef?ciency
decomposition for the two individual stages. For instance, in stage
one, resources (fuel, salaries, and other factors) are used to main-
tain the ?eet size and load factor, and in stage two, the revenue
generating capacity is assessed given the ?eet size and load factors.
The study considers airlines with different cost structures unlike
most of the earlier studies that analyzed only major full service
airlines.
Lee and Worthington (2010) is one of the very few studies in
that included airlines across different cost structures (full services,
low cost and budget airlines), and service types (international and
domestic services). The study investigates whether the inclusion of
low-cost airlines in the data set of international and domestic air-
lines has any impact on the ef?ciency scores of those otherwise
‘ef?cient’ full service airlines. The authors analyzed 53 airlines in
the year 2006 using the non-parametric technique of DEA to
investigate their technical ef?ciency. The study reveals that the
majority of budget airlines are ef?cient relative to the more pres-
tigious full service airlines. Moreover, most airlines identi?ed as
inef?cient are so primarily because of the overutilization of non-
?ight assets.
Michaelides, Belegri-Roboli, Karlaftis, and Marinos (2009) have
employed both Stochastic Frontier Analysis (SFA) and DEA using a
panel data set of 24 world's largest network airlines to estimate
technical ef?ciency in international air transport for the period
1991e2000. The authors observed that the airlines achieved con-
stant returns to scale with technical ef?ciencies ranging from 51%
to 97%. They also observed that ownership (private or public) does
not affect the technical ef?ciency of the airlines. Further, it is
interesting to note that results from both SFA and DEA did not vary
signi?cantly.
Tofallis (1997) used a modi?ed version of DEA called input ef-
?ciency pro?ling method to assess the technical ef?ciency of 14
major full service airlines across all ?ve continents using the same
data set used by Schefczyk (1993). The study demonstrated that
input ef?ciency pro?ling gives better results when compared to
standard DEA in measuring technical ef?ciency of the airlines.
Among the above cited studies and several others (not cited in
detail here) such as Banker and Johnston (1994), Charnes, Galleous,
and Li (1996), Gillen and Lall (1997), Alam, Semenick, and Sickles
(1998), Adler and Golany (2001), and Coli, Nissi and Rapposelli
(2011), none have included airlines operating in an important
market like India. While such omission is perhaps justi?able for
academic studies till 2003 as the Indian civil aviation sector was
very limited with only a few players operating and primarily
dominated by government owned airlines, the more recent studies
also seem to have ignored the Indian civil aviation sector. The
present study contributes to the body of literature in this area by
analyzing all the airlines operating in domestic and international
services in full services, budget/low cost segments.
R.K. Jain, R. Natarajan / Asia Paci?c Management Review 20 (2015) 285e292 287
3. Research methodology
While there are many ways to de?ne and measure the pro-
ductive ef?ciency of an organizational/industrial unit, also called
Decision Making Unit (DMU), there are two widely employed
techniques namely Data Envelopment Analysis (DEA) and Sto-
chastic Frontier Analysis (SFA) (Farrell, 1957). These methods pro-
vide a measure of the technical ef?ciency of a DMU in terms of
radial distance from the best unit on the production frontier rep-
resented by the production function of the ef?cient units. Lesser the
distance greater is the ef?ciency.
DEA is a non-parametric method as it does not impose an
explicit functional form for the production frontier unlike an
econometric method like SFA. DEA method assumes that all de-
viations from the ef?cient frontier are due to inef?ciencies,
whereas the SFA technique assumes that deviations from the ef?-
cient frontier can be a random error or due to realization of in-
ef?ciency or a random shock. In short, DEA measures ef?ciency
relative to a deterministic frontier using linear programming
techniques to envelop observed input/output vectors as tightly as
possible rather than explicitly stating the formof the frontier based
on the de?ned relationship between inputs and outputs.
As seen in the literature on measurement of airline ef?ciency,
DEA method is widely employed for assessing the productive ef?-
ciencies of airlines across the world. This study uses the Banker,
Charnes, and Cooper (BCC) model (Banker, Cooper, Seiford, & Zhu,
2004) which allows for variable returns to scale (VRS) and the
input ef?ciency pro?ling (IEP) model of DEA (Tofallis, 1997).
Globally, airlines industry is characterized by high startup costs
and scale economies. Of course this does not imply that each airline
in the industry will be using a production technology that exhibits
economies of scale due to increasing returns to scale. The tech-
nology for converting inputs to outputs can exhibit different
returns to scale based on the levels of outputs. Some airlines could
be on a production possibility frontier where locally it has dimin-
ishing or increasing returns to scale. Therefore, the scale at which a
particular airline is operating can only be determined empirically.
In the context of DEA, this implies using the Banker, Charnes, and
Cooper (BCC) model (see Banker et al., 2004) which allows for
variable returns to scale (VRS) is more appropriate.
Allowing for variable returns to scale in constructing the ef?-
cient frontier enables separating the ef?ciency of scale from pure
technical ef?ciency. The ef?cient frontiers for the constant returns
to scale (CRS) and the variable returns to scale (VRS) speci?cations
are illustrated in Fig. 3 of the Appendix. When the production
technology allows variable returns to scale (VRS) at different points
on the frontier of the production possibility set, the technical ef?-
ciency (either input- or output-oriented) of a DMU will differ from
its scale ef?ciency. Technical ef?ciency is measured by comparing
the average productivity of a DMU with the corresponding average
productivity at its input- or output-oriented projection onto the
VRS frontier. Scale ef?ciency, on the other hand, compares the
average productivity at the ef?cient input- or output-oriented
projection with the maximum average productivity attained at
the Most Productive Scale Size (MPSS) on the VRS frontier. Fig. 4 in
the Appendix illustrates these concepts. For the de?nition and
discussion of the concept of MPSS see Banker et al. (2004).
The BCC model considers all the inputs together whereas the
effect of each input is considered separately in the IEP model. The
DEA models in general and BCC model in particular has a linear
weighted combination of all the inputs in the denominator of the
objective function. This implies that all the inputs are perfectly
substitutable with the weights as the marginal rates of substitution.
The mathematical formulations (linear programming models) of
these models are given in the Appendix.
The study considers the following two inputs and two outputs,
consistent with the ones used in the earlier academic studies, (Lee
& Worthington, 2010; Schefczyk, 1993; Scheraga, 2004) but ex-
cludes the third input, ‘non-?ight assets’ (NFA), in airlines opera-
tions. Lee and Worthington (2010) in their recent study have
observed that excluding NFA from input set does not seem to have
any signi?cant affect on the overall ef?ciency of the airlines.
Input 1: X1 ¼ Total Available Ton Kilometer (ATKM) (re?ects
airline ?eet capacity in tonnage)
Input 2: X2 ¼ Operating Cost (OC) (all operating cost excluding
aircraft rentals, depreciation and amortization and other capital
expenditure)
Output 1: Y1 ¼ Revenue Passenger Kilometer Performed
(RPKM) (Kilometers performed of revenue paying passenger
traf?c)
Output 2: Y2 ¼ Non-Passenger Revenue (NPR) (revenue from
non-passenger traf?c such as carrying cargo)
The rationale for using the input ef?ciency pro?ling (IEP) model
is as follows. The input pro?ling model is more discriminating than
the conventional DEA model when the inputs are not substitutes.
This applies in the case of airlines where the two inputs i.e., avail-
able ton kilometers (a capacity measure) and operating costs
(which include fuel and labor costs) are not substitutes. Before
Tofallis (1997), Kopp (1981) and Kumbhakar (1988), in their pro-
ductivity studies -which do not use DEA e have considered effects
of each input separately. As Kumbhakar (1988) puts it “knowing the
magnitude of (overall) technical ef?ciency is not enough. It is
important to knowwhich inputs are causing the inef?ciency and to
what extent”. Input pro?ling helps in isolating the inef?ciencies
with respect to each input. This information is useful for managers
to set priorities and input-speci?c targets for improvements.
Input ef?ciency pro?ling model of DEA is particularly relevant to
the situation in the airline industry in India for the following
reasons.
1) The airlines in India want to expand their ?eet by signi?cant
margins (see Table 2 for aircraft on order) (Arushi & Drews,
2011). For instance, IndiGo's aircraft orders in 2011 is more
than six times its ?eet size! This raises questions about utili-
zation of the ATKM input.
2) Fuel costs make up around 40%e50% of total operating expenses
of airline industry in India (Arushi & Drews, 2011). In compari-
son, this cost factor accounted for 33% on average for the global
airline industry in 2008 and 2012 and 26% in 2009 and 2010,
respectively, according to the International Air Transport Asso-
ciation (IATA, 2012).
3) Cost reduction is critical for all airlines in India. Aviation Turbine
Fuel (ATF) costs 60% more in India because of high sales taxes of
state and central governments (Arushi & Drews, 2011). Most
airlines in India have not been able to control costs. The term
LCC is somewhat misleading in India. Among the LCCs in India,
as shown in Fig. 2, in 2007, SpiceJet had the lowest unit cost at
6.2 U.S. cents per Available Seat Kilometer (ASK), which is
comparable with Southwest, Easy Jet, and Jet Blue (CAPA, 2007).
