Business Relations Between The Low Cost

The Journal of Air Transport Studies (JATS – ISSN: 1791-6771) is a peer reviewed journal
aiming at publishing high quality research related to air transport. JATS is interested in
publishing papers primarily focusing on economics, geography, policymaking, management,
marketing, operations, technology, logistics/ supply chain management and modelling.
The Journal is published electronically twice a year, i.e. in January and July by the Hellenic
Aviation Society (www.aviationsociety.gr). The January issue usually contains papers
(subject to changes) originally presented at the Air Transport Research Society
(www.atrsworld.org) Conference of the previous year(s) whereas the July issue may be
occasionally dedicated to a special theme. The Journal is accessible online free-of-charge.
Journal of Air Transport Studies (JATS)

© Hellenic Aviation Society
12, Agiou Charalambous Street, Athens 114 74, Greece.
Telephone: + 30 210 64 24 401
Facsimile: + 30 210 64 24 401
Website: http:// www.aviationsociety.gr
Volume 3, Number 1, January 2012, ISSN: 1791-6771.

Editorial and Advisory Board
‚ Editor-in-Chief: Dr Andreas Papatheodorou, Hellenic Aviation Society and University of
the Aegean, Greece
‚ Associate Editor: Dr Kostas Iatrou, Hellenic Aviation Society, Greece
‚ Assistant Editor: Dr Zheng Lei, Cranfield University, United Kingdom
‚ Book Reviews Editor: Mr Pavlos Arvanitis, University of the Aegean, Greece
‚ Conference Reports Editor: Ms I oulia Poulaki, University of the Aegean, Greece
‚ Honorary Advisor: Dr Taieb Cherif, Former ICAO Secretary General
‚ Scientific and Advisory Board
o Professor Larry Dwyer, University of New South Wales, Australia
o Professor Peter Forsyth, Monash University, Australia
o Professor Sveinn Gudmundsson, Toulouse Business School, France
o Professor Hans-Martin Neimeier, University of Applied Sciences Bremen, Germany
o Professor Tae Oum, President of ATRS, University of British Columbia, Canada
o Professor Paris Tsartas, University of the Aegean, Greece
o Professor Respicio Espírito Santo, Rio de Janeiro Federal University, Brazil
o Professor Pauline Sheldon, University of Hawaii, USA
o Professor Anming Zhang, University of British Columbia, Canada
o Professor Evangelos Christou, Technological Institute of Thessaloniki, Greece
o Dr Anderson Correia, I nstituto Tecnológico de Aeronáutica, Brazil
o Dr Dimitrios Dimitriou, Cranfield University, United Kingdom
o Dr Rafael Echevarne, ACI , Canada
o Dr Triant Flouris, Hellenic American University, Greece
o Dr Anne Graham, University of Westminster, United Kingdom
o Dr Paul Hooper, Department of Transport, Abu Dhabi, UAE
o Dr Panagiotis Karamanos, Athens International Airport, Greece
o Dr Eleftherios Katarelos, Hellenic Civil Aviation Authority, Greece
o Dr Tay Ryang Koo, University of New South Wales, Australia
o Dr Konstantinos Kostopoulos, Hellenic Competition Commission, Greece
o Dr Gui Lohmann, Southern Cross University, Australia
o Dr Christos Markou, I ATA, Canada
o Dr Keith Mason, Cranfield University, United Kingdom
o Dr John F. O’ Connell, Cranfield University, United Kingdom
o Dr Marianna Sigala, University of the Aegean, Greece
o Dr Bijan Vasigh, Embry-Riddle Aeronautical University, USA
o Captain Spyros Jancovich, Hellenic Aviation Society, Greece
o Mr Takis Adamidis, Hellenic Aviation Society, Greece
o Mr Mario Diaz, Atlanta I nternational Airport, USA
o Mr Panagiotis Grigoriou, Hellenic Aviation Society, Greece
o Mr Emmanuel Gyzis, Emm. A. Gyzis & Partners Law Offices, Greece
o Mr Ray Kaduck, Canadian Transportation Agency, Canada
o Mr Emmanuel Keramianakis, Hellenic Aviation Society
o Mr Antonios Simigdalas, Independent Aviation Consulting, Cyprus
o Ms Narjess Teyssier, I CAO, Montreal
o Mr Stamatis Varsamos, Athens I nternational Airport, Greece
o Mr David Young, Eurocontrol, France
ii
Notes for Contributors
JATS publishes the following categories of papers written in scholarly English: a) Full
Research Papers, b) Conference Reports, c) Book Reviews, d) I ndustry Perspectives.
Papers should be submitted electronically to a.papatheodorou@aegean.gr in MS-Word
format ONLY using British spelling, single-column, 1.5 line spacing, Tahoma letters, font size
11. Section headings (and sub-headings) should be numbered and written in capital letters.
Upon acceptance of a paper and before its publication, the corresponding author will be
asked to sign the Transfer of Copyright form on behalf of all identified authors.
Full Research Papers should contain original research not previously published elsewhere.
They should normally be between 4,000 and 7,000 words although shorter or lengthier
articles could be considered for publication if they are of merit. The first page of the papers
should contain the title and the authors’ affiliations, contact details and brief vitae (of about
50 words). Regarding the following pages, papers should generally have the following
structure: a) title, abstract (of about 150 words) and six keywords, b) introduction, c)
literature review, d) theoretical and/or empirical contribution, e) summary and conclusions, f)
acknowledgements, g) references and h) appendices. Tables, figures and illustrations
should be included within the text (not at the end), bear a title and be numbered
consecutively. Regarding the referencing style, standard academic format should be
consistently followed. Examples are given below:
‚ Airbus (2003), Global Market Forecasts 2003-2022, Toulouse: Airbus.

‚ Fragoudaki, A., Keramianakis, M. and Jancovich, S. (2005) The Greek PSO Experience.
4th International Forum on Air Transport in Remoter Regions. Stockholm, May 24-26.
‚ Forsyth P. (2002a), ‘Privatization and Regulation of Australian and New Zealand Airports’,
Journal of Air Transport Management , 8, 19-28.
‚ Papatheodorou, A. (2008) The Impact of Civil Aviation Regimes on Leisure Market. In
Graham, A., Papatheodorou, A. and Forsyth, P. (ed) Aviation and Tourism: Implications
for Leisure Travel , Aldershot: Ashgate, 49-57.
‚ Skycontrol (2007) easyJet welcomes European Commission’s decision to limit PSO abuse
in Italy. 23rd April. Available from: http:/ /www.skycontrol.net/ airlines/easyjet-
welcomes-european-commissions-decision-to-limit-pso-abuse-in-italy/ (accessed on
22/ 08/ 2008).
iii
Conference Reports should be between 1,000 and 1,500 words. They should provide
factual information (e.g. conference venue, details of the conference organizers), present
the various programme sessions and summarize the key research findings.
Book Reviews should be between 1,000 and 1,500 words. They should provide factual
information (e.g. book publisher, number of pages and ISBN, price on the publisher’s
website) and critically discuss the contents of a book mainly in terms of its strengths and
weaknesses.
Industry Perspectives should be up to 1,000 words and provide a practitioner’s point of view
on contemporary developments in the air transport industry. Contributors should explicitly
specify whether their views are espoused by their organization or not.
iv
Table of Contents
EDITORIAL……………………………………………………………………………………………………………….viii
Andreas Papatheodorou, Kostas Iatrou and Zheng Lei
Full Research Papers

1. PEAK-LOAD PRICING AND AIRLINE REACTIONS AT EUROPEAN
AIRPORTS …………………………………………………………………..……….…………………...1-23
Christos Evangelinos, Jacqueline Stangl and Andy Obermeyer
Conventional wisdom in the economics of pricing holds that peak-load pricing can
enhance welfare in cases where demand peaks are clearly identifiable and highly
predictable. However, this pricing tool has not found acceptance among airlines in
the past. In the very few cases in which peak-load pricing has been introduced,
regulators have faced strong opposition from airlines. Recent research has focused
on whether airlines could pass the additional costs associated with peak-load pricing
on to passengers. Expanding on this work, this paper assesses how peak-load
pricing would impact airline costs and forecasts how airlines would react to the
implementation of a peak-load pricing regime. We use a simultaneous
autoregressive model to predict airline pricing reactions. Our findings indicate that
for certain routes, airlines would subsidize revenue decreases in off-peak times with
price increases during peak times. This finding corroborates the perception held by
airlines that a peak-load pricing regime would encourage new competitors to enter
the market at off-peak times.
2. FLYER'S CORNER: SOME SIMPLE TRUTHS ABOUT SCHEDULED AIR

TRAVEL……………………………………………………………………….…….……………..……...24-38
Amalia Polydoropoulou, Evangelia Chortatsiani and Maria Kamargianni
Airlines in a state of crisis need to take steps to ensure revenues keep coming in.
Repeat business and eventually loyalty is critical in that respect. This paper considers
airline customer buying behaviour and preferences and post-flight attitudes with a
view to identifying what makes the air travel product more appealing to customers.
A structured web-based purpose-designed instrument was used to collect travel and
traveller data and perceptions and the findings are prioritised using a consensus
decision making approach. A major issue is identified and a two-step model is
developed to demonstrate its significance for airline customers’ loyalty. Specific
airline customer communications and other priorities and preferences are identified.
v
In addition, it is found that airline customers have a number of complaints, many of
which are not communicated to the airline, and it is demonstrated that these
complaints impact on loyalty. It is therefore argued that airlines and researchers
alike need to study customer complaints in relation to loyalty in greater detail.
Taking into account the bad economic situation in Greece, in combination with the
airline deregulation, the results of this research is of high significance to Greek
airline companies, in order to “tie” their customers and maintain their market share.
3. A COMPLEX SYSTEM APPROACH IN MODELING AIRSPACE CONGESTION

DYNAMICS…………………..……………………………………………………………………………39-56
Soufian Ben Amor and Marc Bui
The dynamical behaviour of groups of airspace sectors, e.g. Functional Airspace
Blocks, is not trivial to be analysed without appropriate theoretical tools. In this
paper, we suggest a discrete model based on cellular automata and multi agent
systems to express the congestion dynamics and complexity in the controlled
airspace. Discrete time simulations have been performed with random selected
scenarios of traffic and with independent sector parameters to investigate the impact
of availability of local sectors on the whole state of the airspace. Obtained results
show the existence of a traffic threshold that leads to a theoretical saturation of
airspace. The test scenario showed a phase transition phenomenon towards the
congestion of the European airspace at the resulting traffic threshold circa 50 000
flights. Validation using real data shows the predictive abilities of the model.
4. BUSINESS RELATIONS BETWEEN THE LOW COST CARRIERS AND AIRPORTS AS A

CONSEQUENCE OF THE AIR TRANSPORT DEREGULATION…..….…………..……….57-77
Eleftherios D. Katarelos and Iason Koufodontis
The deregulation of air transport has increased competition between air carriers,
resulting in lower fares and increased volumes of passengers. Subsequently, the fare
reduction has altered the market structure with the establishment of new carriers,
strategic alliances and mergers, and the bankruptcy of several traditional airlines
which were unable to adapt to the new environment. The emergence of low cost
carriers (LCCs) is one important outcome of the deregulation. LCCs entered the
market by offering a differentiated product based on bare services offered at
significantly lower prices. The main target was travellers with increased sensitivity in
pricing and less demand for all-around services. The rise in terms of passengers and
flights dictated a better utilization of the fleet, requiring reduced turnaround times at
vi
airports. Many central airports had very little flexibility and capacity necessary to
facilitate additional timeslots. As an answer to inadequate capacity combined with
higher taxes and fees, most LCCs have chosen to use secondary or regional airports.
This choice has altered the balance and strategic importance between airports and
increased their importance for air carriers. This paper examines the evolution and
development of LCCs globally, along with the consequences of their expansion to the
traditional carriers, the market and the passengers. Emphasis is given to the
relationship between LCCs and airports which has resulted in an additional increase
in air travel. The prospects of Greece as a market for LCCs are also being discussed.
5. QUALITATIVE ANALYSI S OF THE RELATIONSHIP BETWEEN THE PROFILE OF

DEPARTING PASSENGERS AND THEIR PERCEPTION OF THE AIRPORT
TERMINAL…………………………………………………………….……………………….……….78-102
Michelle Bandeira and Anderson Correia
The opinion of passengers on the degree of importance of the components is
required in order to prioritize services. A low service level can, besides causing
inconvenience for terminal users, increase the waste of resources and increase costs
if there is no adequate planning. Hence, outlining passenger profiles at the airport is
relevant to strategic planning of airport activity management. I t is believed that
individual characteristics could influence opinion on the degree of importance or
about the quality of airport services. This article shows that the check-in and the
departure lounge were considered the most important areas in the airport terminal
by passengers. Finally it was noted that the age and reason for travel influenced the
passengers’ perception about the check-in area and the frequency of flying
influenced the perception of the departure lounge.
vii
Editorial
This issue of the Journal of Air Transport Studies includes five papers. Evangelinos,
Stangl and Obermeyer examine peak-load pricing and airline reactions at European
airports. They find that airlines would subsidize price reductions at off-peak times with
price increases during peak times on certain routes, suggesting a peak-load pricing regime
would encourage new competitors to enter the market at off-peak times. Subsequently,
Polydoropoulou, Chortatsiani and Kamargianni consider airline customer satisfaction
and loyalty. The results of the research indicate that the majority of passenger complaints
are not communicated to the airline; this finding has significant implications for customer
loyalty.
In the following contribution, Ben Amor and Bui adopt a complex system approach to
model airspace congestion dynamics. The test scenario shows a phase transition
phenomenon towards the congestion of the European airspace at the resulting traffic
threshold of circa 50,000 flights. Then, Katarelos and Koufodontis provide useful
insights into the business relations between the low-cost carriers and airports in the context
of the air transport deregulation. Finally, Bandeira and Correia explore the relationship
between the profile of departing passengers and their perception of the airport terminal.
The research shows that the check-in counters and the departure lounge are considered to
be the most important areas in the airport terminal by passengers. The age and reason for
travel influence the passengers’ perception about the check-in area, while the frequency of
flying influences the perception of the departure lounge.
May we take this opportunity to thank all our authors and referees for their support in
publishing this fifth issue of our Journal. Our continuing partnership with Air Transport
News in conjunction with the open access character of the journal aim at ensuring that JATS
can get a significant exposure to the academic and business audience and raise its profile
accordingly. Enjoy reading!
Dr Andreas Papatheodorou, Editor-in-Chief

Dr Kostas I atrou, Associate Editor
Dr Zheng Lei, Assistant Editor
viii
PEAK- LOAD PRI CI NG AND AI RLI NE REACTI ONS AT EUROPEAN AI RPORTS
Christos Evangelinosa,
Institute for Transport and Economics, “Friedrich List” Faculty of Transportation Sciences,
Dresden University of Technology, Germany
Jacqueline Stanglb,
Deutsche Flugsicherung GmbH (DFS), Germany
Andy Obermeyer c
Institute for Transport and Economics, “Friedrich List” Faculty of Transportation Sciences,
Dresden University of Technology, Germany
ABSTRACT
Conventional wisdom in the economics of pricing holds that peak-load pricing can enhance
welfare in cases where demand peaks are clearly identifiable and highly predictable.
However, this pricing tool has not found acceptance among airlines in the past. In the very
few cases in which peak-load pricing has been introduced, regulators have faced strong
opposition from airlines. Recent research has focused on whether airlines could pass the
additional costs associated with peak-load pricing on to passengers. Expanding on this work,
this paper assesses how peak-load pricing would impact airline costs and forecasts how
airlines would react to the implementation of a peak-load pricing regime. We use a
simultaneous autoregressive model to predict airline pricing reactions. Our findings indicate
that for certain routes, airlines would subsidize revenue decreases in off-peak times with
price increases during peak times. This finding corroborates the perception held by airlines
that a peak-load pricing regime would encourage new competitors to enter the market at
off-peak times.
Keywords: Price differentiation, peak-load pricing, special interest groups, pricing behaviour,
airline reactions
a
Christos Evangelinos is currently researcher at the Dresden University of Technology’s Institute for
Transport and Economics, where he also studied. His research interests revolve around
infrastructure charging, airline pricing and valuation techniques for transportation relevant topics,
such as travel time, externalities etc. Corresponding author: Christos Evangelinos
(Christ os.Evangelinos@tu-dresden.de), Phone: + 49 351 463 36708. We wish to thank the
anonymous referees for their helpful comments. Any errors remain our responsibility.
b
Jacqueline Stangl is currently working at the DFS Deutsche Flugsicherung GmbH as a team
member of the Operational Requirement Management. Before, she studied Transport Economics at
the Dresden University of Technology. Her research interests focus on Air Transport and Air Traffic
Control. The paper represents her personal views and not those of the DFS.
c
Andy Obermeyer is a research assistant at the Dresden University of Technology’s I nstitute for
Transport and Economics. He studied Transport and Economics in Dresden and at the University of
Kent at Canterbury. His research interests are valuation of travel time savings and airline pricing.
1. I NTRODUCTI ON
Although the global financial crisis has had a negative impact on the aviation sector, one can
assume that a return to stability will lead the long-run growth trends in air transport to
continue. According to I CAO forecasts (2007) the air transport passenger market is expected
to increase at the rate of 4.6 per cent annually (in terms of passenger-kilometres). The
problems associated with managing this growth are even more acute if one considers air
freight, which is expected to expand at 6.6 per cent annually. I n the absence of sufficient
capacity expansion, this demand growth may be counterproductive for air transport. First,
passengers will likely deal with considerable delays. Second, environmental costs are
expected to rise. Third, air carriers will bear additional costs resulting from delays. At present,
several European Airports already face severe capacity problems. A forecast of demand
growth to the year 2025 without additional capacity growth predicts excess demand of
around 3.7 million flights (see EUROCONTROL, 2004, pp. 2-11). In concrete terms, this
means that in year 2025, more than 60 European airports are expected to face severe
capacity problems in their peak hours and at least 20 airports will have to cope with capacity
problems not only during a few peak hours, but around 10 hours per day. In Germany, for
instance, this would translate in a situation in which all six major airports (including Berlin-
Brandenburg International Airport, which is currently under construction) face excess
demand during peak times (see Röhl, 2007, p. 8). In light of these expected supply
bottlenecks, it is highly necessary to introduce capacity management systems that will
mitigate the negative impacts of excess demand.
Alongside administrative measures to handle increasing capacity, airport expansion

programmes represent one possible solution. Yet expanding an airport is no easy task. In
every single case in Germany in which plans have been made to expand airport capacity,
significant legal and bureaucratic challenges have arisen. In addition, the ability of
environmental organizations to intervene in the legal process with various objections renders
timely airport expansion projects a near impossibility. The construction of the new Munich
airport, for instance, took a total of 29 years: although construction itself was completed in 6
years, 23 years were needed to work through the 5,724 separate legal challenges (see Röhl,
2007). It is clearly recognizable, therefore, that in the short and medium term, capacity
expansions will not be capable of bringing the scarcity problem under control.
From an economic perspective, pricing measures are another means of handling excess
demand. Efficient airport pricing is a well analysed topic in the literature. Wolf (2003, pp.
Journal of Air Transport Studies, volume 3, Issue 1, 2012 Page 2

121-131), for instance, states that marginal-cost pricing (first-best solution) leads to the
well-known problem of deficit. To cope with this issue, he analyses several price
differentiation schemes, concluding that Ramsey pricing schemes are the appropriate
second-best solution for achieving full-cost recovery and the minimization of welfare losses.
In addition, he mentions that in the short run, peak-load pricing could be a valuable
instrument for coping with capacity problems. I n this way, economization measures to ration
demand in the short to medium term could represent a viable coping strategy. Given certain
assumptions concerning demand and technology, airports could price at marginal costs
during off-peak times and at marginal plus capacity costs during peak times. 1 This pricing
structure assumes that peaks are clearly recognizable. Therefore, peak-load pricing is of no
use if an airport is highly congested with continued excess demand during the whole day. I n
addition, peak-load pricing can also function as a signal for capacity expansion. This is the
case when in spite of peak-load pricing, excess demand at certain times exists, thus
indicating that airport capacity is insufficient.
In addition, slot allocation mechanisms, such as auctions or slot trading, are also widely
discussed in the economic literature on scarcity at airports. The current administrative
system in Europe for allocating slots based on grandfather rights is problematic from an
economic perspective, since it does not ensure that slots are allocated to those who value
them most (Menaz & Matthews, 2008). Furthermore, administrative rationing has been
criticised by anti-trust authorities (see Starkie, 1998, p. 113). The auctioning or trading of
slots can, however, generate an efficient (1st Best) outcome as shown, for instance, in
Brueckner (2009), Verheof (2010) and Basso & Zhang (2009). Despite their theoretical
efficiency these mechanisms have been barely applied at all in practice. This might be due to
practical barriers such as the complementary nature of slots or market power concerns (see
Menaz & Matthews, 2008). Furthermore, Forsyth and Niemeier (2008) point out that the
structure of the airport charges is of similar importance as efficient slot allocation processes.
In particular they show that a combination of a slot allocation process and peak-load pricing
can lead to more efficient airport utilisation. In this paper we do not cover the slot allocation
process itself but focus on peak-load pricing instead.
However, pricing measures such as peak-load pricing are extremely difficult to implement.
This is mainly due to lack of acceptance by existing users. Schank (2005, pp. 417-425)
demonstrates that peak-load pricing in Boston and London failed because of lack of
1
For an in-depth treatment of peak-load pricing, see e.g. Crew et al., 1995.
acceptance by user groups, who managed to form effective opposition to the pricing scheme.
In New York, peak-load pricing at La Guardia airport resulted in the relocation of almost all
commuter flights to Teterboro, a regional airport located in New Jersey. These empirical
findings are not only confirmed in individual cases. In general, peak-load pricing is officially
opposed by the IATA (2000, p. 1; Forsyth & Niemeier, 2003, p. 16) based on the argument
of cross-subsidization.
Starkie (2005, p. 6-7) gives the following reasons for the failure to implement efficient
pricing structures:
Ü First, governmental ownership induces a situation in which a maj ority of airports do

not seek profit maximization (see also Forsyth & Niemeier, 2003, pp. 14-15);
Ü Second, it is very difficult for airport managers to reject the traditional charging
scheme, which is based on the partly erroneous assumption that aircraft weight
correlates with runway damage;
Ü Third, airlines oppose such pricing instruments, although they use similar pricing
schemes themselves (yield management);
Ü Fourth, airport managers are unwilling to adopt such pricing schemes as they are
thought to undermine capacity expansion efforts, in turn preventing higher
passenger volumes over the long run.
One institutional argument in particular should be pointed out. The nature of the regulatory
regime in place can play an important role for efficient pricing (see Laffont & Tirole, 2000, pp.
66-67). In this regard, Starkie (2005) highlights the possible inefficiencies of price-cap
regulation. According to empirical observations, price-cap regulated airports tend to engage
in capacity expansion programmes rather than implement peak-load pricing. The regulatory
environment may weaken incentives for the adoption of efficient pricing structures. First, in
several cases, airport price-cap regulations have been accompanied by the introduction of
sliding scales. Second, the majority of price-cap regulated airports are subjected to single-till
regulation. Regardless of the specific regulatory conditions, the role of special interest
groups is essential. In other words, carriers (especially legacy carriers) attach high
importance to the prevention of peak-load pricing and to the preservation of the existing
2
pricing scheme. Looking for reasons as to why carriers oppose peak-load pricing,
researchers have focused lately on the impact that peak-load pricing has on airline profits. I n
2
For an overview of positive economic theory in transport infrastructure pricing, see e.g. Knockaert
et al., 2009.
this respect, it is crucial to identify whether airlines can shift additional cost burdens to
passengers or not. Forsyth (2008) notes in this connection that additional costs during peak
times cannot be fully passed on to passengers, at least not in all cases. By contrast, savings
from lower charges at off-peak times can be fully passed through to passengers (due to
competitive pressures), thus resulting in lower air fares. I t is therefore essential to study the
impact of peak-load pricing on airlines’ costs as well as to analyze which business strategies
can help airlines to cope with peak-load pricing. This paper is organized as follows: Section
2 identifies the effects of peak-load pricing on airline costs; section 3 addresses possible user
reactions (including pricing reactions) to peak-load pricing; and section 4 concludes.
2. THE EFFECTS OF PEAK-LOAD PRI CI NG ON AI RLI NE COSTS
I n order to study the effects of peak-load pricing to airline costs, we must first classify airline
costs. Traditionally, the ICAO takes into account only operating costs and leaves out
extraordinary costs. Operating costs can be useful in benchmarking airline cost efficiency,
and, at the same time, reveal differences between airlines. Table 1 shows the main elements
of airline operating costs for international scheduled operations, for US and European airlines.
Table 1: Operating Airline Cost Shares for I nternational Scheduled Operations

Direct operating costs US [ % ] EU [ % ]
‚ flight crew
Flight operations 41.7 40.5
‚ fuel and oil

8.0
‚ airport and en-route charges

22.7
9.8
Maintenance 10.0 10.5
Capital costs & insurance 6.1 5.5
Rentals 7.9 4.9
Sum 65 .7 61 .4
I ndirect operating costs
Station and ground 17.0 11.5
‚ cabin staff
Passenger-services 5.6 12.3
‚ other passenger services

7.1
5.2
Sales, ticketing and promotions 5.4 11.1
General and administration 6.3 4.7
Sum 34.3 39 .6
Total 100.0 100.0
Source: AEA (2007), ATA (2007)
Direct operating costs are mainly related to aircraft type and represent almost two third of all
operating costs. Within this cost category, flight operations represent the highest cost

element and vary in accordance with distance travelled (mainly due to increasing fuel
consumption). Direct operating costs also vary significantly according to the type of routes
flown and business model of the carrier. For instance, long-haul operations tend to have a
higher operating cost share than short-haul operations. In addition, this cost share rises to
up to 80 per cent for charter and low-cost carriers (mainly due to indirect operating cost
savings).
An alternate, highly instructive approach for classifying costs employs the standard notion of
fixed and variable costs and is grounded in the concept of escapability. According to this
concept, costs are classified into three maj or categories:
‚ The first category is costs related to flight hours (flying costs, representing around 30 to
45 per cent of total costs). These include expenditures for fuel, flight personnel, direct
maintenance, passenger services and finally airport and ATM charges. Such expenditures
are mainly related to aircraft use, which means that they are escapable if a flight does
not take place.
‚ The second category is fleet-related standing costs (representing around 25 to 30 per
cent of total costs). These costs are only escapable in the medium term, which is
typically one year. This cost category includes aircraft capital expenditures, wages, as
well as overhead costs for maintenance. These costs correlate positively with the activity
level of the carrier, which means carriers can save on these expenses only by reducing
their activity level.
‚ The third cost category is fixed indirect costs (representing around 25 to 35 per cent of
total costs). These costs are only escapable in the long run, and include expenses for
administration, sales, marketing as well as ground station activities.
This cost classification scheme seems to be more useful for assessing the impact of peak-
load pricing on airline costs. For ultimately, carriers pay close attention to the revenues
generated by each single flight when scheduling their networks. In order to keep a certain
city pair on a flight schedule, it is essential that the flight cover at least flying costs. 3 For this
reason, carriers aim to achieve high load factors. The introduction of peak-load pricing would
therefore cause a shift in the break-even point towards higher load factors in the peak period
3
Due to the fixed nature of schedules, short run marginal costs are very low. Therefore, every
additionally ticket sold makes an additional contribution to cost recovery. Peak-load pricing is,
however, a pricing scheme which cannot be implemented in the short run, but rather in the
medium run. We therefore regard flying costs to be the relevant factor for airlines in decisions
concerning when to schedule flights under a peak-load pricing scheme.
and vice versa in the off-peak period, since peak-load pricing raises costs for the airline in
the peak period and reduces them in the off-peak period.
However, the additional cost burdens of peak-load pricing will impact carriers in different
ways. To evaluate these impacts, we must therefore begin by differentiating air carriers
according to the following criteria:
1. the degree of slot scarcity faced by the carrier at the airport in question;
2. the significance of airport charges for the carrier;
3. the level of charges at the airport.
The degree of slot scarcity is the first criterion that can have a decisive impact on the level of
peak-load pricing.4 As excess demand at peak times varies at different airports, the amount
of the charge during the peak period will vary respectively. 5 We therefore conclude that
carriers using an airport as their home base that is slot congested at certain times of day will
bear greater financial burdens than other carriers. In addition, when comparing two airports
that both have excess demand in the peak period, we conclude that the carrier at the airport
with higher peak demand will have to pay more for airport charges if peak-load pricing is
implemented. Finally, due to the cost relatedness of the pricing scheme, airport cost
efficiency can result in cost differences for carriers even if airports have similar slot scarcity.
The second criterion is the share of airport charges as a percentage of airline operating costs.
As shown in table 1, airport charges (including ATM charges) represent 9.8 per cent of total
operating costs. This figure is an average value and reflects predominantly the cost situation
of an international carrier. Depending on geographical factors and the carrier’s business
model, this cost share can increase up to 20 per cent. First, airport-charges cost share
increases for short-haul flight operations and decreases for long-haul operations.6 Therefore,
if a carrier offers predominantly short-haul flights, it is expected that the airport-charges cost
share will rise for operating decisions and vice versa. Second, airport charges are the
dominating factor when low-cost carriers decide whether a destination will be served or not.
4
We regard in this case only airports at which peak-load pricing can bring desirable results. This
does not include airports without any capacity problems, or airports with permanent excess
demand.
5
We note in this case that the form of regulation in place can be a serious barrier to the
implementation of peak-load pricing. There are also many cases in which peak-load pricing could
lead to huge profits. Regulatory regimes aiming at cost recovery would prevent the implementation
of peak-load pricing.
6
For the short-haul operations of British Midland and KLM UK, the airport-charge cost shares in
1999 were 15 and 23.4 per cent, respectively (see Doganis, 2002, p. 146).
Third, despite the internationally similar tariff structure (two-part tariff), the charge level
varies immensely from airport to airport. Table 2 shows these differences for a standard
aircraft type.
As depicted in table 2, the implementation of peak-load pricing at a high cost airport like
Paris CDG or Vienna would burden carriers operating from these airports more than others.
In addition to table 2, US carriers have lower airport-charge cost shares (currently 2.2
percent of total operating costs, see ATA, 2010). The reason for this is twofold: First, carriers
in the US often operate their own terminals. Second, parts of airport charges are paid
directly by passengers. I n this way, US carriers currently have a cost advantage compared to
European carriers.
Table 2: Representative Airport Charges for a B747- 400 with 395t MTOW, 335
passengers and 3h parking time for w inter 2010/ 11, in USD
Charge in US $
Ratio of the
Total Charge
Airport Weight based Passenger based Components
Tokyo NRT 8,649 7,537 16,186 53 : 47