But this is more than twice that of the best performers like Air
Asia with unit cost of slightly over 3 cents per ASK. Another
Indian carrier Jet Lite (previously Air Sahara before it was ac-
quired by Jet Airways in 2007) shows up in the Figure at 7.2
cents per ASK. According to Bill Franke (2007), the then Man-
aging Director of leading airline investment ?rm Indigo Part-
ners, “There is not a single airline in India that operates a true
low cost structure. Under the current conditions, it's not
possible.” Similar views were expressed by Naresh Goyal, the
R.K. Jain, R. Natarajan / Asia Paci?c Management Review 20 (2015) 285e292 288
founder and chairman of Jet Airways who questioned the
feasibility of LCCs in India because of lack of infrastructure and
skills. “Here there are no alternative airports.” “India has
nothing called low-cost, only low-fare and low-margin. This is
irrational pricing which will make the whole industry sick,”
(Economist, 2007). Air India, which has accumulated huge los-
ses, is trying to turn the situation around by cutting wages and
using fuel ef?cient aircraft (Boeing's 787 Dreamliner) in some of
its routes, (Sanjai, 2012).
Therefore, in the Indian context, it is worth analyzing the in-
ef?ciencies of the ATKM and OC inputs separately.
4. Sample and datasets
The sample includes all commercial scheduled airline carriers,
with different ownerships (public and private) and cost structure
(full service and lowcost), offering scheduled services on domestic
and international routes operating during the ?ve year period be-
tween 2006 and 2010. A total of twelve airlines were considered for
analysis. Focusing on this period will allow the analysis of the ef-
?ciencies of the new private operators who have entered the in-
dustry since 2005.
Most of the data is drawn fromvarious reports and air transport
statistics (ATS) provided by the Directorate General of Civil Aviation
(DGCA). Further, annual reports and balance sheets of these airlines
were also referred to supplement the data with accounting
Fig. 2. Low Cost Carrier (LCC) unit costs per ASK (including fuel).
Source: Center for Asia Paci?c Aviation (CAPA), 2007.
Fig. 3. Ef?cient frontiers for constant and variable returns to scale DEA speci?cations. Fig. 4. Technical and scale ef?ciencies and returns to scale.
R.K. Jain, R. Natarajan / Asia Paci?c Management Review 20 (2015) 285e292 289
numbers wherever required. Table 3 shows the input and output
data (averaged for the period of study) used to derive ef?ciency
scores.
5. Results
Table 4 provides, in percentages, the technical ef?ciency scores
for the VRS model (denoted by VRS TE), the scale ef?ciencies (SE),
the input ef?ciency pro?ling (IEP) (denoted by IEP (ATKM), IEP
(OC)) respectively for 2006e10. It also identi?es the returns to scale
at which the airline is operating.
Results from Table 4 suggest the following interpretations.
1. Of the six new entrants after 2005 three of them GoAir, In-
diGo, and Air India Express are found to be ef?cient with VRS
technical ef?ciency scores of 100 percent. Two out of the
remaining three new entrants which were inef?cient were
Full Service Carriers (FSC).
2. Budget airlines such as Air Deccan, GoAir, IndiGo, and Air
India Express are ef?cient with 100 percent scores (Table 4).
This indicates that in the Indian context also the cost struc-
ture of an airline in?uences its degree of productive ef?-
ciency, which is consistent with the observations made by
Lee and Worthington (2010).
3. Relatively larger FSCs, Air India, Indian Airlines, and Jet Air-
ways are ef?cient. These results are consistent with the
observation made by Fethi et al. (2001) that size of an airline
does contribute to the degree of ef?ciency.
4. While smaller private sector airlines have been ef?cient, all
the public sector airlines (the larger Air India and Indian
Airlines and the smaller Air India Express) have also been
ef?cient. These results support the conclusions of
Michaelides et al. (2009) that ownership does not affect the
technical ef?ciency of airlines.
5. When the results of the input ef?ciency pro?ling model are
considered, the public sector airlines Air India Express, Air
India and Indian Airlines have performed better than the
private sector airlines with respect to the Operating Cost (OC)
input.
6. Except for Air India and Air India Express, there is greater
inef?ciency with respect to the Operating Cost input, when
the ef?ciencies of the two inputs are compared. The ef?-
ciency for this input was as lowas 25.7 percent (Paramount).
The average ef?ciency with respect to ATKM input is 86.6
while the average ef?ciency with respect to the OC input is
67.6 (Table 4). One possible interpretation for such results is
that while the price war triggered by LCCs such as Air Deccan,
increased the demand for air travel and therefore resulted in
better capacity utilization for most of the airlines, whereas
the increase in fuel cost (which accounts for almost half of
the operating expenses for airline in India), which is gener-
ally beyond the control of the airlines, took its toll on their
operating cost.
7. The IEP scores can help identify areas for improvements in a
way that the results of the standard DEA model cannot. For
instance, for Jet Lite, the IEP model score for input 1 is 88.50
percent, which implies that ATKM can be reduced to 88.50
percent of the current level of that input to achieve current
level of the output. However, according to Table 4, there is
greater inef?ciency in the use of second input, i.e., Operating
Cost (OC), for Jet Lite can reduce that input to 48.9 percent of
its current level and still maintain the current level of output.
The IEP scores suggest that the airlines have to focus on
improving ef?ciency of the Operating Cost input.
8. The returns to scale classi?cation given in Table 4 is based on
the results given in Banker et al. (2004). The DEAP software
(2013) used for this study provides the classi?cation as one
of the outputs of DEAwith VRS speci?cation (the BCC model).
The SE scores in Table 4 indicate that of the new airlines,
GoAir, IndiGo and Air India Express have been operating at
their Most Productive Scale Size (MPSS) in 2006e2010. Air
India, Indian Airlines and Air Deccan have also been able to
operate at MPSS. The means of the SE scores in Table 4
indicate that the average percentage of deviation from
MPSS is 3.8 percent (100e96.2).
9. Jet Airways is classi?ed as operating in diminishing returns
to scale (DRS) during 2006e2010. Usually DRS signify that
either an input such as non-?ight assets (NFA) is not
accounted for or some inputs e.g., pilots, are limited. DRS can
be a short term phenomenon e in the longer term such
limitations can be removed. In airlines, the key input of ?eet
capacity is characterized by indivisibilities. Each aircraft has a
Table 2
Market share, ?eet size, aircrafts on order of scheduled domestic airlines in 2011.
Airline Market share (%) Fleet Orders
King?sher 19 66 130
IndiGo 18.70 39 241
Jet Airways 18 96 29
Indian Airlines 15.80 119 30
SpiceJet 13.80 26 66
Jet Lite 8.1 19 8
GoAir 6.6 10 10
Total 100 375 514
Table 3
Average input e output data for period 2006e2010.
DMU ATKM OC RPKM NPR
Jet Airways 3112.50 81,910.30 16,583.44 12,647.38
Jet Lite 514.94 20,470.22 3827.79 1150.86
Paramount Airways 27.30 2090.13 205.82 12.68
KF Airlines 751.72 38,295.64 5569.27 1657.74
Alliance Air 74.00 4465.15 482.68 942.88
SpiceJet 529.84 14,155.04 3938.75 1002.22
Air India Express 560.80 9854.94 3770.60 1730.54
Air Deccan 578.63 22,899.20 4859.66 2643.33
Air India 4214.60 13,6102.42 18,304.29 40,135.02
Indian Airlines 1693.75 118,040.35 11,569.95 29,004.75
GoAir 151.38 5549.65 1260.96 365.65
IndiGo Airlines 585.35 14,156.03 4311.60 992.18
Table 4
Results for period 2006e2010.
Airlines VRS TE SE IEP (ATKM) IEP (OC)
Jet Airways 100.0 78.0 DRS 65.10 70.0
Jet Lite
a
88.6 99.9 IRS 88.50 48.9
Paramount Airways 100.0 89.8 IRS 89.80 25.7
KF Airlines 92.7 95.2 IRS 88.20 38.0
Alliance Air 100.0 92.0 IRS 90.50 73.2
Spice jet 98.6 99.8 IRS 88.50 72.7
Air India Express 100.0 100.0 MPSS 80.10 100.0
Air Deccan 100.0 100.0 MPSS 100.00 60.5
Air India 1.000 100.0 MPSS 61.70 100.0
Indian Airlines 100.0 100.0 MPSS CRS 100.00 83.3
Go Air 100.0 100.0 MPSS 99.20 59.4
IndiGo airlines 100.0 100.0 MPSS MPCRS 87.70 79.6
Mean 98.3 96.2 86.60 67.60
DRS e Decreasing Returns to Scale; IRS e Increasing Returns to Scale; MPSS e Most
Productive Scale Size; TE e Technical Ef?ciency; SE e Scale Ef?ciency.
a
Jet Lite was called Sahara before the latter was taken over by Jet Airways in the
year 2007.
R.K. Jain, R. Natarajan / Asia Paci?c Management Review 20 (2015) 285e292 290
minimum capacity and each route the airlines decided to
operate also has a minimum distance. Therefore, ?eet ca-
pacity in terms of ATKM will increase in lumps. This could
afford opportunities for increasing returns to scale (IRS) if the
increase in ATKM is more than matched by increase in RPKM
on the demand side. Economies of scale opportunities are
also available with respect to the other input, i.e., Operating
Costs (OC). For instance, inventory levels and costs increase
less than proportionally for any increase in demand, and
increase in costs of construction or leasing of warehouses is
also less than proportional to the increase in the volume of
space required. Such opportunities for IRS are more likely in
the initial stages of expansion of airlines. This has been
observed in the case of Paramount Airlines, a new airline in
the sample, which operated in the southern region offering
only business class service.