London LHR 2,742 12,363 15,105 18 : 82
Paris CDG 3,755 10,359 14,114 27 : 73
Buenos Aires EZE 3,153 10,218 13,371 24 : 76
Vienna 4,387 8,454 12,841 34 : 66
Amsterdam 4,763 7,058 11,821 40 : 60
Frankfurt 1,359 10,379 11,738 12 : 88
Atlanta 1,363 10,050 11,413 12 : 88
Chicago 3,111 7,705 10,816 29 : 71
London LGW 4,410 6,033 10,443 42 : 58
Madrid 5,119 5,026 10,145 50 : 50
Bangkok 1,892 7,739 9,631 20 : 80
Singapore 3,600 5,981 9,581 38 : 62
Manchester 4,261 4,354 8,615 49 : 51
Nairobi 1,880 6,700 8,580 22 : 78
Rome FCO 1,461 6,422 7,883 19 : 81
Hong Kong 3,560 0,990 4,550 78 : 22
Source: Own calculations
The third criterion is the extent to which the peak-load pricing scheme is applied. As table 2
shows, the degree of variability – that is, the ratio between the fixed (aircraft-related) and
variable (passenger-related) components of the charge – fluctuates significantly between
airports. Although lately a shift towards greater variability has occurred, European airports
still charge a certain fixed amount based on aircraft MTOW. Applying peak-load pricing only
to the fixed cost component would severely discourage full-service carriers (FSCs) from

expanding flight frequency (see Brueckner, 2010; Givoni & Rietveld, 2009), as FSCs typically
operate large networks and also commit themselves to offering frequent flights in order to
minimize the passengers’ schedule delay.
3. AI RLI NE REACTI ONS TO PEAK-LOAD CHARGES
In spite of the fixed nature of airport charges over the short run, airlines have certain
opportunities for countering their impact on cost structures. Over the medium-term carriers
can implement various strategies to steer direct and indirect operating costs, thus allowing
the impact of additional peak-load expenses to be mitigated. In this regard, we draw a
distinction between operational measures (such as the choice of aircraft size and location
effects) and pricing measures. In particular, we discuss how airlines can evaluate the best
strategies to implement.
3.1 I NCREASE I N AI RCRAFT SI ZE/ REDUCTI ON OF FLI GHT FREQUENCY
From the regulator’s point of view, one of the desirable airline reactions would be the use of
larger aircraft 7 combined with a reduction in flight frequency during peak periods. Aircraft
size is an important determinant of unit costs per passenger, because it has a direct
influence on operational costs and hourly productivity. Cost advantages are achieved with
larger aircraft due to several factors; aerodynamic benefits and larger, more efficient engines,
for example, reduce fuel consumption per weight unit. Furthermore, a larger aircraft also
leads to higher labour productivity. Thus, costs per seat-kilometre decrease with increasing
aircraft size. Figure 1 illustrates this interrelation.
The average cost curve for smaller aircraft (AC1) reaches its minimum at the point of
maximum seat load capacity. Assuming a utilization factor beyond X, the use of the larger
aircraft would decrease average costs. In addition, the extent to which it is possible to
compensate for increased airport charges depends not only on the cost advantages attained
but also on prevailing passenger preferences and demand characteristics. If the increase in
charges is very large compared to total costs, then these charges can only be offset through
7
A better usage of airport capacity would be achieved not only by a cutback in frequency, but also
by the fact that larger aircraft have lower wake-vortex separation requirements and lower runway
occupancy times. Consequently, larger aircraft typically occupy runways for a shorter time than
lighter ones (see Wolf, 2003, p. 65 ff.; Doganis, 1992, p. 83).
sufficient passenger load and adequate marginal return per seat. 8 Hence, a half-empty
aircraft would be less cost efficient than two smaller, highly utilized aircrafts.
Figure 1: Economies of Size & Fill in Relation to I ncreased Aircraft Size
EUR
AC1
AC2
X
} max. saving
potential
seat capacity (K)

max K1 max K2
Source: Own illustration, based on Button (1982), p. 79.
If an airline pools two flights together during the peak period (assuming identical demand for
each flight), it can be expected that some time-sensitive passengers will be lost. To pre-empt
this disadvantage it is necessary to choose an aircraft according to the future demand
situation. Based on current average load factors of 70–80 per cent (see AEA, 2008, p. 8),
this would mean additional cost benefits (see Wei, 2006). However, if an airline loses many
time-sensitive passengers, financial penalties are likely. In this regard, so-called high-yield
traffic is the most likely customer segment to be lost, as these customers are more sensitive
to flight frequencies (see Hanlon, 1996, p. 167; NERA, 2004, p. 83 ff.). Yet a cutback in
frequencies seems to be a reasonable option for certain sub-segments. On routes dominated
by business travellers or hub flights with quick transfer guarantees, frequency reductions
would lead to lower revenues. In such markets, airlines will be unwilling to change
frequencies. The implementation of such strategies is more probable in the case of point-to-
point short-haul flights due to the larger impact of increased charges.
8
The most cost-effective combination in general is maximum range with a full payload. Hence, a
suitable traffic density is necessary to tap the full cost advantages from maximum load (see
Doganis, 2002, p. 122).
High frequencies imply both greater flexibility and more flight hours. The higher the aircraft
utilization, the lower the average costs. Consequently, lower frequencies are more costly and
a cutback in frequency can have negative effects on productivity, especially on short- and
medium-haul flights (see Doganis, 2002, p. 133 ff.). Ultimately, therefore, the airline reaction
will be determined by the interplay of these various factors. However, there are two
additional factors that hinder a possible implementation of lower frequencies:
1. Airport schedules: Existing schedules mean frequency changes can only be

implemented in the medium to long run.
2. Large aircraft availability: Not all carriers can switch to larger aircraft. Because of
their homogenous fleet structures, low-cost carriers in particular have a limited
ability to introduce larger aircraft compared to FSCs.
In summary, the feasibility of introducing lower frequencies and larger aircraft depends
strongly on specific market and demand characteristics. For example, Givoni & Rietveld
(2009) have shown that service frequencies are not only significant in terms of the time and
price elasticity of passengers. 9 They are also an important instrument in competition, and
can strongly influence a carrier’s choice of aircraft.
3.2 TEMPORAL AND SPATI AL RELOCATI ON EFFECTS

A further possible reaction of carriers is the reassignment of flights to off-peak times. While
such a reassignment is a primary goal of peak-load charges, it can only be achieved if
monetary incentives are strong enough to motivate a rescheduling of arrival and departure
times. The primary aims in flight scheduling are to achieve high aircraft utilization; the
optimal timing of flights to cater to passenger time preferences; and high market shares.
These considerations as well as several operational and external conditions (e.g. night-flight
restrictions, maintenance requirements and the availability of slots) can considerably impede
flexible flight planning (see Lüking, 1993, p. 249, 253 ff.).
I n general, peak flights are strongly favoured by passengers. Therefore, the loss in revenue
connected with rescheduling to off-peak times should not be underestimated. Alongside this
expected commercial disadvantage, aircraft size is also of importance. On routes with high
demand, rescheduling to lower demand periods can require flights to be combined or the
operation of smaller aircraft in order to reach an adequate load factor. Hence, extensive
9
For a theoretical discussion, see Fischer, 1997, pp. 101-114.
rescheduling towards off-peak periods would be connected with demand losses and cost
disadvantages due to a decrease in the usage of aircraft and flight crews. Because of the
required adjustment in fleet structure, rescheduling seems to be a less attractive solution as
a response to peak charges. Savings in the off-peak period have to be substantial to
compensate for the operational and commercial disadvantages.
Similarly, in the case of frequency reductions, the ability of rescheduling to be implemented

is limited by existing slot allocation procedures in Europe. Currently allocation of slots at
Community airports primarily takes place according to Council Regulation (EEC) No 95/ 93. At
airports with serious capacity shortages (declared as coordinated airports) air carriers need
permissions to use the facilities for take-off and landing at a particular time, which are
allocated bi-annually by an independent coordinator for the entire respective flight plan
(summer or winter season). 10 Therefore, at slot coordinated airports, the rescheduling of
flights can only be implemented in the next period while taking into account the associated
condition of slot pairing. Furthermore, rescheduling to off-peak at the airport of origin can
potentially lead to increased activity during peak times at the destination airport. Hence, it
might become very difficult to find slots at the destination airport.
According to the current allocation principles air carriers can claim slots in the next
scheduling period if they are utilized for at least 80% otherwise the slots will be returned to
the slot pool for reallocation to competitors (so-called grandfather right and use-it-or-lose-it
rule). From this it follows that, because of competition issues, there are serious doubts that
airlines will be willing to give up their valuable peak-time slots. There is evidence that
established carriers use slots as a barrier to entry in order to increase demand for their own
services (see Starkie, 1998, p. 113). This argument is even stronger if one takes into account
that during peak periods, carriers realize scarcity rents (see for instance Menaz and
Matthews, 2008).
Another alternative mainly applicable to low-cost carriers is the relocation of operations to

less congested secondary airports with available capacity and no peak charges. In contrast
to the relocation of flights to off-peak periods, this approach enables more attractive flight
times. However, due to expected demand losses (especially in transfer traffic), airlines have
10
The mentioned regulation act should ensure, that the allocation happens in a neutral, transparent
and non-discriminatory way. For further details see Council Regulation (EEC) No 95/ 93 of January
1993 on common rules for the allocation of slots at Community airports and amending acts,
basically Regulation (EC) No 793/ 2004.
only shown minor interest in pursuing such a strategy to date. The use of peripheral airports
is a reasonable alternative only for charter airlines or some low-cost airlines, as their supply
of short- and medium-haul point-to-point flights is mainly directed at the price-sensitive
segment of leisure travellers. We would not expect FSCs that use the airport as base station
to implement such a strategy.
3.3 PRI CI NG STRATEGI ES

Compared to other operational conditions, pricing strategies can be changed relatively fast.
The question as to whether carriers can shift additional cost burdens to passengers without
significantly eroding demand hinges on several factors. The demand characteristics and
preferences of customers are of high relevance in this regard, yet also important are route
lengths, the commercial and operational significance of a route in an airline’s network as well
as the market structure and competitive environment in the given city pair.
In terms of demand characteristics, it is assumed that long- and short-haul passengers will
be affected differently depending on the customer segment and airline business model in
question. The share of business and leisure travellers that fly a route determines to large
degree how much an airline can increase fares without incurring revenue losses. Low-cost
and charter airlines, which cater first and foremost to the price-sensitive group of leisure
travellers, would be particularly limited in their ability to pass additional costs to passengers.
Therefore, compared to FSCs, they would have to bear a large proportion of increased costs
themselves, which also means they would be faced with a competitive disadvantage (for an
analytical and simulative analysis, see e.g. Fu et al., 2006).
Differences may also arise in the ability to shift costs in relation to route distance. Given the
fixed character of airport charges, a peak premium will affect ticket prices very differently.
First, in the case of long-haul flights, supplementary charges represent a lower percentage of
the overall ticket cost, and there are possible advantages due to economies of size and fill.
Apart from this, flight distance is a key determinant of demand elasticity, which is lower for
long-haul flights because of the limited number of alternatives (see e.g. Brons et al., 2002,
p. 172). As a result, in long-haul markets the potential to pass on costs is much larger than
in short-haul ones. For short-haul operations, it is crucial to consider both the degree of
competition with other modes of transport as well as the ratio of business to leisure
passengers on a certain route. The higher the share of business travellers, the easier it is to
shift additional costs. In domestic markets with a low share of business travellers, cost
shifting is apparently difficult to implement. Here, complete cost shifting to passengers would
imply that the required price premium per passenger is sufficiently high enough to offset
both the peak charge and losses due to decreased demand. If the cost increase is really high
(and this can be expected in short-haul markets), then a result would be the phasing out of
some routes, especially point-to-point flights without a commuter function.
Aside from demand related issues, factors like market structure and competitive behaviour
can considerably affect an airline’s scheduling and pricing policy. The oligopolistic market
structures and tendencies towards collusive behaviour that often characterize the airline
sector (see Starkie, 2002, p. 64) seem in general to provide carriers with possibilities for fare
increases. Depending on the commercial importance of a route and the intensity of
competition, certain strategic relationships among the actors can also limit the potential for a
rise in prices. If there is a leader–follower situation, the follower would prefer a limited scope
for cost shifting, because this would compel the leader to maintain fare levels. Such a
dynamic can be observed in the case of feeder flights (see Stangl, 2008. pp. 75–88). I n
addition, the current slot allocation system in Europe seems to enhance airlines’
opportunities for increasing fares. Quite in contrast to other situations of scarcity, the slot
allocation system has the ability to weaken competition among carriers, thus offering a
certain leeway for price increases (see Lüking, 1993, p. 271).
The complete transfer of savings to passengers in the off-peak period also seems to be an
unlikely outcome. A carrier has little motivation to reduce fairs if sufficient load at current
fares can be achieved. In this regard, competition takes place predominantly with regard to
non-pricing criteria such as flight frequency and service amenities. However, market entries
can change such an equilibrium (e.g. when low-cost airlines enter the market; see Forsyth &
Niemeier, 2003, p. 11). Furthermore, if sub-markets are characterized by low passenger
volumes, then a partial pass-through of cost savings would seem to be reasonable in order
to capture additional demand. In-house capacity policies as well as strategic interactions
among market actors are thus significant determinants of airline behaviour. The extent to
which an airline is ultimately able to pass on cost increases to customers in the medium to
the long term therefore varies according to the structure of the market and demand factors,
which vary in relation to the sub-market.
For this reason, conclusions about the degree of cost shifting that will be possible can only
be drawn for the different sub-markets. Conceivably, carriers might prefer to first exploit

internal cost-saving potentials on short-haul routes instead of adopting pricing measures that
risk a decrease demand and loss of market share. This would result in an increase in market
concentration (due to relocation or market entries and exits), in turn encouraging higher
fares (especially those charged by the hub carrier). The attendant increase in competition in
other markets could encourage fare reductions (see De Wit & Burghouwt, 2007, p. 111).
Such displacement effects are possible across specific routes or flights, as carriers seek to
fully tap price-inelastic customer segments. This is confirmed by experiences at Heathrow
airport, where peak passenger charges were redistributed not directly but passed on to all
passengers (see Doganis, 1992, p. 97). Finally, the degree to which cost shifting is possible
determines the ability of a scarcity-based charging policy to enhance relocation and the more
efficient use of airport capacity. Particularly in long-haul segments and for routes with a
large share of price-inelastic business travellers, the effectiveness of peak-load pricing seems
to be very limited. The low time sensitivity of passengers and the necessity of changes within
the hub and spoke scheduling facilitate the shifting of costs to ticket prices, which already
tend to be very high in the absence of opportunities for market entry (see Lüking, 1993, p.
123f.). We therefore conclude that peak-load pricing will have only limited effects with
respect to a change in supply behaviour.
As pointed out previously, the introduction of peak-load pricing schemes at airports can
reduce profits if the airlines are unable to offset additional capacity costs in the peak period
by means of operational cost reductions or pricing measures. Yet the fact that airlines
oppose the introduction of peak-load pricing might also be driven by additional competitive
considerations. Under a peak-load pricing scheme, airport capacity during off-peak periods is
priced only at marginal costs. Airlines may fear that this will encourage additional carriers,
particularly low-cost airlines, to enter the market.
We hypothesize that airlines face tough competition over passengers in off-peak periods.
Consequently, each airline will react to a competitor’s price adjustments. If so, we should
observe significant pricing interdependencies in off-peak periods, with this effect weakening
during peak periods. To test our hypothesis, we investigated the pricing behaviour of
Lufthansa (LH) and Air Berlin (AB) for the airport-pair Berlin-Tegel (TXL) – Frankfurt/Main
(FRA). For this purpose, we collected the lowest offered fares of both airlines in a peak and
an off-peak period starting three months prior to departure. 11 The main elements of the
chosen flights can be seen in table 3.
11
More precisely, we collected prices over 91 days prior to departure for flights on 25 August 2008.
Table 3: Flight times and aircraft types for the Berlin- Tegel – Frankfurt/ Main route
Airport Pair Peak Flight Time Off- Peak Flight Time Carrier
( Aircraft) ( Aircraft)
17: 20 (A 321) 21: 15 (A 321) Lufthansa (LH)
TXL - FRA
16: 55 (B 737-800) 21:25 (B737-800) Air Berlin (AB)
Figure 2: Plotted peak and off- peak prices of Lufthansa and Air Berlin
400
Ticket price in €
350
300
250 Price LH - Off-Peak

Price AB - Off-Peak
200
Price LH - Peak
Price AB -Peak
150
100
50
0
91 84 77 70 63 56 49 42 35 30 27 24 21 18 15 12 9 6 3 0
Days before departure
The gathered peak and off-peak ticket prices for Lufthansa and Air Berlin are plotted in
Figure 2. From the figure it is evident that the peak and off-peak prices follow nearly the
same trend line until three weeks before departure. Furthermore, as the departure date
comes closer, peak prices rise faster than off-peak prices. I n terms of differences,
Lufthansa’s prices increase very sharply as the departure comes closer, while Air Berlin
increases its prices more gradually. The summarized statistics in table 4 shed light on the
average price and distribution. Average peak prices are higher than average off-peak prices.
Lufthansa’s average prices and price dispersion are generally higher than Air Berlin’s.
Table 4: Descriptive statistics of peak and off- peak prices for Lufthansa and Air Berlin
Variable Arithmetic Average Coefficient of Variation

挑張
Lufthansa off-peak price (喧墜捗捗 ) 83.54 0.76
凋喋
Air Berlin off-peak price (喧墜捗捗 ) 77.19 0.22
Lufthansa peak price (p宅滝丹奪叩谷 ) 106.62 0.73
凋喋
Air Berlin peak price (喧椎勅銚賃 ) 96.71 0.33

With the help of an econometric model, we analyse if carrier price-setting behaviour is
substantially different between peak and off-peak periods. For our analysis we employ a
seemingly unrelated regression model (SUR) analogous to the one used by Pels and Rietveld
(2004) to analyse the London—Paris market. Our econometric model is described by the
following SUR equations.
凋喋凋喋凋喋凋喋凋喋凋喋凋喋挑張
(1) 喧墜捗捗 = 糠墜捗捗 + 紅墜捗捗穴欠検嫌痛 + 紅墜捗捗穴欠検嫌態痛 + 紅墜捗捗喧墜捗捗 + 紅墜捗捗喧墜捗捗 +
痛怠態戴痛貸怠替痛貸怠
凋喋凋喋凋喋挑張
紅墜捗捗喧椎勅銚賃 + 紅墜捗捗喧椎勅銚賃
泰痛貸怠滞痛貸怠
挑張挑張挑張挑張挑張凋喋挑張挑張
(2) 喧墜捗捗 = 糠墜捗捗 + 紅墜捗捗穴欠検嫌痛 + 紅墜捗捗穴欠検嫌態痛 + 紅墜捗捗喧墜捗捗 + 紅墜捗捗喧墜捗捗 +
挑張凋喋挑張挑張
紅墜捗捗喧椎勅銚賃 + 紅墜捗捗喧椎勅銚賃
凋喋凋喋凋喋凋喋凋喋凋喋凋喋挑張
(3) 喧椎勅銚賃 = 糠椎勅銚賃 + 紅椎勅銚賃穴欠検嫌痛 + 紅椎勅銚賃穴欠検嫌態痛 + 紅椎勅銚賃喧墜捗捗 + 紅椎勅銚賃喧墜捗捗 +
凋喋凋喋凋喋挑張
紅椎勅銚賃喧椎勅銚賃 + 紅椎勅銚賃喧椎勅銚賃
挑張挑張挑張挑張挑張凋喋挑張挑張
(4) 喧椎勅銚賃 = 糠椎勅銚賃 + 紅椎勅銚賃穴欠検嫌痛 + 紅椎勅銚賃穴欠検嫌態痛 + 紅椎勅銚賃喧墜捗捗 + 紅椎勅銚賃喧墜捗捗 +
挑張凋喋挑張挑張
紅椎勅銚賃喧椎勅銚賃 + 紅椎勅銚賃喧椎勅銚賃
沈沈
The variables 喧墜捗捗 and 喧椎勅銚賃 denote the price charged by carrier i in time period t for an
痛痛
off-peak and a peak flight, respectively. Each single price is regressed on the number of days
until departure as well as on the lagged off-peak and peak prices of the considered carrier
and its competitor. This allows us to investigate the airline’s price responses in the short-run.
The estimation results are shown in table 5. A regression of the residuals on the lagged
residuals and the other explanatory variables does not reveal any autocorrelation.12 The high
values for the adjusted R² imply a good fit of the model.
Table 5: Estimation results for price setting behaviour
Equation (TXL-FRA) AB Off-Peak LH Off-Peak AB Peak LH Peak

Variable Estimate (Standard Error)
Constant 18.9033 (6.23)* * * -124.094 (29.32)* * * 35.8761 (10.67)* * * -34.7766 (41.96)
Days until departure 0.0998 (0.10) -0.0216 (0.48) -0.2874 (0.17) -1.2550 (0.68)*
2
(Days until departure) -0.0016 (0.001) 0.0055 (0.004) 0.0016 (0.002) 0.0146 (0.01)* *
AB Off-Peak 0.3728 (0.11)* * * 1.1838 (0.48)* * -0.4101 (0.17)* * -0.6141 (0.68)
LH Off-Peak 0.0488 (0.01)* * * 0.9472 (0.07)* * * 0.0235 (0.02) 0.1823 (0.09)*
AB Peak 0.2431 (0.08)* * * 0.5984 (0.04) 0.9724 (0.13)* * * 1.6349 (0.52)* * *
LH Peak 0.0269 (0.02) -0.2320 (0.10)* * 0.0527 (0.04) 0.3405 (0.15)* *
adj
R 0.9752 0.9589 0.9774 0.9432
* significant at 10%; * * significant at 5%; * * * significant at 1%
12
For details on the methodology, see Pels & Rietveld, 2004. Results are not presented here but are
available upon request.
The constants are highly significant except for Lufthansa in the peak period. The number of
days until departure and its squared value are significant only for Lufthansa in the peak
period. However, the p-value for the respective parameter of Air Berlin is close to the 10%
significance level. Therefore, it seems that Air Berlin and Lufthansa significantly adjust their
prices in the peak period but not in the off-peak period according to the number of days
before departure. Even though, this adjustment process is different for the two carriers;
Lufthansa increases its price in a quadratic fashion as the departure day comes closer while
Air Berlin follows a more linear price trend.
While in the peak period the number of days before departure but not the lagged prices of
the competing carrier seem to be the predominant influence on price setting, this picture
reverses in the off-peak period. In the off-peak period Air Berlin adjusts its price in period 建
according to its own prices in the former peak and off-peak period as well as in response to
the off-peak price of Lufthansa. Moreover, Lufthansa reacts to the off-peak price of Air Berlin
and to its own peak and off-peak prices. In the peak period Air Berlin independently sets its
own peak and off-peak prices and does not react to Lufthansa’s price while Lufthansa takes
both Air Berlin’s peak price and its own lagged prices into account. Apart from the two
挑張凋喋
exceptions 紅墜捗捗 and 紅椎勅銚賃 , all significant price reactions on the lagged variables are
滞戴
positive.
We can draw three major conclusions from these results. First, in the peak period Lufthansa
as well as Air Berlin increase their prices significantly as the departure date comes closer.
This effect is reinforced by the positive reactions to their own lagged peak-period prices.
Second, both carriers are close competitors in the off-peak since they react on each other’s
prices positively, e.g. if Lufthansa reduces its price in the off-peak period Air Berlin will follow
and vice versa. But this effect is less significant in the peak period, when only Lufthansa
reacts to Air Berlin’s price but Air Berlin does not react to Lufthansa’s price. Third,
interpreting the negative impact of the Lufthansa’s lagged peak price on its off-peak price
and the negative impact of Air Berlin’s off-peak price on its peak price is less obvious. These
price trends are an indication of cross-subsidization between off-peak and peak flights.
The results support our thesis that airlines try to attract passengers in off-peak periods by
adjusting prices according to the prices of their competitors. This effect is less evident during
peak periods. Furthermore, the results indicate that tickets during off-peak periods might be
cross-subsidized by higher prices during the peak period. Hence, we can infer from this

result that if airlines face competition in the off-peak period, they will tend to cross-subsidize
between peak and off-peak periods. Yet to do so, an increase in peak prices is probably
necessary. Particularly in the case of routes with a high proportion of price sensitive
customers, 13 airlines will struggle to increase prices substantially without incurring significant
passenger losses. For this reason, carriers may fear that peak-load pricing schemes will
induce increased competition during the off-peak periods.
4. CONCLUSI ONS
In this paper we sought to identify the reasons for airline opposition to peak-load pricing.
We first considered the effects of peak-load pricing on airline costs. Due to varying business
models and differences in route networks, airlines have different cost structures. This, in
turn, means the introduction of peak-load pricing will “hurt” carriers in different ways. Trip
length (short haul vs. long haul), service quality (FSC vs. LCC) and the geographic base
(European vs. non-European) are significant factors that lead to different cost structures.
We subsequently reviewed possible airline reactions to peak-load pricing. Although airlines
have a relatively strong ability to influence direct operational costs, they often face diverse
external constraints, which hinder an effective response to pricing signals. Due to operational
factors (e.g. fleet management and vehicle schedules), regulatory conditions, as well as
company-specific restrictions, carrier freedom of action is constrained in the near term.
However, in the medium to long run, airlines have a wide spectrum of opportunities to react
effectively to the implementation of peak pricing structures at airports.
Given the very low marginal costs for the transport of additional passengers and the
relevance of pricing policies that encourage high load, demand circumstances play a key
role. They define not only the potential for price reductions, but also the extent to which
price adjustments are feasible. Furthermore, carrier flexibility to introduce route and flight
time changes varies considerably. It is assumed that carriers who offer international long-
haul services are less flexible with respect to flight time adjustments. However, due to the
lack of travel alternatives for passengers, there is greater flexibility for pricing measures in
the long-haul market than there is in the short-haul one.
13
Here, for instance, we think of routes with a low share of business customers and a high share of
leisure customers, respectively.
In contrast to the use of larger aircraft, there is little practical incentive to rescheduling
operations to off-peak periods. The operational and commercial disadvantages of
rescheduling to off-peak turn out to be higher than the operating cost savings. The
disadvantages also tend to outweigh the benefits in the case of secondary airport use; only
cost-oriented carriers, such as charter airlines or LCCs, can be expected to derive a net
benefit from relocating flights to secondary airports.
A key constraint to the implementation of operational changes is the current slot allocation
procedure. Because of the existence of so-called “grandfather rights” and resulting
tendencies to strategically hoard slots, legacy carriers in particular have few incentives to
adjust their operating schedules. If the monetary inducements of peak charges are not
sufficient to mitigate hoarding behaviour, subsidiary measures such as a slot reservation fee,
the tightening of the so called “use-it-or-lose-it” rule, as well as the allowance of slot trading
would be beneficial. Such measures would augment the monetary incentives of peak
charges. 14 Taking into account these difficulties, it is assumed that under the current
regulatory environment, limited options are available to motivate carriers to change their
operations. In this connection, additional important factors include the role of an airport in a
carrier’s network as well as the availability and accessibility of adequate secondary airports.
The ability of carriers to implement price changes hinges to large extent on the sub-market
in question. If increases in costs can be passed along to customers, carriers will probably
prefer to do this rather than extensively re-structure their operations. Consequently, given
the unchanged slot rents during peak periods, rescheduling to less usage intensive times
probably won’t take place. The goal of peak charges – to force the effective rationing of
demand – would not be achieved.
Furthermore, it is still an open question as to whether carriers oppose peak-load pricing