10. The scale at which an airline operates also depends on its
expansion path relative to increase in demand for its ser-
vices. Many airlines in India have ambitious expansion ?eet
plans (see Table 2). But ?eet capacity expansion has to be
commensurate with expansion of demand. Both Air Deccan
and King?sher from their inception have expanded their
operations very aggressively aiming to become national
carriers. In the case of Air Deccan, the ?rst discount airline in
India, the demand increased because of its low fares but in
the case of King?sher (KF) which operated as a Full Service
Carrier (FSC) charging higher fares, it did not. Due to its
reckless expansion and its inability to control costs, King-
?sher went bankrupt in 2012.
11. Maintaining a dynamic balance between capacity and de-
mand is relatively easier for mature airlines such as Air India
and Indian Airlines because of their sheer size. The ?eet ca-
pacity increments are smaller in percentage terms relative to
overall existing capacity and same would be true of any
?uctuations in demand. These airlines are expected to
exhibit CRS and this is validated in Table 4 by the fact that
they are operating at their MPSS.
6. Conclusions
The following broad conclusions can be drawn from the study.
1. Budget airlines and low-cost airlines are found to be more
ef?cient when compared to the full service airlines in the private
sector. This supports the conclusion drawn by Lee and
Worthington (2010) that cost structure of airlines matters in
their ability to achieve better productive ef?ciencies.
2. While the new, private, low-cost airlines have been ef?cient, all
the public sector airlines have also been ef?cient. This is inter-
esting because ?nancially all the public sector airlines have only
made losses during the period of the study. This conclusion
supports Scheraga (2004) that relative technical ef?ciency does
not imply superior ?nancial performance.
3. It is also found that size and scale of the airline in?uences their
ef?ciency scores, thus supporting the conclusions drawn by
Fethi et al. (2001). Larger airlines (which in India happen to be
publicly owned) have been found to be ef?cient in the VRS
model. In the Input Ef?ciency Pro?ling models, they were found
to be more ef?cient when compared to the smaller airlines.
4. Input Ef?ciency Pro?ling model results indicate that, of the two
inputs considered in the study, almost all the airlines need to
achieve greater ef?ciency in the use of operating costs (of which
fuel costs are a signi?cant component). Such inef?ciency with
respect to operating cost can be arguably attributed to the cost
of highly regulated Aviation Turbine Fuel.
5. Further studies e along the lines of Good et al. (1995) in the
context of European airlines e need to be conducted to assess if
liberalization of the airline sector and the consequent entry of
new airlines in India has contributed to the improvement of
productive ef?ciencies of the airlines.
Appendix
1. The DEA formulation, according to Banker, Charnes, and Cooper
(1984) of input minimization with an assumption of variable
returns to scale (VRS) to calculate the ef?ciency scores is given
below.
The relative ef?ciency of a Decision Making Unit DMUo is ob-
tained from the following linear programming (LP) model:
Minq
0
subject to
q
0
x
ij0
À
X
n
j¼1
l
j
x
ij
0; i ¼ 1; …; m
X
n
j¼1
l
j
y
rj
y
rj0
; r ¼ 1; …; s
X
n
j¼1
l
j
¼ 1; j ¼ 1; …; nðconvexity constraintÞ
l
j
0; j ¼ 1; …; n
where, y
rj
is the amount of the r
th
output of DMU j, x
ij
is the amount
of the i
th
input to DMU j, l
j
are the weights of DMU j and q is the
shrinkage factor.
The model seeks a set of non-negative l values which add up to
1 and which minimizes q
0
toq
*
0
and identi?es a point within the
production possibility set which uses the lowest proportion q
*
0
of
input levels of DMU j while offering output levels which are at least
as high as those of DMU j. This point is a composite DMU corre-
sponding to the linear combination of ef?cient DMUs:
X
n
j¼1
l
*
j
x
ij
;
X
n
j¼1
l
*
j
y
rj
; with i ¼ 1; :::; m and r ¼ 1; :::; s
It can be said that:
X
n
j¼1
l
*
j
x
ij
;
X
n
j¼1
l
*
j
y
rj
outperformsðq
0
; X
j0
; y
jo
Þ when q
)
0
TE
NIRS
while DMU Q
is too big and operating at DRS because TE
VRS
¼ TE
NIRS
, and DMU C
is at optimal size (MPSS) because TE
VRS
¼ TE
CRS
DMU C is on both the CRS and VRS ef?cient frontiers. It is
operating at a level of output where IRS changes to DRS and has the
maximum possible economy of scale. It has a scale ef?ciency of 100
percent and is operating at its Most Productive Scale Size (MPSS).
References
Adler, N., & Golany, B. (2001). Evaluation of deregulated airline networks using data
envelopment analysis combined with principal component analysis with an
application to Western Europe. European Journal of Operational Research, 132(2),
260e273.
Alam, I., Semenick, M., & Sickles, R. C. (1998). The relationship between stock
market returns and technical ef?ciency innovations: evidence from the US
Airline Industry. The Journal of Productivity Analysis, 9(1), 35e51.
Arushi, & Drews, S. (2011). Aviation and environment (Working paper). New Delhi,
India: Center for Science and Environment.
Banker, R., Charnes, A., & Cooper, W. W. (1984). Some models for estimating tech-
nical and scale inef?ciencies in data envelopment analysis. Management Sci-
ence, 30, 1078e1092.
Banker, R., Cooper, W. W., Seiford, L. M., & Zhu, J. (2004). Returns to scale in DEA. In
W. W. Cooper, L. M. Seiford, & J. Zhu (Eds.), Handbook of data envelopment
analysis (pp. 41e70). Boston, MA: Kluwer Academic Publishers.
Banker, R., & Johnston, H. (1994). Evaluating the impacts of operating strategies on
ef?ciency in the US airline industry. In A. Charnes, W. W. Cooper, A. Lewin, &
L. Seiford (Eds.), Data envelopment analysis: Theory, methodology, and application
(pp. 97e128). Boston, MA: Kluwer Academic Publishers.
Centre for Asia Paci?c Aviation (CAPA) Report (2007).http://indiaaviation.aero/
news/airline/7591/59/Unit-costs-fallinge-airlines-in-drive-for-ef?ciency-.
Accessed 27.03.13.
Charnes, A., Galleous, A., & Li, H. (1996). Robustly ef?cient parameter frontiers: an
approximation via the multiplicative DEA model for domestic and international
operations of the Latin American airline industry. European Journal of Opera-
tional Research, 88, 525e536.
Coli, M., Nissi, E., & Rapposelli, A. (2011). Ef?ciency evaluation in an airline com-
pany: some empirical results. Journal of Applied Sciences, 11, 737e742.
DEAP. (2013). DEAP V2.1 a data envelopment analysis (Computer) program. http://
www.uq.edu.au/economics/cepa/deap.php. Accessed 27.03.13.
Farrell, M. J. (1957). The measurement of productive ef?ciency. Journal of Royal
Statistical Society, Series A (General), 120, 253e290.
Fethi, M. D., Jackson, P. M., & Weyman-Jones, T. G. (September 2001). European
airlines: a stochastic DEA study of ef?ciency with market liberalization. Paper
presented at the Seventh European Workshop on Ef?ciency and Productivity
Analysis (7EWEPA), University of Oviedo, Oviedo, Spain.
Franke, B. (2007). No true LCCs in India e indigo partners. In Keynote address at the
4th annual India and Middle East Low Cost Airline Symposium. http://
centreforaviation.com/members/download/avan/20070614.htm. Accessed
27.03.13.
Gillen, D., & Lall, A. (1997). Developing measures of air transport productivity and
performance: an application of data envelopment analysis. Transportation
Research, 33, 261e273.
Good, D. H., Roller, L.-H., & Sickles, R. C. (1995). Airline ef?ciency differences be-
tween Europe and the US: implications for the pace of EC integration and do-
mestic regulation. European Journal of Operations Research, 80, 510e518.
ICRA Research Report. (2012). Indian aviation industry: Through turbulent times, FDI
relaxation alone not a game changer.http://www.icra.in/Files/ticker/Indian%
20Aviation%20Industry%20%28NEW%29.pdf. Accessed 27.03.13.
International Air Transport Association (IATA) (2012). Fact sheet: Fuel.http://www.
iata.org/pressroom/facts_?gures/fact_sheets/Pages/fuel.aspx. Accessed 27.03.13.
International Air Transport Association (IATA) (2013). Air passenger market analysis.http://www.iata.org/whatwedo/Documents/economics/Passenger-Analysis-
Jan-2013.pdf. Accessed 27.03.13.
Kaul, K. (2007). Aviation consolidation looming in India. Centre for Asia Paci?c
Aviation (CAPA).http://www.centreforaviation.com/news/press-releases/2007/
03/07/aviation-consolidation-looming-in-india-outlook-2007-report-released/
page1. Accessed 27.03.13.
Kazmin, A. (February 21, 2013). Foreign airlines see beyond clouds in India. Financial
Times.http://www.ft.com/cms/s/0/38d491ee-7be5-11e2-99f0-00144feabdc0.
html#axzz2f9Va9mue. Accessed 27.03.13.
Kopp, R. J. (1981). The measurement of productive ef?ciency: a reconsideration.
Quarterly Journal of Economics, 96(3), 477e504.