schemes merely because of their potential to reduce profits, or whether additional
competitive considerations play a role, for marginal-cost pricing during off-peak periods
seems to be an invitation for low-cost carriers to enter the market. Using a simultaneous
pricing model for a domestic airport pair in Germany, we showed that competition in the off-
peak period may lead to cross-subsidization between peak and off-peak periods. Although
this may not hold for all airport-pairs the finding remains that incumbent airlines will suffer
14
Furthermore, remuneration mechanisms are possible, e.g. certain discounts, bonuses, or a
lowering of other charging elements would offer additional incentives towards a modification of
operating patterns.
losses. 15 Precisely this finding seems to be in-line with Forsyth’s (2008) conclusion that the
introduction of peak-load pricing leads to higher consumer surpluses and lower airline
profits.
REFERENCES
‚ AEA- Association of European Airlines (2008), State of the Industry, Brussels: May 2008.
‚ AEA-Association of European Airlines (2007), Operating Economy of AEA Airlines 2007,
Brussels: December 2007. URL: http:/ /files.aea.be/RI G/Economics/DL / SumRep07.pdf.
‚ ATA-Air Transport Association (2010), ATA Quarterly Cost Index: U.S. Passenger Airlines,
Washington DC. URL: http:// www.airlines.org/ Economics/ DataAnalysis/Pages/Quarterly
CostIndex.aspx
‚ ATA-Air Transport Association (2007), 2007 Economic report , Washington DC. URL:
http://www.airlines.org/Economics/ReviewOutlook/Documents/ 2007AnnualReport.pdf.
‚ Basso, L.J., Zhang, A. (2009), ` Pricing vs Slot Policies When Airport Profits Matter´ , CTS
Working Paper 2009-1.
‚ Brons, M., Pels, E., Nijkamp, P., Rietveld, P. (2002), ` Price elasticities of demand for
passenger air travel: A meta-analysis´ , Journal of Air transport Management, 8, 165-175.
‚ Brueckner, J.-K. (2010), ` Schedule Competition Revisited´ , Journal of Transport
Economics and Policy, 44, 261-285.
‚ Brueckner, J.-K. (2009), ` Price vs. Quantity-Based Approaches to Airport Congestion
Management´ , Journal of Public Economics, 93, 681-690.
‚ Button, K.J. (1982), Transport Economics, London: Heinemann.
‚ Crew, M.A., Fernando, C.S., Kleindorfer, P.R. (1995), ` The Theory of Peak-Load pricing: A
Survey´ , Journal of Regulatory Economics, 8 (3), 215-248.
‚ De Wit, J., Borghouwt, G. (2007), ` The impact of secondary slot trading at Amsterdam
Airport Schiphol´ , SEO Economic Research, 957, A report for The Netherlands Ministry of
Transport, DGTL, Amsterdam, March.
‚ Doganis, R. (2002), Flying off course: The economics of international airlines, 3rd edition,
London: Routledge.
‚ Doganis, R. (1992), The airport business, London/ New York: Routledge.
‚ Eurocontrol (ed.) (2004), Challenges to Growth 2004 Report (CTG04) , Brussels:
December. URL: http:/ /www.eurocontrol.int/eatm/gallery/content/public/ library/CTG04
_report.pdf.
‚ European Economic Community, (1993), Council Regulation (EEC) No 95/ 93 of 18 January
1993 on common rules for the allocation of slots at Community airports, official journal L
014, 22/ 01/ 1993, Brussels, as well as amending acts. URL: http://eur-
15
This result also suggests that there are some possible demand interdependencies between peak
and off-peak periods, an issue which is not touched in this paper, however.
lex.europa.eu/smartapi/ cgi/ sga_doc?smartapi!celexplus!prod!CELEXnumdoc&numdoc= 39
3R0095&lg= en.
‚ Fischer, R. (1997), ` Time Sensitivity of Passengers and Market Structure in the Airline
Industry: A Model of International Air Transport´ , In Lang, P. (ed.), Europäische
Hochschulschriften, V, Volks- und Betriebswirtschaft, 2135, Bern/Berlin.
‚ Forsyth, P., Niemeier, H.-M. (2008), Prices and Regulation in Slot Constrained Airports. I n
Czerny, A.I., Forsyth, P., Gillen, D., Niemeier, H.-M. (ed.), Airport Slots, International
Experiences and Options for Reform, Aldershot: Ashgate, 127-148.
‚ Forsyth, P. (2008), ` The I mpact of Emerging Aviation Trends on Airport Infrastructure´ ,
‚ Fu, X., Lijesen, M., Oum, T. H. (2006), ` An Analysis of Airport Pricing and Regulation in
the Presence of Competition Between Full Service Airlines and Low Cost Carriers´ , Journal
of Transport Economics and Policy, 40 (3), 425-447.
‚ Givoni, M., Rietveld, P. (2009), ` Airline’s choice of aircraft size – Explanations and
implications´ , Transportation Research Part A, 43, 500-510.
‚ Hanlon, P. (1996), Global Airlines: Competition in a Transnational Industry, 1st ed.,
Oxford: Butterworth Heinemann.
‚ International Air Transport Association (IATA) (2000), ` Peak/Off-Peak Charges´ , Ansconf
Working Paper, 82, ICAO, Montreal.
‚ International Civil Aviation Organization (ICAO) (2007), I CAO Environmental Report 2007.
URL: http:/ /www.icao.int/icao/en/env/pubs/env_report_07.pdf
‚ Knockaert, J., Evangelinos, C., Rietveld, P. and Wieland, B. (2009), ` Differentiated
Infrastructure Charging: A Comparison of Theory and Practice´ , European Transport/
Transporti Europei, 43, 4-34.
‚ Laffont, J.-J. and Tirole, J., (2000), Competition in Telecommunications, , Cambridge,
Massachusetts: MIT Press.
‚ Lüking, J. (1993), ` Angebotsplanung und Fluggastverhalten im überlasteten
Luftverkehrssystem´ , In Lang, P. (ed.), Europäische Hochschulschriften, V, Volks- und
Betriebswirtschaft, 1409, Bern.
‚ Menaz, B., Matthews, B. (2008), Economic Perspectives on the Problem of Slot Allocation,
In Czerny, A.I., Forsyth, P., Gillen, D., Niemeier, H.-M. (ed.), Airport Slots, International
Experiences and Options for Reform, Aldershot: Ashgate, 21-39.
‚ NERA – National Economic Research Associates (ed.) (2004), ` Study to assess the effects
of different slot allocation schemes´ , A Report for European Commission, DG TREN,
January.
‚ Pels, E., Rietveld, P. (2004), ` Airline Pricing Behaviour in the London-Paris Market´ ,
‚ Röhl, K.-H. (2007), ` Das System der deutschen Flughäfen – fit für die Zukunft?´ I W –
Trends, March.

‚ Schank, J.L. (2005), ` Solving airside congestion: Why peak runway pricing is not
working´ , Journal of Air Transport Management , 11, 417-425.
‚ Stangl, J. (2008), ` Auswirkungen der Spitzenlasttarifierung von Flughäfen auf die
Fluggesellschaften´ , Diploma Thesis, Dresden.
‚ Starkie, D. (2005), ` Making airport regulation less imperfect´ , Journal of Air Transport
Management , 11, 3-8.
‚ Starkie, D. (2002), ` Airport Regulation and competition´ , Journal of Air Transport
Management , 8, 63-72.
‚ Starkie, D. (1998), ` Allocating airport slots: a role for the market?´ , Journal of Air
Transport Management , 4, 111-116.
‚ Verhoef, E.T. (2010), ` Congestion pricing, slot sales and slot trading in aviation´ ,
Transportation Research Part B, 44, 320-329.
‚ Wei, W. (2006), ` I mpact of landing fees on airlines’ choice of aircraft size and service
frequency in duopoly markets´ , Journal of Air Transport Management , 12, 288-292.
‚ Wolf, H. (2003), ` Privatisierung im Flughafen Sektor: Eine ordnungspolitische Analyse´ ,
Kieler Studien, Springer, Berlin.

FLYER'S CORNER: SOME SI MPLE TRUTHS ABOUT SCHEDULED AI R TRAVEL
Amalia Polydoropoulou1
Department of Shipping, Trade and Transport (STT), University of the Aegean, Greece
Evangelia Chortatsiani 2
Maria Kamargianni 3
ABSTRACT
Airlines in a state of crisis need to take steps to ensure revenues keep coming in. Repeat
business and eventually loyalty is critical in that respect. This paper considers airline customer
buying behaviour and preferences and post-flight attitudes with a view to identifying what
makes the air travel product more appealing to customers. A structured web-based purpose-
designed instrument was used to collect travel and traveller data and perceptions and the
findings are prioritised using a consensus decision making approach. A major issue is identified
and a two-step model is developed to demonstrate its significance for airline customers’ loyalty.
Specific airline customer communications and other priorities and preferences are identified. In
addition, it is found that airline customers have a number of complaints, many of which are not
communicated to the airline, and it is demonstrated that these complaints impact on loyalty. It is
therefore argued that airlines and researchers alike need to study customer complaints in
relation to loyalty in greater detail. In view of both the economic situation in Greece and airline
deregulation, this research is of significance to Greek airlines seeking to “tie” their customers
and maintain market presence.
Keywords: Airline Customers, Satisfaction, Complaints, Loyalty
1
Amalia Polydoropoulou is a Professor at the Department of Shipping Trade and Transport of the
University of the Aegean. She received a Civil Engineer Diploma from NTUA, and a MSc. and PhD
in Transportation Systems and Decision Sciences from MIT. Her areas of research include transportation
systems analysis, transportation demand modeling, market research and econometrics. Contact details:
polydor@aegean.gr
2
Evangelia Chortatsiani holds a lectureship at the Department of Shipping, Trade and Transport of the
University of the Aegean. She is a services economist, graduate of the Aristotle University of Thessaloniki,
holds an MSc from the London School of Economics and a PhD in Marketing and I nnovation Management
from I mperial College. She specializes in product development, costing and marketing, competition and
market regulation in telecommunications, transport, shipping and financial services. Contact details:
echor@aegean.gr
3
Maria Kamargianni is a PhD candidate at the Department of Shipping Trade and Transport of the
University of the Aegean. She received a Diploma in Business Administration from University of
Macedonia, and a MSc. in Shipping, Trade and Transport from STT. Her areas of research include travel
behavior, transportation demand modeling and sustainable transport systems analysis. Contact details:
kamargianni@aegean.gr
1. I NTRODUCTI ON
For a great deal of sectors, “crisis” is interpreted as a sharp decline in GDP or indeed spending
specific to the sector. A “crisis” may be seen as the end of a business cycle, provided this is
signalled by some prominent event or events. Depending on sector among others, it can be
overcome in due course, aided by such initiatives as Keynesian approaches (Keynes, 1936;
Woodford, 1999), or other government or corporate efforts to re-establish equilibrium.
For the aviation world and commercial airlines in particular, crisis may mean a sharp change in
revenue or cost streams or volatility that may last from a few weeks to more than a year. These
can have a multitude of originating factors. Reduced load factors of some duration, due to a
broader economic “crisis” or downturn (with their impact exacerbated by the significant fixed
cost element of aircraft leases, trained personnel payroll, slots rental and the like), fuel price
volatility, industrial action, airspace disruptions (such as those that resulted from the Icelandic
volcano eruption in 2010 or the World Trade Centre attacks in 2001) can all lead to a “crisis”.
A company’s typical response to a crisis is to minimise costs, followed by some efforts to

maintain revenues. For airlines, this is where loyalty comes in. On the critical assumption that
passengers have a real choice of carrier, and given that occasional travellers have been found to
be significantly more price sensitive (Gomez et al., 2006), airlines need to strive to maintain the
regular patronage of regular flyers. Indeed, to this precise end they have almost invariably
devised some incentive programme and most have now joined an international alliance one of
the primary purposes of which is to extend the reach of such programmes.
This paper builds upon the multi-step airline choice (Suzuki, 2007) and traveller perception,
satisfaction and loyalty models (Gomez et al. 2006; Oyewole, 2001; Hess et al., 2007; Espino et
al., 2008) literature to explore the real issues of concern to travellers. It is argued that there is a
significant wealth of information on matters of interest to travellers (delays, luggage handling
etc.) that can affect re-purchase and loyalty. This takes the shape of complaint material and
appears to be understudied in terms of utilizing models for traveller behaviour. One or a series
of specific negative travel experiences, it is argued, can act as a disincentive to repurchase or
recommend an airline that can easily counterbalance costly positive actions such as price
discounts or incentive programmes.

Section 2 considers the study's design, section 3 presents key passenger buying habits and
perceptions, section 4 builds a customer loyalty model to demonstrate the concepts discussed
and section 5 presents the study's conclusions and implications.
2. RESEARCH METHODOLOGY
To collect the data, a structured web-based purpose-designed instrument was used. It begins by
collecting some general information about respondent's general profile and long distance travel
behavior. The second questionnaire section asks about the airline travel offering, as perceived
by the traveller, including her involvement in its production/consumption. The third section
focuses on the respondent's last trip. This helps to sharpen information and perception
recollections, an approach used extensively in the transportation literature (Ben-Akiva et al.,
2002). The fourth section considers matters to do with the current economic crises. The last
section collects socioeconomic control data identified by the literature as pertinent to air travel
analysis (Dolan et al., 2006; Polydoropoulou et al. 2010). Most questions were either multiple
choice or answerable on a 7-point scale (from -3: not at all to + 3: completely).
The data was collected in May 2011 and 188 responses were received, essentially from within
Greece. As can be seen from Table 1, a reasonably stratified sample was achieved according to
a number of dimensions, except for a higher than average number of student traveller cases
(48% ). These are not necessarily all from the authors' research location base (i.e. the island of
Chios) since only 8.5% of reported “last trips” landing or departing at Chios. Given sufficient
controls, these were all maintained in the sample. With an average of 8 trips in the last year,
respondents generally appear to be experienced air travellers, although most of them (62%) are
not Frequent Flyer Programme (FFP) (incentive programme) members.
In the exploratory section of the study, tabulation based diagrams were used to aid concept
presentation. To prioritise the issues to consider, a consensus decision making approach was
employed, with people combining transport research skills, air traveller profiles and general
business acumen singling-out those results that appear below. Tests were subsequently carried
out to discern the robustness of results where named factors were ranked and the significance
of any differences to the values attached to each studied.
In the explanatory section of the study, a two-step model is used to demonstrate factor
interaction.

Table 1: Key Sample Characteristics
Percent Percent
( N. obs. 18 8) ( N. obs. 1 88)
Gender Male 55% Complaints Yes 13%
Female 45% No 87%
Age 18-26 54% Satisfaction 1 (Dissatisfied) 10%
27-40 23% 2 (Quite satisfied) 19%
More than 41 23% 3 (Satisfied) 26%
Educational level Bachelor 66% 4 (Very satisfied) 45%
Master 27% Air Company Aegean 40%
( last trip)
Doctoral 7% Olympic 27%
Occupation Public Services 10% Other 33%
Trips in the 8 trips
Private sector 30%
last 1 2 months on average
Businessperson 12% FFP Yes 38%
membership
Student 48% No 62%
Monthly family < €1000 22% Monthly family
income income ( cont.)
€1000-€2000 16% €3000-€4000 18%
€2000-€3000 18% More than €4000 26%
3. DATA ANALYSI S
3.1 TRAVEL BEHAVI OR AND ECONOMI C CRI SI S
From the airline customers' views, the current economic crisis (in Greece) has affected both the
demand for air travel and the price sensitivity of travellers. Experienced travellers (incentive
programme members) exhibit those behaviours marginally less, as they report tougher time
constraints and a lower transport mode substitution (Figure 1).
3.2 BOOKI NG
The front end of the air travel business has clearly moved to the e-business era. Some 94% of
respondents use the web to inform themselves about available flights and fares while 53% use
it to book a flight (Figures 2, 3). Interestingly, it is not young age that makes the “modern”
traveller, with on-line buyers being on average older (at 33 years) than traditional buyers (at
29.5 years).

Figure 1: Attitudes about air trips during economic crisis
To the extent that e-booking can reduce airline costs and that extending it does not lead to
reduced availability of more traditional ways of booking for those that need them, there is still
ample scope for increasing the take-up of e-booking.
Figure 2: I nformation about air tickets
Subsequent research can look into ways of encouraging this, including dealing with price
incentives, differential approaches to cancellations and ticket modifications, trust and confidence
issues with regard to e-payments and other similar parameters (Papola and Polydoropoulou,
1996).

Figure 3: Air ticket booking
3.3 AI RLI NE ADVERTI SEMENT AND I NFORMATI ON POLI CY
Flyers feel that traditional advertising is inconsequential for services pertaining to a specific
journey. Any impact should therefore probably be sought in brand profile building (Figure 4).
Figure 4: Advertisements' effects
Passengers have clear and specific preferences on advertising communications. They wish to
receive ticket offer information by email (75% overall and 90% for incentive programme
members) and incentive programme information on their mobile telephone (88%) (Figures 5, 6).
Given that incentive programme information is generally not more urgent or critical than ticket
offers, an explanation might be that programme members are prepared to allow a more
“intimate” or close communication with their airline. Carefully designed such communications
can logically be expected to positively impact loyalty among others.