Kumbhakar, S. C. (1988). Estimation of input-speci?c technical and allocative in-
ef?ciency in stochastic frontier models. Oxford Economic Papers, 40, 535e549.
Land, K. C., Lovell, C. A. K., & Thore, S. (1993). “Chance-constrained Data Envelop-
ment Analysis”. Managerial and Decision Economics, 14, 541e554.
Lee, B. L., & Worthington, A. C. (2010). The relative ef?ciency of international, do-
mestic, and budget airlines: Nonparametric evidence (Working Paper No. 2010-
02). Australia: Grif?th University.
Michaelides, P. G., Belegri-Roboli, A., Karlaftis, M., & Marinos, T. (2009). Interna-
tional air transportation carriers: evidence from SFA and DEA technical ef?-
ciency results (1991e2000). European Journal of Transportation and
Infrastructure Research (EJTIR), 4(9), 347e362.
Sanjai, P. R. (2012). Air India to cut wages to bring down costs. Live mint.http://www.
livemint.com/Companies/xraOS5yWgAnS9P5Y2n0yUJ/Air-India-to-cut-wages-
to-bring-down-costs.html. Accessed 27.03.13.
Schefczyk, M. (1993). Operational performance of airlines: an extension of tradi-
tional measurement paradigms. Strategic Management Journal, 14(4), 301e317.
Scheraga, C. A. (2004). Operational ef?ciency versus ?nancial mobility in the global
airline industry: a data envelopment and Tobit analysis. Transportation Research,
Part A, 38, 383e404.
Taking Flight. (2007, September 6). Face value. The Economist.
The Association of Private Airport Operators. (2013). A paper on civil aviation policy.http://civilaviation.gov.in/cs/groups/public/documents/newsandupdates/moca_
001404.pdf. Accessed 05.09.13.
Tofallis, C. (1997). Input ef?ciency pro?ling: an application to airlines. Computers
and Operations Research, 24(3), 253e258.
Zhu, J. (2011). Airlines performance via two-stage network DEA approach. Journal of
CENTRUM Cathedra, 4(2), 260e269.
R.K. Jain, R. Natarajan / Asia Paci?c Management Review 20 (2015) 285e292 292
doc_968355221.pdf
The objective of this study is to investigate the technical and scale efficiencies of all airlines across service
type, size and ownership structures operating in India during 2006e2010. The variable returns to scale
(VRS) model of Data Envelopment Analysis (DEA) with two inputs and two outputs is used. For additional
insights, Input Efficiency Profiling (IEP) model of DEA is also used. The key findings are: a large majority
of budget airlines have been found to be efficient; while smaller private sector airlines have been efficient,
both the larger and smaller public sector airlines have also been efficient; the public sector airlines,
although incurring financial losses, are also operating at their most productive scale size; of the two
inputs, there is greater inefficiency with respect to the operating cost input. These findings are consistent
with other studies of airlines that found size, type of service and ownership to impact efficiency.
A DEA study of airlines in India
Ravi Kumar Jain
a, *
, Ramachandran Natarajan
b, 1
a
Symbiosis Institute of Business Management, Symbiosis International University, India
b
College of Business, Tennessee Technological University, USA
a r t i c l e i n f o
Article history:
Received 23 September 2013
Accepted 3 March 2015
Keywords:
Airlines in India
DEA
Input ef?ciency pro?ling
Technical ef?ciency
Scale ef?ciency
a b s t r a c t
The objective of this study is to investigate the technical and scale ef?ciencies of all airlines across service
type, size and ownership structures operating in India during 2006e2010. The variable returns to scale
(VRS) model of Data Envelopment Analysis (DEA) with two inputs and two outputs is used. For additional
insights, Input Ef?ciency Pro?ling (IEP) model of DEA is also used. The key ?ndings are: a large majority
of budget airlines have been found to be ef?cient; while smaller private sector airlines have been ef?-
cient, both the larger and smaller public sector airlines have also been ef?cient; the public sector airlines,
although incurring ?nancial losses, are also operating at their most productive scale size; of the two
inputs, there is greater inef?ciency with respect to the operating cost input. These ?ndings are consistent
with other studies of airlines that found size, type of service and ownership to impact ef?ciency.
© 2015 College of Management, National Cheng Kung University. Production and hosting by Elsevier
Taiwan LLC. All rights reserved.
1. Introduction
The civil aviation industry in India has come a long way since the
Air Corporation Act was repealed in the year 1994 allowing private
airlines to operate in scheduled services category. Several private
operators showed interest and were granted the status of sched-
uled carriers in the year 1995. However, many of those private
airlines soon shut down. Only Jet Airways and Sahara Airlines
survived and continued to offer scheduled services (Table 1). Till
the years 2001e2003, the Indian civil aviation sector was charac-
terized by the domination of government owned national carriers
like Indian Airlines (domestic) and Air India (international),
excessive regulations (e.g., control of aviation fuel) and taxation. All
this resulted in high cost of operating airline services which acted
not only as an entry barrier for the private players but also made air
travel an expensive affair giving it an image of elitist mode of travel.
However, the scenario has changed rapidly over the last decade
and the sector has witnessed a signi?cant growth not only in terms
of the entry of new private players but also in terms of increase in
passenger traf?c (16 percent CAGR between 2001 and 2011 with a
more rapid growth at an average of 19 percent in the latter half of
the decade from 2006 to 2010, Fig. 1). All this can be attributed
arguably to the introduction of structural reforms, entry of new
private airlines especially with different cost structures e.g., low
cost e no frills, airport modernizations, and improvement in ser-
vice standards (ICRA Research Report; March 2012). Also, as the
Indian economy began to grow faster, international passenger
traf?c into and out of India also began to grow (Fig. 1).
In the year 2003, India's ?rst low-cost carrier (LCC), Air Deccan,
entered the market. This was a landmark event for this industry,
which triggered an entry of several other private players. The year
2005 was also a watershed year for the Indian civil aviation sector
as several private low cost and full services carriers (such as
King?sher, IndiGo, SpiceJet, GoAir, and Paramount) commenced
their operations (Table 1). This led to the rapid increase in capacity
as measured by Available Seat Kilometers (ASK). Some of the new
entrants e Air Deccan, Spice Jet, GoAir and IndiGo e were pursuing
a different strategy and competed as low cost carriers (LCC).
The entry of the LCCs has signi?cantlyexpandedthe civil aviation
market by making air travel both affordable and accessible to the
middle class. Low fares offered by LCCs have made air travel very
attractive, prompting travelers to switch to air travel fromroad and
rail travel. In the early days of Air Deccan 40 percent of its passen-
gers were ?rst-time ?yers. SpiceJet, for instance, targeted passen-
gers who were traveling by air-conditioned classes in Indian
Railways. LCCs ushered in a newera of competition among airlines.
For instance, LCCs competitive pricing set off a price war with the
* Corresponding author. Kothur Mandal, Mamidipalli village, Mahabubnagar
District, Hyderabad 509216, India.
E-mail address: [email protected] (R.K. Jain).
Peer review under responsibility of College of Management, National Cheng
Kung University.
1
Tel.: þ1 931 372 3001.
HOSTED BY
Contents lists available at ScienceDirect
Asia Paci?c Management Review
j ournal homepage: www. el sevi er. com/ l ocat e/ apmrvhttp://dx.doi.org/10.1016/j.apmrv.2015.03.004
1029-3132/© 2015 College of Management, National Cheng Kung University. Production and hosting by Elsevier Taiwan LLC. All rights reserved.
Asia Paci?c Management Review 20 (2015) 285e292
incumbent Full Service Carriers (FSCs) such as Indian Airlines, Jet
Airways, and Air Sahara. This compelled the FSCs to discount their
fares by as much as 60e70 percent in some routes to match
the prices of LCCs. While the aggressive pricing strategy of LCCs has
both deepened and widened demand which has percolated to non-
metrohttp://www.domain-b.com/Aerospace/recommend.aspx
towns and Tier-II cities, this has hurt not only the pro?tability of
LCCs but also the revenues, margins, and market shares of FSCs. For
instance, Jet Airways, which controlled about 50 percent of the
domestic market in 2003, saw its share (including that of its
acquisition Jet Lite) drop to about one-third by 2007.
According to Kaul (2007) of Centre for Asia Paci?c Aviation, “…
the aggressive expansion of the LCC segment comes at a cost to the
whole sector. India's airlines are expected to post a combined loss
of approximately USD$500 million in the current ?nancial year
ending 31-Mar-07,”
Aggressive expansion of capacity and inability to control costs
were other factors which contributed to the mounting losses. These
conditions can lead only to two types of outcomes for the air-
linesdeither some of them go bust in a market shake-out or they
merge/get acquired by other airlines or business groups. Whereas
in the 1990s, many private carriers went bust, this time around the
industry has witnessed a wave of consolidation. Year 2007 became
a landmark year in the industry when major consolidation took
place (Table 1).
In 2008, there was a steep fall in the domestic air travel due to
the slowdown in the Indian economy, the H1N1 ?u scare, and the
terrorist attack in Mumbai in November 2008 (Fig. 1). There was
excess capacity all around and the airlines responded by developing
plans to lay off employees and by offering deeply discounted fares
to stimulate demand. Rival airlines Jet Airways and King?sher
formed a strategic alliance for code sharing and cutting costs. The
trend was to shift more capacity to LCC operations. The Indian
economy slowed in 2008e09 but there was no recession and as the
economy picked up in the second half of 2009, the demand for air
travel made a comeback (Fig. 1). Interestingly, the “pure” LCCs like
SpiceJet and IndiGo made pro?ts while carriers like King?sher, Jet
Airways, Air India and Indian Airlines operated both FSC and LCC
services, experienced huge losses. The anticipated cost reduction
due to synergies from the mergers could not be realized. In 2012,
the slowing economy affected the overall demand for air travel and
the domestic demand was down by 4.9 percent compared to 2011.
The airlines were also adjusting their capacities downwards to
more realistic levels (IATA, 2013).
Despite these setbacks and massive losses, the long term pros-
pects for the industry appear to be quite good. The reason why
foreign airlines such as Air Asia (Asia's largest lowcost airline wants
to take 49 percent stake in a joint venture with the Tata Group) and
Abu Dhabi-based Etihad (which announced to buy a stake in Jet
Airways in early 2013) ?nd the Indian market attractive is the huge
potential for growth in air travel (Kazmin, 2013). India accounts for
only about 2 percent of global air traf?c. Only about 4 percent of
India's population of over 1.1 billion people had ever been on a
?ight. Industry estimates suggest that the total passenger traf?c
will growfrom143 million in 2010e11 to 290e300 million by 2020
making India the third largest civil aviation market in the world. To
meet this demand, a ?eet of over 1000 aircrafts will be required
with around 350e400 operational airports across the country (The
Association of Private Airport Operators, 2013). This makes Indian
civil aviation industry very attractive. Low cost and budget airlines
are better positioned to dominate this market which is highly price
sensitive.
The above backdrop raises the research question whether the
developments described have brought about any signi?cant change
Fig. 1. Domestic and international passenger traf?c growth.
Source: Centre for Asia Paci?c aviation (2010) preparing for long term growth of Indian
aviation, New Delhi. Arushi and Stefan Drews (2011).
Table 1
Evolution of the Indian civil aviation industry.
Year Major milestones
1953 Nine Airlines existed including Indian Airlines & Air India
1953 Nationalization of all private airlines through Air Corporations Act;
1986 Private players permitted to operate as air taxi operators
1994 Air Corporation Act repealed; Private players allowed to operate scheduled services
1995 Jet, Sahara, Modiluft, Damania, East West granted scheduled carrier status
1997 4 out of 6 operators shut down; Jet & Sahara continue
2001 Aviation Turbine Fuel (ATF) prices decontrolled
2003 Air Deccan starts operations as India's ?rst LCC
2005 King?sher, SpiceJet, IndiGo, GoAir, Paramount start operations
2007 Industry consolidates; Jet acquired Sahara; King?sher acquired Air Deccan
2010 SpiceJet starts international operations
2011 IndiGo starts international operations, King?sher exits LCC segment. Air India and Indian
airlines merger completed.
2012 Government allows direct ATF imports, FDI proposal for allowing foreign carriers
to pick up to 49 percent stake under consideration. King?sher goes bankrupt.
2014 Falling international crude prices lead to sharp cuts in the price of aviation turbine fuel (ATF)
or jet fuel. Debt-ridden SpiceJet cancels ?ights, cuts third of its ?eet and seeks help from the government.
Source: ICRA research.
R.K. Jain, R. Natarajan / Asia Paci?c Management Review 20 (2015) 285e292 286
in productive ef?ciency of the airlines. This study investigates the
technical and scale ef?ciencies of all airlines across cost, size and
ownership structures e including budget and full-service, private
and state owned, large and small e operating in the country of-
fering scheduled services on domestic and international routes. The
scale at which they operate is identi?ed. While there are several
airline ef?ciency studies conducted globally, none to our knowl-
edge have attempted to study the Indian airlines industry. The
present study, by addressing this gap, contributes to the existing
body of literature in this area.
The rest of the paper is organized as follows. In Section 2, the
existing literature is reviewed. The research methodology is dis-
cussed in Section 3. The sample and the data sets are presented in
Section 4. Results are presented and discussed in Section 5. Con-
clusions follow in Section 6.
2. Literature review
The existing literature in this area has covered major airlines in
Asia, Europe and the USA across different cost structures (full ser-
vice carriers, budget airlines and low cost carriers); size (large and
small), and service type (domestic and international). Some of the
important studies are cited below.
Schefczyk (1993), used DEA technique, apparently for the ?rst
time, to analyze and compare operational performance of 15 in-
ternational airline using non-?nancial data such as Available Ton
Kilometer, Revenue Passenger Kilometer etc. The study demon-
strated that DEA can be a very useful tool to assess the technical
ef?ciency of international airlines which otherwise was dif?cult to
do using ?nancial data because: (1) most airlines lease a substantial
fraction of their ?eet; and (2) different accounting and taxation
rules in various countries result in different impacts of leased assets
on pro?t and balance-sheet information of airlines.
Good, Roller, and Sickles (1995) examined the performance of
the eight largest European and the eight largest American airlines
for a ten year period between 1976 and 1986 using two methods e
one a parametric using statistical estimation and another non-
parametric (DEA) using linear programming. The authors
observed discrepancy in the productive ef?ciency of European
airlines even under the conditions of deregulations and liberaliza-
tion of the airline industry. The study suggested that if European
airlines were as productively ef?cient as their American counter-
parts then they could save approximately $4 billion per year (in
1986 dollars).
Fethi, Jackson and Jones (2001) studied ef?ciency across a panel
of 17 European airlines in the 1990s during the early phase of
liberalization using stochastic DEA constructs (which develops
production frontiers that incorporate both inef?ciency and sto-
chastic error). They use the Land, Lovell, and Thore (1993) model
incorporating information on the covariance structure of inputs
and outputs. Stochastic DEA was chosen to achieve a closer envel-
opment of the mean performance of the companies in the sample
and reduce, if not eliminate, the effect of extreme outliers. This
model computes the relative ef?ciencies after allowing for sto-
chastic error. They concluded that the airlines that were ef?cient in
1995 resembled those that were ef?cient in 1993 but not those in
1991. Interestingly, they also found that larger size airline com-
panies were ef?cient thus implying that size of an airline does
contribute to their degree of ef?ciency.
Scheraga (2004) investigated the structural drivers of opera-
tional ef?ciency as well as the ?nancial position of airlines in a
industry ridden by ?nancial crises post-September 11, 2001. The
study analyzed 38 airlines from North America, Europe, Asia and
the Middle East and across different cost structures (low cost and
budget airlines included for the ?rst time in such a study) to
investigate whether relative operational ef?ciency implied superior
?nancial performance. The author used DEA technique to derive
ef?ciency scores for individual airlines followed by Tobit analysis to
investigate the underlying structural drivers of ef?ciency. The study
concluded that the traditional framework developed in the litera-
ture still provided reasonable explanatory power for realized
relative operational ef?ciency. The author also observed that rela-
tive operational ef?ciency did not inherently imply superior
?nancial performance and airlines that had chosen relatively ef?-
cient operational strategies found themselves vulnerable in terms
of ?nancial performance and thus suffered the consequences in the
post-September 11 environment.
Zhu (2011) measured performance ef?ciency of 21 airlines in the
US during 2007e2008 using the centralized ef?ciency model, a
two-stage process used to optimize performance simultaneously,
instead of a standard DEA technique thus generating ef?ciency
decomposition for the two individual stages. For instance, in stage
one, resources (fuel, salaries, and other factors) are used to main-
tain the ?eet size and load factor, and in stage two, the revenue
generating capacity is assessed given the ?eet size and load factors.
The study considers airlines with different cost structures unlike
most of the earlier studies that analyzed only major full service
airlines.
Lee and Worthington (2010) is one of the very few studies in
that included airlines across different cost structures (full services,
low cost and budget airlines), and service types (international and
domestic services). The study investigates whether the inclusion of
low-cost airlines in the data set of international and domestic air-
lines has any impact on the ef?ciency scores of those otherwise
‘ef?cient’ full service airlines. The authors analyzed 53 airlines in
the year 2006 using the non-parametric technique of DEA to
investigate their technical ef?ciency. The study reveals that the
majority of budget airlines are ef?cient relative to the more pres-
tigious full service airlines. Moreover, most airlines identi?ed as
inef?cient are so primarily because of the overutilization of non-
?ight assets.
Michaelides, Belegri-Roboli, Karlaftis, and Marinos (2009) have
employed both Stochastic Frontier Analysis (SFA) and DEA using a
panel data set of 24 world's largest network airlines to estimate
technical ef?ciency in international air transport for the period
1991e2000. The authors observed that the airlines achieved con-
stant returns to scale with technical ef?ciencies ranging from 51%
to 97%. They also observed that ownership (private or public) does
not affect the technical ef?ciency of the airlines. Further, it is
interesting to note that results from both SFA and DEA did not vary
signi?cantly.
Tofallis (1997) used a modi?ed version of DEA called input ef-
?ciency pro?ling method to assess the technical ef?ciency of 14
major full service airlines across all ?ve continents using the same
data set used by Schefczyk (1993). The study demonstrated that
input ef?ciency pro?ling gives better results when compared to
standard DEA in measuring technical ef?ciency of the airlines.