Figure 5: Flyer’s willingness to receive information
Figure 6: FFP member’s willingness to receive advertisements
3.4 CRI TERI A FOR SELECTI NG AN AI RLI NE TO FLY
Air travellers consider multiple factors each time they select an airline carrier (Polydoropoulou et
al., 2007; Hess et al., 2007; Nako, 1992).
Interestingly, this research shows that incentive programme membership is ranked lower than
even environmental footprint. This is likely the result of a moderately experienced (in terms of
average trips per year or incentive programme membership) air travel study sample, as
described in section 2, and the recent redesign of major Greek airlines' incentive programmes to
essentially reward only truly regular or premium class flyers, or in non-financial ways (eg. lounge
access). If it takes some 25 low cost fares to accumulate incentive mileage sufficient for a 26th
journey (equivalent to a 4% discount) and then fees and charges are still due, the discount may
fall to under 3% or even 2% in some cases. Hardly an incentive to stick to a carrier and allowing

all sorts of factors such as specific ticket price, punctuality etc. to make the difference for a
broad range of passengers.
3.5 SATI SFACTI ON FROM AI RLI NE SERVI CES AND COMPLAI NTS
Passenger satisfaction is a compound concept (Figure 7). There exists a very interesting
combination of generally reported satisfaction (Figure 8) and a strong stream of airline reported
and non-reported complaints (Figure 9).
Figure 7: Ranking of factors affecting satisfaction
Airlines are apparently leaving flyers content with overall service provision, the highest scoring
factors being booking, welcoming on board and the courtesy of stewardesses, and the lowest
scoring being price, food & beverages and on-board entertainment. The booking process, in
particular, is an excellent example of how the sector can push forward on matters it sets its
mind to (en bloc, as Ott (1993) found). A key component is automated reservation systems,
which was one of the early tools airlines used to (perhaps forcefully) enhance customer loyalty
with the onset of deregulation in Europe and elsewhere (Lee et al., 1996). This was seamlessly
evolved to the e-business era with the advent of (cost reducing) e-booking and more recently
and for limited circumstances, e-check in. Another example is the development of (the now
omni-present) incentive programmes (Clemes et al., 2008; Miller, 1993). In both cases, the
initiatives were taken in the early 1990s in response to deregulation and rising competition, so

one could anticipate that an increase in (market) pressures of various sorts might drive the
development of further satisfaction and loyalty seeking initiatives.
Figure 8: Flyer’s satisfaction from airline services
Figure 9: Emphasis of complaints (% in terms of passengers who have complaints)
The arguably unfinished state of this process then, perhaps so ad-infinitum, or indeed
management or other issues, may be among the reasons why, at the same time, travellers have
a significant number of concerns and complaints. Indeed, some 13% of respondents reported a

complaint about their last flight, this percentage obviously rising significantly if, say, data had
been collected for the last 5 or 10 flights. Slightly more than half of those who had a complaint
(55% ) were experienced travellers (incentive programme members). Most interestingly, the
majority of passengers with complaints (61%) did not bother to report them to the airline.
Customer complaints are personal. They can be voiced to company staff or even put on paper.
They are also by their nature specific, much more so that “low satisfaction” is. This means they
are to easier understand and, more importantly, set about to dealing with them, although as the
data collected during this study suggests, they remain largely (61%) unreported and therefore
outside the sphere of airline knowledge.
It was therefore decided to take this finding one step further and consider whether complaints
affect customer loyalty. There is significant academic scope for research in this area, with
current research carefully mapping complaints without linking them to subsequent subject
behaviours (eg. Özlem, 2007).
4. A PRELI MI NARY PARAMETRI C MODEL TO PREDI CT CUSTOMER LOYALTY
This section presents a preliminary model developed for predicting a customer loyalty. A loyalty
index is created as the outcome of a two-step process. In the first step a factor analytic model is
estimated to produce a factor from two variables namely customer repurchase and
recommendation to others. In the second step the fitted values of the loyalty index is used as
the dependent variable of a regression model. Independent variables are travel characteristics
(chosen airline company), satisfaction level (overall satisfaction, complaints) and socioeconomic
characteristics (income).
The resulting model is as follows:

LOYALTY = -1.860 + 0.459* Satisfaction - 0.567* Complaints – 0.329* I ncome
+ 0.612* Aegean + 0.374* Olympic
Table 2 presents the model's estimation results. The model's overall fit is moderate. However,
the coefficients are statistically significant at a 95% level of confidence, while their signs are
intuitively correct. Flyer satisfaction and complaints impact loyalty as expected (Fornell et al.,
1996; Spreng, et al., 1995; Garow, 2010). “High” income has a negative effect on loyalty,
possibly due to reduced time flexibility (time of flight determines airline), a different booking
process (eg. through a secretary) or other factors. The airline each passenger last flew also re-

enforces loyalty provided it was a major one, perhaps on the basis of “once you will fly with us,
you will always fly with us” and/ or limited alternatives (for the 21 islands with airports, traveller
choices are essentially a sub-hour flight or a multi-hour boat trip (Kitrinou et at., 2010;
Polydoropoulou et al., 2011) while the market is big enough to support only a very limited
number of airlines). In general, Aegean customers seem to be more loyal than Olympic Air
customers compared to customers of all other airlines.
Table 2: Model Estimation Results
Coefficient
Coefficient Names t-test
Estimates
Satisfaction (overall satisfaction)
0.459 8.001
(7pt Likert Scale: -3= completely dissatisfied,..,+ 3= completely satisfied)
Complaints
-.567 -3.155
(1= yes, 0= o/ w)
I ncome (Monthly Family Income more than €4000)
-0.329 -2.64
(1= yes, 0= o/w)
Aegean (passengers who travelled with Aegean at their last trip)
0.612 4.566
(1= yes, 0= o/w)
Olympic (passengers who travelled with Olympic at their last trip)
0.374 2.495
(1= yes, 0= o/w)
Constant -1.860 -4.638
Statistics
Number of Observations 188
R-square 0.403
Adjusted R-square 0.387
Although the sample used in this research is not extensive, on the basis of the above there is
evidence to suggest that customer complaints do have a predictive value for airline passenger
behavioural loyalty, as satisfaction does. Moreover, the results obtained create several avenues
for future research.
5. CONCLUSI ONS
Airlines in a state of crisis need to take steps to ensure revenues keep coming in. Repeat
business and eventually loyalty is critical in that respect. This paper considers airline customer
buying behaviour and preferences and post-flight attitudes. Specific communications and other
priorities and preferences are identified.
Analysis shows that the current economic crisis has affected both the demand for air travel and
the price sensitivity of travellers. However, experienced travellers and incentive programme
members exhibit those behaviours marginally less, due to tougher time constraints and limited
travel mode alternatives.

Nowadays passengers seem to be familiar with the e-services that airline companies offer. The
vast majority of the participants use the internet to be informed about flights and fares, while
they are willing to receive ticket offer information by email and incentive programme information
on their mobile telephone. I nteresting is the fact that the participants who use e-booking and e-
services are older than traditional buyers. Airlines should encourage more of their passengers to
become members of incentive programmes in order to achieve a more close communication
with them.
Concluding, it is found that airline customers have many complaints, many of which are not
communicated to the airline, and it is demonstrated that these complaints impact on loyalty. It is
therefore argued that airlines and researchers alike need to study customer complaints in
relation to loyalty in greater detail. Such efforts may benefit from methodological approaches
segmenting customers and/or providing value-added propositions based on the Delta Model
(Hax et al., 2001) or the Rhombus Model (Litinas et al., 2010).
REFERENCES
‚ Anderson, E., Fornell, C., and Lehmann, D.R. (1994). Customer satisfaction, Market share,
and Profitability: Findings from Sweden. Journal of Marketing , Vol. 58 (July), pp. 53-66.
‚ Athanassopoulos, A.D. (2000). Customer Satisfaction Cues to Support Market Segmentation
and Explain Switching Behavior. Journal of Business Research, Vol. 47 (3), pp. 191-207.
‚ Bailey, E. R., Graham, D.R. and Kaplan, D. (1985). Deregulating the Airlines. MIT Press
Series on Regulation of Economic Activity, Cambridge, MA: MIT Press.
‚ Ben-Akiva, M., Walker, J., Bernardino, A.T., Gopinath, D.A., Morikawa, T., and
Polydoropoulou, A. (2002). Integration of Choice and Latent Variable Models. In Perpetual
Motion: Travel Behaviour Research Opportunities and Application Challenges. Editor: Hani
Mahmassani. Publisher: Elsevier Science, May 2002.
‚ Boshoff, C., 2005. A re-assessment and refinement of RECOVSAT. Managing Service Quality,
15, pp. 410-425.
‚ Clemes, M.D., Gan, C., Kao, T. H., and Choong, M. (2008). An empirical analysis of customer
satisfaction in international air travel. Innovative Marketing, Vol. 4, (2), pp. 49-62.
‚ Clemes, M.D., Ozanne, L.K., and Laurenson, W.L. (2001). Patients’ Perceptions of Service
Quality Dimensions: An Empirical Examination of Health Care in New Zealand. Health
Marketing Quarterly, Vol. 19 (1), pp. 3-22.

‚ Cronin, J.J. and Taylor, S.A. (1992). Measuring service quality: a re-examination and
extension. Journal of Marketing , Vol. 56, July, pp. 55-68.
‚ Dennet, C.E., Ineson, M., Stone, G.J., and Colgate, M. (2000). Pre-Bookable Services in the
Chartered Airline Industry: Increasing Satisfaction through Differentiation. The Service
Industries Journal, April, Vol. 20 (2), pp. 82-94.
‚ Dolan, P. and White. M. (2006). Dynamic well-being: connecting indicators of what people
anticipate with indicators of what they experience. Social Indicators Research, 75, pp. 303–
333.
‚ Dolnicar,S., Grabler, K., Grün, B., Kulnig, A. (2011). Key drivers of airline loyalty. Tourism
Management , 32 (2011) 1020-1026.
‚ Duarte, A., Garcia, C., Limão, S. Polydoropoulou, A., (2009). Experienced and Expected
Happiness in Transport Mode Decision Making Process. CD Proceedings of the 88th Annual
Meeting of the Transportation Research Board, Washington DC, 2009.
‚ Espino, R., Martin, J. C., & Roman, C. (2008). Analyzing the effect of preference
heterogeneity on willingness to pay for improving service quality in an airline choice context.
Transportation Research, Part E, 44, 593-606.
‚ Fornell, C. Johnson, M.D., Anderson, E.W., Char, J., and Bryant, B.E. (1996). The American
Customer Satisfaction Index: Nature, Purpose and Findings. Journal of Marketing , Vol. 60
(October), pp. 7-18.
‚ Garow, L.A. (2010). Discrete choice modelling and air travel demand. Ashgate Publishing
Limited, ISBN: 978-0-7546-7051-3, England and USA.
‚ Gilbert, D., Wong, K.C. (2002). Passenger expectations and airline services: a Hong Kong
based study. Tourism Management , Vol. 24, April, pp. 519-532.
‚ Gomez, B. G., Arranz, A. G., & Cillan, J. G. (2006). The role of loyalty programs in
behavioural and affective loyalty. Journal of Consumer Marketing, 23(7).
‚ Hax, A., and Wilde, D. (2001). The Delta Project . Palgrave, ISBN 978-1-4419-1479-8,
Cambridge MA, USA.
‚ Hess, S., Adler, T., Polak, J. W. (2007). Modelling airport and airline choice behaviour with
the use of stated preference survey data. Transportation Research, Part E, 43, pp. 221-233.
‚ Jones, T.O., Sasser, W.E., (1995). Why satisfied customers defect. Harvard Business Review,
Vol. 73, pp. 88-99.
‚ Keynes, J.M. (1936). The General Theory of Employment, Interest and Money. UK: Palgrave
Macmillan.
‚ Kiran Karande, K., Magnini, V.P., and Tam, L. (2007). Recovery Voice and Satisfaction After
Service Failure An Experimental Investigation of Mediating and Moderating Factors. Journal
of Service Research, Volume 10, No. 2, pp. 187-203.
‚ Kitrinou, E., Polydoropoulou A., and Bolduc D., (2010). Development if Integrated Choice
and Latent Variable Models for the Residential Relocations Decision in Island Areas. Choice
Modelling: The State of the art and the state of practice. Proceedings from the Inaugural
International Choice Modeling Conference. Chapter 27, pp 593-618. Edited by Stephane
Hess and Andrew Daly. 2010 Emerald Group Publishing Limited 2010. ISBN: 978-1-84950-
772-1.
‚ Lee, M., and Cunningham, L.F. (1996). Customer Loyalty in the Airline Industry.
Transportation Quarterly, Vol. 50, No 2, Spring, pp. 57-72.
‚ Levin, M. E. (1987). Airline Competition in Deregulated Markets: Theory, Firm Strategy, and
Public Policy. The Yale Journal on Regulation , Vol. 4, 2 (Spring), pp. 393-494.
‚ Litinas, N., Kapros, S., Polydoropoulou, A. (2010). Change of Values and Transportation
Needs: Towards a New Transportation Strategic Planning Methodology. Chapter 6 Applied
Transport Economics - A Management and Policy Perspective. Eddy Van de Voorde and
Thierry Vanelslander. De Boeck Publishing, http:// hoger.deboeck.com/ titres?id= 80905_1. In
print.
‚ McDougall, G.H.G., Levesque, T., (2000). Customer satisfaction with services: putting
perceived value into the equation. Journal of Service Marketing. Vol. 14, pp. 340-392.
‚ Miller, C. (1993). Rewards for the Best Customers: Loyalty Programs Increase as Marketers
try to Build Relationships. Marketing News (July 5), pp. 1-6.
‚ Nako, S. M. (1992). Frequent flyer programs and business travellers: an empirical
investigation. Logistics and Transportation Review, 28, pp. 395-414.
‚ Oliver, R.L. (1997). Satisfaction: A Behavioral Perspective on the Consumer . Boston:
McGraw-Hill/Irwin.
‚ Ostrowski, R.L., O’Brien T.V., and Gordon G.L. (1993). Service quality and customer loyalty
in the commercial airline industry. Journal of Travel Research , Vol. 32 (2), pp. 16-24.
‚ Ott, J. (1993). Airline Customer Service Rated ‘Average’ in Survey. Aviation Week and Space
Technology, February 1, pp. 31.
‚ Oyewole, P. (2001). Consumer's Socio-Demographic Characteristics and Satisfaction with
Services in the Airline Industry. Services Marketing Quarterly, 23: 2, pp. 61-80.
‚ Özlem, A., (2007). Customer complaints about airline service: a preliminary study of Turkish
frequent flyers. Management Research News, Vol. 30 I ss: 6, pp.409 – 419.
‚ Papola, A., and Polydoropoulou, A. (2006). Shopping-Related Travel in an Rich ICT Era: A
Case Study on the Impact of e-Shopping on Travel Demand. Paper Presented at the 85th
TRB Meeting, January 22nd-26th, Washington DC, 2006.

‚ Polydoropoulou, A., Litinas N. Kamargianni M. (2010). Airline Passengers’ Expectations and
Evaluation of Airline Services. Paper presented at the 1st Conference “Airtransport of Today
and Tomorrow” of Hellenic Aviation Society.
‚ Ramamurti, R., Sarathy, R. (1997). Deregulation and Globalization of Airlines. The
International Trade Journal, 11 (3), pp. 389-432.
‚ Spreng, R.A., Gilberd D. and Mackoy R. D. (1995). Service Recovery: Impact on Satisfaction
and Intentions. Journal of Services Marketing, Vol. 9(1), pp. 15-23.
‚ Suzuki, Y. (2007). Modeling and testing the “two-step” decision process of travellers in
airport and airline choices. Transportation Research, Part E, pp. 43.
‚ Yu-Wei, Ch. & Yu-Hern Ch. (2010). Does service recovery affect satisfaction & customer
loyalty? An empirical study of airline services. Journal of Air Transport Management 16, pp.
340-42.
‚ Woodford, M. (1999). Revolution and evolution in twentieth century macroeconomics.
Frontiers of the mind in the twenty-first century conference, Washington.
‚ Zeithaml, V.A. and Bitner, M.J. (1996). Services Marketing. 1st ed. New York: McGraw-Hill.