Among the above cited studies and several others (not cited in
detail here) such as Banker and Johnston (1994), Charnes, Galleous,
and Li (1996), Gillen and Lall (1997), Alam, Semenick, and Sickles
(1998), Adler and Golany (2001), and Coli, Nissi and Rapposelli
(2011), none have included airlines operating in an important
market like India. While such omission is perhaps justi?able for
academic studies till 2003 as the Indian civil aviation sector was
very limited with only a few players operating and primarily
dominated by government owned airlines, the more recent studies
also seem to have ignored the Indian civil aviation sector. The
present study contributes to the body of literature in this area by
analyzing all the airlines operating in domestic and international
services in full services, budget/low cost segments.
R.K. Jain, R. Natarajan / Asia Paci?c Management Review 20 (2015) 285e292 287
3. Research methodology
While there are many ways to de?ne and measure the pro-
ductive ef?ciency of an organizational/industrial unit, also called
Decision Making Unit (DMU), there are two widely employed
techniques namely Data Envelopment Analysis (DEA) and Sto-
chastic Frontier Analysis (SFA) (Farrell, 1957). These methods pro-
vide a measure of the technical ef?ciency of a DMU in terms of
radial distance from the best unit on the production frontier rep-
resented by the production function of the ef?cient units. Lesser the
distance greater is the ef?ciency.
DEA is a non-parametric method as it does not impose an
explicit functional form for the production frontier unlike an
econometric method like SFA. DEA method assumes that all de-
viations from the ef?cient frontier are due to inef?ciencies,
whereas the SFA technique assumes that deviations from the ef?-
cient frontier can be a random error or due to realization of in-
ef?ciency or a random shock. In short, DEA measures ef?ciency
relative to a deterministic frontier using linear programming
techniques to envelop observed input/output vectors as tightly as
possible rather than explicitly stating the formof the frontier based
on the de?ned relationship between inputs and outputs.
As seen in the literature on measurement of airline ef?ciency,
DEA method is widely employed for assessing the productive ef?-
ciencies of airlines across the world. This study uses the Banker,
Charnes, and Cooper (BCC) model (Banker, Cooper, Seiford, & Zhu,
2004) which allows for variable returns to scale (VRS) and the
input ef?ciency pro?ling (IEP) model of DEA (Tofallis, 1997).
Globally, airlines industry is characterized by high startup costs
and scale economies. Of course this does not imply that each airline
in the industry will be using a production technology that exhibits
economies of scale due to increasing returns to scale. The tech-
nology for converting inputs to outputs can exhibit different
returns to scale based on the levels of outputs. Some airlines could
be on a production possibility frontier where locally it has dimin-
ishing or increasing returns to scale. Therefore, the scale at which a
particular airline is operating can only be determined empirically.
In the context of DEA, this implies using the Banker, Charnes, and
Cooper (BCC) model (see Banker et al., 2004) which allows for
variable returns to scale (VRS) is more appropriate.
Allowing for variable returns to scale in constructing the ef?-
cient frontier enables separating the ef?ciency of scale from pure
technical ef?ciency. The ef?cient frontiers for the constant returns
to scale (CRS) and the variable returns to scale (VRS) speci?cations
are illustrated in Fig. 3 of the Appendix. When the production
technology allows variable returns to scale (VRS) at different points
on the frontier of the production possibility set, the technical ef?-
ciency (either input- or output-oriented) of a DMU will differ from
its scale ef?ciency. Technical ef?ciency is measured by comparing
the average productivity of a DMU with the corresponding average
productivity at its input- or output-oriented projection onto the
VRS frontier. Scale ef?ciency, on the other hand, compares the
average productivity at the ef?cient input- or output-oriented
projection with the maximum average productivity attained at
the Most Productive Scale Size (MPSS) on the VRS frontier. Fig. 4 in
the Appendix illustrates these concepts. For the de?nition and
discussion of the concept of MPSS see Banker et al. (2004).
The BCC model considers all the inputs together whereas the
effect of each input is considered separately in the IEP model. The
DEA models in general and BCC model in particular has a linear
weighted combination of all the inputs in the denominator of the
objective function. This implies that all the inputs are perfectly
substitutable with the weights as the marginal rates of substitution.
The mathematical formulations (linear programming models) of
these models are given in the Appendix.
The study considers the following two inputs and two outputs,
consistent with the ones used in the earlier academic studies, (Lee
& Worthington, 2010; Schefczyk, 1993; Scheraga, 2004) but ex-
cludes the third input, ‘non-?ight assets’ (NFA), in airlines opera-
tions. Lee and Worthington (2010) in their recent study have
observed that excluding NFA from input set does not seem to have
any signi?cant affect on the overall ef?ciency of the airlines.
Input 1: X1 ¼ Total Available Ton Kilometer (ATKM) (re?ects
airline ?eet capacity in tonnage)
Input 2: X2 ¼ Operating Cost (OC) (all operating cost excluding
aircraft rentals, depreciation and amortization and other capital
expenditure)
Output 1: Y1 ¼ Revenue Passenger Kilometer Performed
(RPKM) (Kilometers performed of revenue paying passenger
traf?c)
Output 2: Y2 ¼ Non-Passenger Revenue (NPR) (revenue from
non-passenger traf?c such as carrying cargo)
The rationale for using the input ef?ciency pro?ling (IEP) model
is as follows. The input pro?ling model is more discriminating than
the conventional DEA model when the inputs are not substitutes.
This applies in the case of airlines where the two inputs i.e., avail-
able ton kilometers (a capacity measure) and operating costs
(which include fuel and labor costs) are not substitutes. Before
Tofallis (1997), Kopp (1981) and Kumbhakar (1988), in their pro-
ductivity studies -which do not use DEA e have considered effects
of each input separately. As Kumbhakar (1988) puts it “knowing the
magnitude of (overall) technical ef?ciency is not enough. It is
important to knowwhich inputs are causing the inef?ciency and to
what extent”. Input pro?ling helps in isolating the inef?ciencies
with respect to each input. This information is useful for managers
to set priorities and input-speci?c targets for improvements.
Input ef?ciency pro?ling model of DEA is particularly relevant to
the situation in the airline industry in India for the following
reasons.
1) The airlines in India want to expand their ?eet by signi?cant
margins (see Table 2 for aircraft on order) (Arushi & Drews,
2011). For instance, IndiGo's aircraft orders in 2011 is more
than six times its ?eet size! This raises questions about utili-
zation of the ATKM input.
2) Fuel costs make up around 40%e50% of total operating expenses
of airline industry in India (Arushi & Drews, 2011). In compari-
son, this cost factor accounted for 33% on average for the global
airline industry in 2008 and 2012 and 26% in 2009 and 2010,
respectively, according to the International Air Transport Asso-
ciation (IATA, 2012).
3) Cost reduction is critical for all airlines in India. Aviation Turbine
Fuel (ATF) costs 60% more in India because of high sales taxes of
state and central governments (Arushi & Drews, 2011). Most
airlines in India have not been able to control costs. The term
LCC is somewhat misleading in India. Among the LCCs in India,
as shown in Fig. 2, in 2007, SpiceJet had the lowest unit cost at
6.2 U.S. cents per Available Seat Kilometer (ASK), which is
comparable with Southwest, Easy Jet, and Jet Blue (CAPA, 2007).
But this is more than twice that of the best performers like Air
Asia with unit cost of slightly over 3 cents per ASK. Another
Indian carrier Jet Lite (previously Air Sahara before it was ac-
quired by Jet Airways in 2007) shows up in the Figure at 7.2
cents per ASK. According to Bill Franke (2007), the then Man-
aging Director of leading airline investment ?rm Indigo Part-
ners, “There is not a single airline in India that operates a true
low cost structure. Under the current conditions, it's not
possible.” Similar views were expressed by Naresh Goyal, the
R.K. Jain, R. Natarajan / Asia Paci?c Management Review 20 (2015) 285e292 288
founder and chairman of Jet Airways who questioned the
feasibility of LCCs in India because of lack of infrastructure and
skills. “Here there are no alternative airports.” “India has
nothing called low-cost, only low-fare and low-margin. This is
irrational pricing which will make the whole industry sick,”
(Economist, 2007). Air India, which has accumulated huge los-
ses, is trying to turn the situation around by cutting wages and
using fuel ef?cient aircraft (Boeing's 787 Dreamliner) in some of
its routes, (Sanjai, 2012).
Therefore, in the Indian context, it is worth analyzing the in-
ef?ciencies of the ATKM and OC inputs separately.
4. Sample and datasets
The sample includes all commercial scheduled airline carriers,
with different ownerships (public and private) and cost structure
(full service and lowcost), offering scheduled services on domestic
and international routes operating during the ?ve year period be-
tween 2006 and 2010. A total of twelve airlines were considered for
analysis. Focusing on this period will allow the analysis of the ef-
?ciencies of the new private operators who have entered the in-
dustry since 2005.
Most of the data is drawn fromvarious reports and air transport
statistics (ATS) provided by the Directorate General of Civil Aviation
(DGCA). Further, annual reports and balance sheets of these airlines
were also referred to supplement the data with accounting
Fig. 2. Low Cost Carrier (LCC) unit costs per ASK (including fuel).
Source: Center for Asia Paci?c Aviation (CAPA), 2007.
Fig. 3. Ef?cient frontiers for constant and variable returns to scale DEA speci?cations. Fig. 4. Technical and scale ef?ciencies and returns to scale.
R.K. Jain, R. Natarajan / Asia Paci?c Management Review 20 (2015) 285e292 289
numbers wherever required. Table 3 shows the input and output
data (averaged for the period of study) used to derive ef?ciency
scores.
5. Results
Table 4 provides, in percentages, the technical ef?ciency scores
for the VRS model (denoted by VRS TE), the scale ef?ciencies (SE),
the input ef?ciency pro?ling (IEP) (denoted by IEP (ATKM), IEP
(OC)) respectively for 2006e10. It also identi?es the returns to scale
at which the airline is operating.
Results from Table 4 suggest the following interpretations.
1. Of the six new entrants after 2005 three of them GoAir, In-
diGo, and Air India Express are found to be ef?cient with VRS
technical ef?ciency scores of 100 percent. Two out of the
remaining three new entrants which were inef?cient were
Full Service Carriers (FSC).
2. Budget airlines such as Air Deccan, GoAir, IndiGo, and Air
India Express are ef?cient with 100 percent scores (Table 4).
This indicates that in the Indian context also the cost struc-
ture of an airline in?uences its degree of productive ef?-
ciency, which is consistent with the observations made by
Lee and Worthington (2010).
3. Relatively larger FSCs, Air India, Indian Airlines, and Jet Air-
ways are ef?cient. These results are consistent with the
observation made by Fethi et al. (2001) that size of an airline
does contribute to the degree of ef?ciency.
4. While smaller private sector airlines have been ef?cient, all
the public sector airlines (the larger Air India and Indian
Airlines and the smaller Air India Express) have also been
ef?cient. These results support the conclusions of
Michaelides et al. (2009) that ownership does not affect the
technical ef?ciency of airlines.
5. When the results of the input ef?ciency pro?ling model are
considered, the public sector airlines Air India Express, Air
India and Indian Airlines have performed better than the
private sector airlines with respect to the Operating Cost (OC)
input.
6. Except for Air India and Air India Express, there is greater
inef?ciency with respect to the Operating Cost input, when
the ef?ciencies of the two inputs are compared. The ef?-
ciency for this input was as lowas 25.7 percent (Paramount).
The average ef?ciency with respect to ATKM input is 86.6
while the average ef?ciency with respect to the OC input is
67.6 (Table 4). One possible interpretation for such results is
that while the price war triggered by LCCs such as Air Deccan,
increased the demand for air travel and therefore resulted in
better capacity utilization for most of the airlines, whereas
the increase in fuel cost (which accounts for almost half of
the operating expenses for airline in India), which is gener-
ally beyond the control of the airlines, took its toll on their
operating cost.
7. The IEP scores can help identify areas for improvements in a
way that the results of the standard DEA model cannot. For
instance, for Jet Lite, the IEP model score for input 1 is 88.50
percent, which implies that ATKM can be reduced to 88.50
percent of the current level of that input to achieve current
level of the output. However, according to Table 4, there is
greater inef?ciency in the use of second input, i.e., Operating
Cost (OC), for Jet Lite can reduce that input to 48.9 percent of
its current level and still maintain the current level of output.
The IEP scores suggest that the airlines have to focus on
improving ef?ciency of the Operating Cost input.
8. The returns to scale classi?cation given in Table 4 is based on
the results given in Banker et al. (2004). The DEAP software
(2013) used for this study provides the classi?cation as one
of the outputs of DEAwith VRS speci?cation (the BCC model).
The SE scores in Table 4 indicate that of the new airlines,
GoAir, IndiGo and Air India Express have been operating at
their Most Productive Scale Size (MPSS) in 2006e2010. Air
India, Indian Airlines and Air Deccan have also been able to
operate at MPSS. The means of the SE scores in Table 4
indicate that the average percentage of deviation from
MPSS is 3.8 percent (100e96.2).
9. Jet Airways is classi?ed as operating in diminishing returns
to scale (DRS) during 2006e2010. Usually DRS signify that
either an input such as non-?ight assets (NFA) is not
accounted for or some inputs e.g., pilots, are limited. DRS can
be a short term phenomenon e in the longer term such
limitations can be removed. In airlines, the key input of ?eet
capacity is characterized by indivisibilities. Each aircraft has a
Table 2
Market share, ?eet size, aircrafts on order of scheduled domestic airlines in 2011.
Airline Market share (%) Fleet Orders
King?sher 19 66 130
IndiGo 18.70 39 241
Jet Airways 18 96 29
Indian Airlines 15.80 119 30
SpiceJet 13.80 26 66
Jet Lite 8.1 19 8
GoAir 6.6 10 10
Total 100 375 514
Table 3
Average input e output data for period 2006e2010.
DMU ATKM OC RPKM NPR
Jet Airways 3112.50 81,910.30 16,583.44 12,647.38
Jet Lite 514.94 20,470.22 3827.79 1150.86
Paramount Airways 27.30 2090.13 205.82 12.68
KF Airlines 751.72 38,295.64 5569.27 1657.74
Alliance Air 74.00 4465.15 482.68 942.88
SpiceJet 529.84 14,155.04 3938.75 1002.22
Air India Express 560.80 9854.94 3770.60 1730.54
Air Deccan 578.63 22,899.20 4859.66 2643.33
Air India 4214.60 13,6102.42 18,304.29 40,135.02
Indian Airlines 1693.75 118,040.35 11,569.95 29,004.75
GoAir 151.38 5549.65 1260.96 365.65
IndiGo Airlines 585.35 14,156.03 4311.60 992.18
Table 4
Results for period 2006e2010.
Airlines VRS TE SE IEP (ATKM) IEP (OC)
Jet Airways 100.0 78.0 DRS 65.10 70.0
Jet Lite
a
88.6 99.9 IRS 88.50 48.9
Paramount Airways 100.0 89.8 IRS 89.80 25.7
KF Airlines 92.7 95.2 IRS 88.20 38.0
Alliance Air 100.0 92.0 IRS 90.50 73.2
Spice jet 98.6 99.8 IRS 88.50 72.7
Air India Express 100.0 100.0 MPSS 80.10 100.0
Air Deccan 100.0 100.0 MPSS 100.00 60.5
Air India 1.000 100.0 MPSS 61.70 100.0
Indian Airlines 100.0 100.0 MPSS CRS 100.00 83.3
Go Air 100.0 100.0 MPSS 99.20 59.4
IndiGo airlines 100.0 100.0 MPSS MPCRS 87.70 79.6
Mean 98.3 96.2 86.60 67.60
DRS e Decreasing Returns to Scale; IRS e Increasing Returns to Scale; MPSS e Most
Productive Scale Size; TE e Technical Ef?ciency; SE e Scale Ef?ciency.
a
Jet Lite was called Sahara before the latter was taken over by Jet Airways in the
year 2007.
R.K. Jain, R. Natarajan / Asia Paci?c Management Review 20 (2015) 285e292 290
minimum capacity and each route the airlines decided to
operate also has a minimum distance. Therefore, ?eet ca-
pacity in terms of ATKM will increase in lumps. This could
afford opportunities for increasing returns to scale (IRS) if the
increase in ATKM is more than matched by increase in RPKM
on the demand side. Economies of scale opportunities are
also available with respect to the other input, i.e., Operating
Costs (OC). For instance, inventory levels and costs increase
less than proportionally for any increase in demand, and
increase in costs of construction or leasing of warehouses is
also less than proportional to the increase in the volume of
space required. Such opportunities for IRS are more likely in
the initial stages of expansion of airlines. This has been
observed in the case of Paramount Airlines, a new airline in
the sample, which operated in the southern region offering
only business class service.
10. The scale at which an airline operates also depends on its
expansion path relative to increase in demand for its ser-
vices. Many airlines in India have ambitious expansion ?eet
plans (see Table 2). But ?eet capacity expansion has to be
commensurate with expansion of demand. Both Air Deccan
and King?sher from their inception have expanded their
operations very aggressively aiming to become national
carriers. In the case of Air Deccan, the ?rst discount airline in
India, the demand increased because of its low fares but in
the case of King?sher (KF) which operated as a Full Service
Carrier (FSC) charging higher fares, it did not. Due to its
reckless expansion and its inability to control costs, King-
?sher went bankrupt in 2012.
11. Maintaining a dynamic balance between capacity and de-
mand is relatively easier for mature airlines such as Air India
and Indian Airlines because of their sheer size. The ?eet ca-
pacity increments are smaller in percentage terms relative to
overall existing capacity and same would be true of any
?uctuations in demand. These airlines are expected to
exhibit CRS and this is validated in Table 4 by the fact that
they are operating at their MPSS.
6. Conclusions
The following broad conclusions can be drawn from the study.
1. Budget airlines and low-cost airlines are found to be more
ef?cient when compared to the full service airlines in the private
sector. This supports the conclusion drawn by Lee and
Worthington (2010) that cost structure of airlines matters in
their ability to achieve better productive ef?ciencies.
2. While the new, private, low-cost airlines have been ef?cient, all
the public sector airlines have also been ef?cient. This is inter-
esting because ?nancially all the public sector airlines have only
made losses during the period of the study. This conclusion
supports Scheraga (2004) that relative technical ef?ciency does
not imply superior ?nancial performance.
3. It is also found that size and scale of the airline in?uences their
ef?ciency scores, thus supporting the conclusions drawn by
Fethi et al. (2001). Larger airlines (which in India happen to be
publicly owned) have been found to be ef?cient in the VRS
model. In the Input Ef?ciency Pro?ling models, they were found
to be more ef?cient when compared to the smaller airlines.