A COMPLEX SYSTEM APPROACH I N MODELI NG AI RSPACE CONGESTI ON DYNAMI CS
Soufian Ben Amor 1

PRISM Lab, University of Versailles Saint-Quentin-en-Yveline
Marc Bui2
Complex Systems Modelling and Cognition, Eurocontrol and EPHE Joint Research Lab
ABSTRACT
The dynamical behaviour of groups of airspace sectors, e.g. Functional Airspace Blocks, is
not trivial to be analysed without appropriate theoretical tools. In this paper, we suggest a
discrete model based on cellular automata and multi agent systems to express the
congestion dynamics and complexity in the controlled airspace. Discrete time simulations
have been performed with random selected scenarios of traffic and with independent sector
parameters to investigate the impact of availability of local sectors on the whole state of the
airspace. Obtained results show the existence of a traffic threshold that leads to a theoretical
saturation of airspace. The test scenario showed a phase transition phenomenon towards
the congestion of the European airspace at the resulting traffic threshold circa 50 000 flights.
Validation using real data shows the predictive abilities of the model.
Keywords: Air Traffic Management, Air Traffic Control, Complex Systems, Airspace
availability, complex networks, Multi Agent Systems, cellular automata.
"""""""""""""""""""""""""""""""""""""""" """"""""""""""""
1
Soufian Ben Amor received his PhD in Computer Science in 2008 from Ecole Pratique des Hautes
Etudes. His PhD was performed at EUROCONTROL (the European Organisation for the Safet y of Air
Navigation). He is currently Assistant Professor at University of Versailles Saint-Quentin-en-Yvelines.
He is author and co-author of more than 27 papers and conference papers on the topics complex
systems modeling and operation research. Contact Details: soufian.ben-amor@uvsq.fr
2
Marc Bui received his PhD in Computer Science in 1989 from University Paris 11. He is full
professor of Computer Science at University of Paris 8 and at the Ecole Pratique des Hautes Etudes,
where he is the head of a research team, the CSMC. Prof. Marc Bui is co-editor for Studia Informatica
Universalis journal. He has chaired and/ or served on the advisory boards and program committees of
various international conferences and workshops. He is author and co-author of more than 60 papers
and conference papers on the topics of distributed systems and complex systems modelling. Contact
Details : marc.bui@ephe.sorbonne.fr
"
1. I NTRODUCTI ON
Most of new policies trying to improve the Air Traffic Management (ATM) system tend to
maintain its classical structure. Innovation in this field is basically focused on equipment,
information and communication technologies, task automatization and improvement of
Human-machine interfaces [Watkins et al., 2002] . Nevertheless, different actors of airspace
traffic consider that the current system has attained its limits and congestion is more and
more difficult to resorb. Empirical studies show that more and more network-effects are
observed in the operational context demonstrating the qualitative changes in the airspace
availability ([Mayer et al., 2003], [Daniel, 1995], [ Brueckner et al., 1992]).
Delay cost is evaluated to be between 7 and 11 billion euros per year and according to the
I nstitute of Air Transportation (I AT) 60% is due to ATC (Air Traffic Control) [ITA, 2000]. The
ATM system is composed of numerous processes and various actors having different and
divergent objectives: pilots try to be on time, companies focus on economic aspects
(reducing costs and maximizing benefits) and controllers must guarantee the security of the
traffic. In our approach, even if we tried to include indirectly some ATM aspects, we are
basically concerned with the ATC subsystem, in particular the en route control. In fact, en
route control in Europe is the main responsible of traffic delays leading to costs of several
billion euros per year [Golaszewsk, 2002]. This is not the case in the USA where delays are
caused by the airport saturation.
Air Traffic Management (ATM) can be modelled as a set of components of different

subsystems in mutual interactions in order to accomplish the mission of simultaneously
maintaining safety and sustaining growth. The ATM system is considered as a complex
system because its behaviour depends on a complex combination of various sub-systems
performing complicated functions. The evaluation of the impact of each function on the
overall ATM system cannot be performed unless a specific approach is used. Understanding
the mechanisms by which complexity may be reduced in the particular domain of ATM may
provide important solutions to optimize the dynamics of the system and its structure.
2. OVERVI EW OF THE ATM SYSTEM

The Air Traffic Management system is a complex network composed of several
heterogeneous and mutually interacting subsystems. The complexity of the ATM can be
"
related to the following factors: system size, diversity of users, safety constraints and
uncertainty (weather, human factor, technical factor...). This complexity can be also related
to the Air Traffic Control (ATC) subsystem representing the rigidly structured air space and
the largely centralised, human operated control hierarchy ([Delahaye et al., 2005], [Histon et
al., 2002]). ATC, in which we observe complex phenomena, is composed of services
provided by the controllers on the ground to ensure the safety and the efficiency of aircraft’s
motion, and are provided throughout the controlled sectors. I n fact, aircraft tend to fly along
fixed corridors and at specific altitudes, depending on their route. The entire path of the
aircraft is pre-planned (flight plan) and only minor changes are permitted online. The ATC is
in complete command of the air traffic and ultimately responsible for safety. All requests by
the aircraft have to be cleared by the ATC.
Airspace is composed of controlled airspace and uncontrolled airspace. A controlled airspace

is a set of controlled sectors, each of which being associated to a team of air traffic
controllers. These air traffic controllers are persons who operate the air traffic control system
to expedite and maintain a safe and orderly flow of air traffic, and help prevent mid-air
collisions. They apply separation rules to keep each aircraft apart from others in their area of
responsibility and move all aircraft efficiently through “their” airspace and on to the next.
[Tran Dac, 2004]
Aircraft follow a planned trajectory to join two airports. They are monitored and guided
throughout the whole flight time by air traffic controllers. Computers, communication links
and radar screens all provide up-to-date information. Technology quite often has not one but
two back-up systems to cover any possible breakdowns. The whole organisation is based
upon international regulations and determined routines. During the flight different services
are furnished by three kinds of control activities: Tower Control where controllers direct
aircraft that are taking off or landing at airports, Approach Control where controllers handle
aircraft that are transitioning from the en-route portion of flight into the airspace around or
near an airport and En-route Control where controllers handle aircraft that are operating on
the main travel portion of their flights, typically at a high altitude.
As specified in the introduction, in this work, we are especially interested in the study of the
behaviour of the en route sectors because, in Europe, en route control is the main
responsible for the airline delays and traffic congestion. We include the effects of approach
"
and tower control sectors using a non-local interaction rules. As we can see in Figure 1, the
initial traffic demand exceeds the declared capacity of a sector at a specific moment. The
Central Flow Management Unit (CFMU), an operational unit of EUROCONTROL3, manages air
traffic demand in order to avoid airspace congestion due to this difference and to optimize
the utilization of resources. Despite the regulation and the planning made by the CFMU there
are always differences between planned and real traffic. These differences are specific to
each sector and the resulting effect of these local states and their interactions between them
on the whole availability of airspace are difficult to determine.
By considering the densely interconnected system of ATM as a network where components

properties are heterogeneous and individual and by applying appropriate theories we are
able to model the emergence of global properties in the system from the local behaviour of
its component (typically the availability and congestion of the airspace). By this way we can
take into account the coordination requirements representing the interactions between
controllers in adjacent sectors, which is an important factor in ATC complexity [Histon et al.,
2002]. It is also important to note that these interactions are closely dependent of the
airspace design. I n fact the topological structure of airspace defines the structure of the
sectors coordination network. The space-time analysis that we propose in this paper is a
general approach focusing on the intricate relation between these two fundamental aspects.
Figure 1 Real versus planned traffic in a congested sector
"""""""""""""""""""""""""""""""""""""""" """"""""""""""""
3
EUROCONTROL is the European Organization for the Safety of Air Navigation
"
3. BOTTOM- UP MODELI NG OF THE ATM SYSTEM
ATM simulation requires a modelling approach and simulation framework taking into account
particularities and properties of this system. ATM being a complex system where the
objective is to guarantee the security and fluidity of the traffic by optimizing the use of the
shared resources between different actors having divergent constraints (companies, air
traffic controllers, pilots, passengers...) needs to be studied using appropriate Tools
[Boccara, 2004].
In various natural and artificial contexts, we observe phenomena of high complexity.

However, research in physics, biology and in other scientific fields showed that the
elementary components of complex systems are quite simple. It became crucial for scientific
research dealing with complex systems to determine the mathematical mechanisms to
understand how a certain number of such elementary components, acting together, can
produce the complex behaviours observed in these systems.
Cellular automata studied by Stephen Wolfram [Wolfram, 1984] represent an attempt to

design the simplest mathematical model able to generate a high complexity. One of the most
important current problems consists in finding general laws being able to be applied to study
the majority of complex systems. A cellular automaton is, in the simplest case, one line made
up of empty boxes. Each box carries one value 0 or 1. Thus, the system configurations are
an ordered sequence of 0 and 1 evolving over time. At each time step, the value of each site
is updated according to a specific rule. The rule depends on the value of a cell, and of its
two closest neighbours.
According to Wolfram [Wolfram, 1986], [ Wolfram, 1994] Cellular Automata (CA) are
microscopic models for complex natural systems containing large numbers of simple identical
components with local interactions. Even if the construction of the cellular automata is very
simple, their behaviour can be very complex [Wolfram, 1994], [Wolfram, 2002]. There are
fundamental reasons showing that there is no general method which can universally be
applied to predict the behaviour of these systems. Compared with reality the cellular
automata appear simplistic. However, they are currently considered as a fundamental tool in
modelling and simulating complex phenomena, in particular concerning the auto-organized
systems. The use of the cellular automaton makes it possible to reduce the complexity of
modelling to what is necessary to generate the phenomenon. It is a paradox of complex
"
systems: the behaviour of the system is unpredictable and complex (at a long term level)
whereas the laws (or rules) which controlling it are simple and deterministic. Moreover,
cellular automata represent a powerful simulation tool. In fact a convincing simulation of
large dataset requires computing power of parallel computers. However, the local nature of
interactions between cells makes the programming of cellular automata easy “to be
parallelized”. The dynamic theory of systems was developed to describe the global properties
of the solutions of equations.
A combination of Cellular Automata (CA) formalism and Multi Agent Systems (MAS) allows a
coherent mathematical and computational representation of the physical model. In fact, CA
permit the representation of the entities composing the system and the evolution of their
state over time whereas MAS are well adapted to express the interactions between the
entities and their behaviour [ Weiss, 1999] , [ Fikes, 1982] . We will use in our model this
combination to rebuild the real system from the basic components.
a. WHY CA AND MAS IN ATM MODELING?

The interest in combining MAS and CA is to introduce the mobility of the components
representing the cells of CA. These components are called agents in MAS and they are able
to move, communicate, transmit information, take decisions and influence their environment.
Generally, agents are used in social sciences to represent individual or collective decisions in
a population and more generally they have socio-economical attributes. But multi agent
simulations in social sciences are often non-spatial which is not the case in ATM. In fact, in
ATM modelling we need to represent:
The physical and geographical system (airspace) and its properties:
‚
‚
structure dynamics : evolution of the merging and splitting schemes of the sectors.
routes and sectors topology (shape of the sectors, average number of neighbours,
‚
routes configuration,...).
‚
the technical system : aircrafts, communication systems...
human system : controllers, pilots, ...
In order to provide an efficient and realistic simulation of the ATM behaviour it is important
"
to include the relations between its three basic subsystems. In fact, the components of
technical subsystem (aircrafts) interact with the physical subsystem (sectors) and the human
subsystem (controllers, pilots). The human subsystem is particularly important because it
supervises the two other subsystems in order to accomplish the global mission of the ATM
system: manage the continuous increase of the traffic volume while guaranteeing the
security and the fluidity of the traffic. These aspects can be easily integrated by combining
CA and the multi agent paradigm.
Actually, complexity of ATM (combination of a natural and an artificial complex system) is

such that even a MAS/CA simulation is insufficient to capture all the aspects and specificities
of the system. A realistic representation of ATM needs the representation of hierarchy and
heterogeneity of the different subsystems. Nevertheless, in our work we are especially
interested in observing qualitatively the behaviour of the system while reproducing in a
simplified way its basic mechanisms.
Here is the list of the important aspect which must be considered in the modelling of the
ATM system:
‚ the different kind of entities in the system;
‚ the different hierarchical levels in the system;
‚ the topology of the entities;
‚ the different kind of relations between the entities;
‚ the process determining the state of the entities;
‚ the process determining the changes in their spatial location.
The simulation of the behaviour of such a system needs a rigorous formulation of these
aspects. The simulation of the management of shared resources requires also the integration
of the interaction between the agents and the dynamical resources. A first method that could
be used to represent these interactions emphasize on processes determining interactions
between agents and resources. These agents are cognitive agents having a representation of
the resource and possess their own rules to reach their objectives. Each agent acts on the
resource according to his rules and modify the resource for other agents. In our context we
are facing the problem of the management of renewable common resources (airspace) in
confrontation with different actions and situations which may lead to a satisfactory use (or
not) of the resource for the different agents (pilots, controllers, companies,...).
"
The multi-agents universe offers also an interesting ability to simulate the message
exchange in communication networks (controller/controller communication and
pilot/controller communication). These communications may represent information
exchange, negotiation between the agents (collaborative decision making) or services
exchange.
b. MULTI AGENT MODEL FOR CONGESTION DYNAMICS IN THE ENROUTE AIRSPACE

In the following, the physical model corresponds to the en route part of the controlled
airspace. The mathematical model is represented by the objective correspondence between
sectors and cells of a cellular automaton (Figure 2). The computational model corresponds to
a multi agent system implementing the functions of the cellular automaton. Each sector is an
environment agent integrating the operational rules of air traffic control. Similarly, each
aircraft is a mobile agent following predetermined trajectory and communicating with
sectors.
As seen in section 2, there is a significant difference between the planned traffic and the
realized one [Gwiggner et al., 2006]. This difference leads to the congestion of a certain
number of control sectors. I n order to reduce the congestion and to keep a certain fluidity of
the traffic, the controllers in the saturated sectors may reduce the speed of the aircraft or
deviate from its trajectory to an available control sector. To be able to take into account
these particularities of the ATC system, we integrated these aspects in the rules
implemented in the cellular automaton where cells represent the controlled sectors.
Figure 2: Correspondence betw een sectors and sites
The algorithm for implementing air traffic control rules is as follows:
"
‚ Each sector is an agent modelling its behaviour of air traffic control in the operational
context.
‚ The state of this binary valued agent: 0 if it is available (able to provide control service to
an entering aircraft) and 1 if not (the sector is congested).
‚ An aircraft entering in sector sa at time t is transferred to the following sector sb
Fa Fa
according to the flight plan and the following rules :
– at time t + if sb is available; where is the needed time to cross the
sector sa ;
– if sb is congested at time t + F a , the aircraft is delayed (by decreasing the
speed) by one time unit, then transferred to sector sb if sb is available at
time t + Fa + 1;
‚ Otherwise, the aircraft is rerouted to one of the neighbouring and non-congested sectors
with probability p1 .
‚ An aircraft may be subject to delays other than those imposed by sectors for security
reason. That is why an aircraft may have randomly a delay while arriving to a sector with
probability p2 (this allows to take into account uncertainties related to the
meteorological conditions, take-off delays, ...)
‚ The aircraft may increase its speed if it was already delayed with probability p3 .
c. SIMULATION OF THE MODEL