4. Input Ef?ciency Pro?ling model results indicate that, of the two
inputs considered in the study, almost all the airlines need to
achieve greater ef?ciency in the use of operating costs (of which
fuel costs are a signi?cant component). Such inef?ciency with
respect to operating cost can be arguably attributed to the cost
of highly regulated Aviation Turbine Fuel.
5. Further studies e along the lines of Good et al. (1995) in the
context of European airlines e need to be conducted to assess if
liberalization of the airline sector and the consequent entry of
new airlines in India has contributed to the improvement of
productive ef?ciencies of the airlines.
Appendix
1. The DEA formulation, according to Banker, Charnes, and Cooper
(1984) of input minimization with an assumption of variable
returns to scale (VRS) to calculate the ef?ciency scores is given
below.
The relative ef?ciency of a Decision Making Unit DMUo is ob-
tained from the following linear programming (LP) model:
Minq
0
subject to
q
0
x
ij0
À
X
n
j¼1
l
j
x
ij
0; i ¼ 1; …; m
X
n
j¼1
l
j
y
rj
y
rj0
; r ¼ 1; …; s
X
n
j¼1
l
j
¼ 1; j ¼ 1; …; nðconvexity constraintÞ
l
j
0; j ¼ 1; …; n
where, y
rj
is the amount of the r
th
output of DMU j, x
ij
is the amount
of the i
th
input to DMU j, l
j
are the weights of DMU j and q is the
shrinkage factor.
The model seeks a set of non-negative l values which add up to
1 and which minimizes q
0
toq
*
0
and identi?es a point within the
production possibility set which uses the lowest proportion q
*
0
of
input levels of DMU j while offering output levels which are at least
as high as those of DMU j. This point is a composite DMU corre-
sponding to the linear combination of ef?cient DMUs:
X
n
j¼1
l
*
j
x
ij
;
X
n
j¼1
l
*
j
y
rj
; with i ¼ 1; :::; m and r ¼ 1; :::; s
It can be said that:
X
n
j¼1
l
*
j
x
ij
;
X
n
j¼1
l
*
j
y
rj
outperformsðq
0
; X
j0
; y
jo
Þ when q
)
0
TE
NIRS
while DMU Q
is too big and operating at DRS because TE
VRS
¼ TE
NIRS
, and DMU C
is at optimal size (MPSS) because TE
VRS
¼ TE
CRS
DMU C is on both the CRS and VRS ef?cient frontiers. It is
operating at a level of output where IRS changes to DRS and has the
maximum possible economy of scale. It has a scale ef?ciency of 100
percent and is operating at its Most Productive Scale Size (MPSS).
References
Adler, N., & Golany, B. (2001). Evaluation of deregulated airline networks using data
envelopment analysis combined with principal component analysis with an
application to Western Europe. European Journal of Operational Research, 132(2),
260e273.
Alam, I., Semenick, M., & Sickles, R. C. (1998). The relationship between stock
market returns and technical ef?ciency innovations: evidence from the US
Airline Industry. The Journal of Productivity Analysis, 9(1), 35e51.
Arushi, & Drews, S. (2011). Aviation and environment (Working paper). New Delhi,
India: Center for Science and Environment.
Banker, R., Charnes, A., & Cooper, W. W. (1984). Some models for estimating tech-
nical and scale inef?ciencies in data envelopment analysis. Management Sci-
ence, 30, 1078e1092.
Banker, R., Cooper, W. W., Seiford, L. M., & Zhu, J. (2004). Returns to scale in DEA. In
W. W. Cooper, L. M. Seiford, & J. Zhu (Eds.), Handbook of data envelopment
analysis (pp. 41e70). Boston, MA: Kluwer Academic Publishers.
Banker, R., & Johnston, H. (1994). Evaluating the impacts of operating strategies on
ef?ciency in the US airline industry. In A. Charnes, W. W. Cooper, A. Lewin, &
L. Seiford (Eds.), Data envelopment analysis: Theory, methodology, and application
(pp. 97e128). Boston, MA: Kluwer Academic Publishers.
Centre for Asia Paci?c Aviation (CAPA) Report (2007).http://indiaaviation.aero/
news/airline/7591/59/Unit-costs-fallinge-airlines-in-drive-for-ef?ciency-.
Accessed 27.03.13.
Charnes, A., Galleous, A., & Li, H. (1996). Robustly ef?cient parameter frontiers: an
approximation via the multiplicative DEA model for domestic and international
operations of the Latin American airline industry. European Journal of Opera-
tional Research, 88, 525e536.
Coli, M., Nissi, E., & Rapposelli, A. (2011). Ef?ciency evaluation in an airline com-
pany: some empirical results. Journal of Applied Sciences, 11, 737e742.
DEAP. (2013). DEAP V2.1 a data envelopment analysis (Computer) program. http://
www.uq.edu.au/economics/cepa/deap.php. Accessed 27.03.13.
Farrell, M. J. (1957). The measurement of productive ef?ciency. Journal of Royal
Statistical Society, Series A (General), 120, 253e290.
Fethi, M. D., Jackson, P. M., & Weyman-Jones, T. G. (September 2001). European
airlines: a stochastic DEA study of ef?ciency with market liberalization. Paper
presented at the Seventh European Workshop on Ef?ciency and Productivity
Analysis (7EWEPA), University of Oviedo, Oviedo, Spain.
Franke, B. (2007). No true LCCs in India e indigo partners. In Keynote address at the
4th annual India and Middle East Low Cost Airline Symposium. http://
centreforaviation.com/members/download/avan/20070614.htm. Accessed
27.03.13.
Gillen, D., & Lall, A. (1997). Developing measures of air transport productivity and
performance: an application of data envelopment analysis. Transportation
Research, 33, 261e273.
Good, D. H., Roller, L.-H., & Sickles, R. C. (1995). Airline ef?ciency differences be-
tween Europe and the US: implications for the pace of EC integration and do-
mestic regulation. European Journal of Operations Research, 80, 510e518.
ICRA Research Report. (2012). Indian aviation industry: Through turbulent times, FDI
relaxation alone not a game changer.http://www.icra.in/Files/ticker/Indian%
20Aviation%20Industry%20%28NEW%29.pdf. Accessed 27.03.13.
International Air Transport Association (IATA) (2012). Fact sheet: Fuel.http://www.
iata.org/pressroom/facts_?gures/fact_sheets/Pages/fuel.aspx. Accessed 27.03.13.
International Air Transport Association (IATA) (2013). Air passenger market analysis.http://www.iata.org/whatwedo/Documents/economics/Passenger-Analysis-
Jan-2013.pdf. Accessed 27.03.13.
Kaul, K. (2007). Aviation consolidation looming in India. Centre for Asia Paci?c
Aviation (CAPA).http://www.centreforaviation.com/news/press-releases/2007/
03/07/aviation-consolidation-looming-in-india-outlook-2007-report-released/
page1. Accessed 27.03.13.
Kazmin, A. (February 21, 2013). Foreign airlines see beyond clouds in India. Financial
Times.http://www.ft.com/cms/s/0/38d491ee-7be5-11e2-99f0-00144feabdc0.
html#axzz2f9Va9mue. Accessed 27.03.13.
Kopp, R. J. (1981). The measurement of productive ef?ciency: a reconsideration.
Quarterly Journal of Economics, 96(3), 477e504.
Kumbhakar, S. C. (1988). Estimation of input-speci?c technical and allocative in-
ef?ciency in stochastic frontier models. Oxford Economic Papers, 40, 535e549.
Land, K. C., Lovell, C. A. K., & Thore, S. (1993). “Chance-constrained Data Envelop-
ment Analysis”. Managerial and Decision Economics, 14, 541e554.
Lee, B. L., & Worthington, A. C. (2010). The relative ef?ciency of international, do-
mestic, and budget airlines: Nonparametric evidence (Working Paper No. 2010-
02). Australia: Grif?th University.
Michaelides, P. G., Belegri-Roboli, A., Karlaftis, M., & Marinos, T. (2009). Interna-
tional air transportation carriers: evidence from SFA and DEA technical ef?-
ciency results (1991e2000). European Journal of Transportation and
Infrastructure Research (EJTIR), 4(9), 347e362.
Sanjai, P. R. (2012). Air India to cut wages to bring down costs. Live mint.http://www.
livemint.com/Companies/xraOS5yWgAnS9P5Y2n0yUJ/Air-India-to-cut-wages-
to-bring-down-costs.html. Accessed 27.03.13.
Schefczyk, M. (1993). Operational performance of airlines: an extension of tradi-
tional measurement paradigms. Strategic Management Journal, 14(4), 301e317.
Scheraga, C. A. (2004). Operational ef?ciency versus ?nancial mobility in the global
airline industry: a data envelopment and Tobit analysis. Transportation Research,
Part A, 38, 383e404.
Taking Flight. (2007, September 6). Face value. The Economist.
The Association of Private Airport Operators. (2013). A paper on civil aviation policy.http://civilaviation.gov.in/cs/groups/public/documents/newsandupdates/moca_
001404.pdf. Accessed 05.09.13.
Tofallis, C. (1997). Input ef?ciency pro?ling: an application to airlines. Computers
and Operations Research, 24(3), 253e258.
Zhu, J. (2011). Airlines performance via two-stage network DEA approach. Journal of
CENTRUM Cathedra, 4(2), 260e269.
R.K. Jain, R. Natarajan / Asia Paci?c Management Review 20 (2015) 285e292 292
doc_968355221.pdf