The model was implemented using Repast (Recursive Porous Agent Simulation Toolkit)
[North et al., 2006]. A part of the controlled airspace, representing the en route part is
modelled using a square grid of size spaceSizeX* spaceSizeY, each corresponds to a sector.
The crossing time of sectors is uniformly distributed in [minCrossingTime,
maxCrossingTime]. The sector status is determined as follows:
‚ a sector s is available at time t if it contains a number of aircrafts inferior to its capacity
Cs ; the maximum number of aircraft that a controller is able to manage simultaneously.
‚ sector capacities are distributed uniformly in [min-Capacity,maxCapacity].
Traffic pattern was generated randomly. The variable nbPairsOrigDest represents the
number of origin-destination couples ( input-cell, output-cell ). The trajectory obtained is a
"
segment having as extremities Orig. cell and Dest. cell. The flight plan generated is
consequently composed of the list of sectors crossed by the segment. In order to take into
account the fluctuation of the traffic during the day we used a particular distribution of
flights where the traffic is doubled in two different time windows.
ÂN
24
Let nbFlights = t be the total number of flight of the day (crossing the studied en route
0
airspace). The distribution of the flights is introduced such that the traffic is doubled during
the intervals (Figure 3): [ t1 = 6h; t2 = 8h] and [ t3 = 18h; t4 = 20h].
Figure 3 The distribution of the daily Traffic
The simulation shows the existence of a phase transition phenomenon concerning the
congestion of the airspace due to a critical density of the traffic [Ben Amor et al., 2007]. For
example, for a given parameterization we notice that while varying the number of flights we
obtain a behaviour of the system totally different when a certain threshold is reached. In
fact, for n < 50,000 (in particular for n = 30000 representing the mean volume of the daily
traffic in Europe) and the other parameters being fixed according to the mean observed
values in the real operational context, we obtain some local congestions that are quickly
resorbed by the collaboration between sectors (Figure 4). When n 50,000, we can identify
a phase transition phenomenon where the system is trapped in a situation where the
congestion propagates through the whole area and local rules are unable to resorb this
congestion (Figure 5). This phenomenon reflects situations where the system needs an
external help to resorb the congestion (delaying take-off at the airports, change the routing
plans for aircrafts on the ground, etc...)
"
Figure 4: The system absorbs the local congestion w hen nbFlights
is inferior to the threshold
Figure 5: Phase transition phenomenon observed w hen the threshold of the

parameter nbFlights is reached
"
4. VALI DATI ON WI TH REAL DATA
In order to validate the model, from the qualitative point of view4, we used the real data
relative to a one-day traffic. To be useful we applied some treatments on these data in order
to gather from different files the needed information implementing the model. We used in
particular the following files:
‚ the file ALL-FT.20070624,traffic file of the CFMU giving the profile of all flight of the day
(June 24th 2007). This file contains among other information the following parameters
for each flight: departure and arrival airport, aircraft identity, company, type of aircraft,
followed route, requested flight level, regulations (in particular rerouting), ATFM sent and
received messages.
‚ the file Airblock.296, environment file of the CFMU giving the set of the elementary
structural units of airspace, and their geographical coordinates.
‚ the file Aircraft.296 giving for each aircraft its identity, type and performance.
‚ the file Airport.296 containing the name, ICAO code and geographical coordinates of all
airports in the world.
‚ the Airspace.296 giving for each airspace entity its identity, name, type and the number
of elementary sectors composing it.
‚ the file Capacity.296 giving the capacity and specifying the type of the concerned
element (control centre, elementary sector, composed sector...) and the time unit.
‚ the file Configuration.296 giving for each control centre its configurations during the day
and the name of the sectors in the different configurations.
‚ the file Flow.296 giving the traffic flows existing between the different airports.
‚ the file NavPoint.296 giving the number of beacons and for each its name, type and
coordinates.
‚ the file OpeningScheme.296 giving the opening schemes (merging and splitting of the
sectors) of different control centres during the day.
‚ the file routes.296 giving the available routes network during the day.
‚ the file Sector.296 giving the number of the elementary sectors and for each its name
and air blocks composing it.
‚ the file TrafficVolume.296 dividing the traffic into traffic volumes and giving
complementary information about the flows.
‚ the file reroutingStats giving statistics about the realized rerouting procedures.
"""""""""""""""""""""""""""""""""""""""" """"""""""""""""
4
Correspondence concerning the simulated scenario and the realized traffic between the shape of the graphs
giving the evolution of the number of congested sectors.
"
‚ the file OverloadHourly giving the total number of hours of congestion where the traffic
exceeded the capacity by 1%, 20%, 40%, 60% et 80%. It provides also the different
kind of regulation procedures realized to resorb the exceeding traffic.
In order to rebuild the realized traffic and represent the evolution of congestion level over
time we need to elaborate a simulation scheme and manage the dependencies between the
data contained in the different files. For example, the evolution of capacity depends on the
opening schemes of the different centres which implies the reading of the file
OpeningScheme and determine the structural entity to which the capacity is applied.
Similarly, to determine the geographical location of a sector we need to open the file
containing the list of airblocks composing it. More generally, the figure 6 gives the global
map of dependencies between the different files. The figure 7, provides a general view of
the simulation interface.
Figure 6 : File Dependencies Map
"
Figure 7: Evolution of the Congestion in the Airspace
The simulation using real data showed a similarity between the shape of the graph giving
the observed congestion level of en route sectors in the European airspace (figure 9) and the
graph given by the simulated scenario of the model using the following parameterization
(figure 8) :
‚ probability of delay of aircraft on take-off : 0,25.
‚ capacity interval : [ 12-20].
‚ probability of regulation using speed : 0,15.
‚ sector crossing time interval : [ 5-15].
‚ rerouting probability : 0,18.
Nevertheless, even if the two graphs (observed and simulated congestion) are similar, they
have mainly two differences:
‚ at the quantitative level, there is an important difference between the observed and
the simulated congestion. By comparing the graph representing the evolution of the
real number of congested sectors to the graph simulated congestion, we notice that
the real congestion level is clearly inferior to the simulated congestion.
‚ the sharpness of certain peaks in the congestion are more important in the simulated
"
congestion.
The difference between the observed congestion and the simulated one is not surprising.
Considering the sensitivity of the ATM system (as any complex system) to the initial
conditions, the quantitative prediction is very hard to establish. The main objective of our
model is essentially oriented to the reconstruction of congestion dynamics (the aspect of the
congestion graph).
This quantitative difference may be explained also by the use of the instantaneous capacity
(number of aircraft simultaneously present in the sector) and we do not integrate the hourly
capacity (amount of the traffic that could be managed by a sector in one hour). Concerning
the small differences in the peaks related the abrupt changes in the number of congested
sectors we can provide these two elements of explanation:
1) the difference is basically due to the difference in the rerouting procedure used in the
model compared to the real procedures. In fact, in the model we considered only the tactical
rerouting but in the real operational context the flow managers using short term predictive
tools are able to display specific online procedures to apply strategic rerouting schemes.
2) other real factors which are difficult to capture in the model may also provide a part of
the explanation of this difference, e.g. the traffic management by controllers. Actually,
controllers do not systematically apply a rerouting scheme when the sector is overloaded. It
was shown by empirical studies that controllers are able to manage sometimes a certain
traffic load which is more important than the declared theoretical capacity.
Despite these differences, our proposed model reproduced dynamics of the congestion which
is very close to the real context. More, it allows testing hypothesis and different scenarios by
varying the simulation parameters. Thus we noticed concordant observations with empirical
studies. In particular, we tested the effects of the variation of the size of the sectors and
noticed that there is a minimal size of sectors under which the propagation is amplified.
"
Figure 8: Simulated congestion - Traffic of June 24th 2007
Figure 9: Observed congestion - Traffic of june 24th 2007
5. CONCLUSI ON AND FUTURE WORK

The systemic approach we used to model the dynamics of availability in ATC showed the
complex nature of the behaviour of the system illustrated by the phase transition
phenomenon which occurs when specific thresholds of key parameters (i.e. number of flights
and crossing time) are reached.
Validation using real data of sectors shows the ability of the model to reproduce congestion
dynamics similar to the real system. In a future work we aim to provide a mathematical
model providing more precise quantitative predictions. To achieve this goal we need, from a
mathematical point of view, to formalize and generalize the neighbourhood concept using
pre-topology theory in order to express different kind of connections between sectors and to
consider a more realistic neighbourhood basis.
"
Although airspace is a common resource, ATM in the European Union is still organised in a
fragmented way. Every time a plane enters the airspace of a Member State, it is serviced by
a different air navigation service provider on the basis of different rules and operational
requirements. In order to improve capacity and efficiency while minimizing costs of air
navigation services, European Member States provided a key mechanism integrated to Single
European Sky (SES) and called Functional Airspace Blocs (FABs). This implies an operational
organisation of airspace independently from country boundaries.
From a managerial and operational point of view, our model showed in particular the interest
of the single sky and Functional Airspace Blocs (FABs) concepts. In fact, according to the
simulations it is clear that a functional and operational segmentation of the controlled
airspace is more efficient to guarantee a performing traffic management, by reducing
conflicts due the heterogeneity of rules and operational requirements.
REFERENCES
‚ Ben Amor S., Tran Dac H., Bui M., & Duong V., (2007) Simulating Dynamic ATM Network
Effects Using Cellular Automata, In Proceedings of the European Conference on Complex
Systems, ECCS’07, Dresden, Germany, October.
‚ Boccara N., (2004) Modeling complex systems. Springer-Verlag, New York.
‚ Brueckner J.K., Nichola J. D., and Pablo T. S., (1992), Fare Determination in Airline Hub-
and-Spoke Networks, Rand Journal of Economics, 23(3), 309-34.
‚ Daniel, J.I., (1995), Congestion Pricing and Capacity of Large Hub Airports: A Bottleneck
Model with Stochastic Queues, Econometrica, 63(2), 327-70.
‚ Delahaye, D., Puechmorel, S. (2005) Air Traffic Complexity Map based on Non Linear
Dynamical Systems. In 4th EUROCONTROL Innovative Research Workshop & Exhibition,
12/ 2005, Brétigny sur Orge, France.
‚ Fikes R.E., (1982) A Commitment-Based Framework for Describing Informal Cooperative
Work. Cognitive Science 6, pp. 331-347.
‚ Golaszewsk R., (2002) Reforming air traffic control: an assessment from the American
perspective, Journal of Air Transport Management, Volume 8, Issue 1, January 2002,
Pages 3-11.
‚ Gwiggner, C., Duong, V., (2006) Averages, Uncertainties and Interpretation in Flow
Planning. Second International Conference on Research in Air Transportation,
"
ICRAT2006, 06/2006, Belgrade.
‚ Histon, J. M., Hansman, R. J, (2002) The Impact of Structure on Cognitive Complexity in
Air Traffic Control . Report No. ICAT-2002-4June 2002, MIT International Center for Air
Transportation, Cambridge, United States.
‚ ITA, (2000) Costs of air transport delay in Europe, Rapport technique, Institut du
Transport Aérien, novembre 2000.
‚ Mayer C., Sinai T., (2003) Network Effects, Congestion Externalities, and Air Traffic
Delays : Or Why Not All Delays Are Evil The American Economic Review, Vol. 93, No. 4
(Nov., 2003), pp. 1194-1215.
‚ North, M.J., Collier, N.T., Vos, J.R., (2006) Experiences Creating Three Implementations
of the Repast Agent Modeling Toolkit . ACM Transactions on Modeling and Computer
Simulation, Vol. 16, I ssue 1, pp. 1-25, ACM, New York, New York, USA.
‚ Tran Dac H. T., (2004) Sectorisation contrainte de l’espace aérien, thèse de Doctorat de
l’Universié de Technologie de Compiègne.
‚ Watkins O. J., Lygeros J., (2002) Safety relevant operational cases in Air Traffic
Management , Report D1.1 HYBRIDGE, IST-2001-32460, Distributed Control and
Stochastic Analysis of Hybrid Systems Supporting Safety Critical Real-Time Systems
Design.
‚ Weiss, G., (1999) Multi-agent Systems, A Modern Approach to Distributed Artificial
Intelligence, The MIT Press, Cambridge, Massachusetts.
‚ Wolfram, S. (1984) Universality and complexity in cellular automata, Physica D, 10.
‚ Wolfram, S. (1986) Theory and Application of Cellular Automata. Reading, MA: Addison-
Wesley.
‚ Wolfram, S. (1994) Cellular Automata and Complexity: Collected Papers. Reading, MA:
Addison Wesley.
‚ Wolfram, S. (2002) A New Kind of Science. Champaign, I L: Wolfram Media.
"
BUSI NESS RELATI ONS BETWEEN THE LOW COST CARRI ERS AND AI RPORTS AS A
CONSEQUENCE OF THE AI R TRANSPORT DEREGULATI ON
Eleftherios D. Katarelos, 1
Department of Shipping Trade and Transport, University of the Aegean, Greece
Iason Koufodontis, 2
Department of Business Administration, University of the Aegean, Greece
ABSTRACT
The deregulation of air transport has increased competition between air carriers, resulting in
lower fares and increased volumes of passengers. Subsequently, the fare reduction has
altered the market structure with the establishment of new carriers, strategic alliances and
mergers, and the bankruptcy of several traditional airlines which were unable to adapt to
the new environment. The emergence of low cost carriers (LCCs) is one important outcome
of the deregulation. LCCs entered the market by offering a differentiated product based on
bare services offered at significantly lower prices. The main target was travelers with
increased sensitivity in pricing and less demand for all-around services. The rise in terms of
passengers and flights dictated a better utilization of the fleet, requiring reduced turnaround
times at airports. Many central airports had very little flexibility and capacity necessary to
facilitate additional timeslots. As an answer to inadequate capacity combined with higher
taxes and fees, most LCCs have chosen to use secondary or regional airports. This choice
has altered the balance and strategic importance between airports and increased their
importance for air carriers. This paper examines the evolution and development of LCCs
globally, along with the consequences of their expansion to the traditional carriers, the
market and the passengers. Emphasis is given to the relationship between LCCs and airports
which has resulted in an additional increase in air travel. The prospects of Greece as a
market for LCCs are also being discussed.
Keywords: Low Cost Carriers, Airports, Greek Air Travel
1
Eleftherios D. Katarelos (corresponding author) is Adjunct Lecturer at the University of the
Aegean and also Air Traffic Controller at the Hellenic Aviation Authority. Contact details:
e.kat arelos@stt.aegean.gr
2
I ason Koufodontis is Adjunct Lecturer at the University of the Aegean. His research focuses in the
utilization of information technologies and air transportation in the fields of tourism and development
in island regions. Contact details: ikouf@aegean.gr
1. I NTRODUCTI ON
The new global regulations have effectively deregulated the air transport market. The
establishment of a competitive environment opened the way for Low Cost Carriers (LCCs) as
a cheaper and simpler alternative model for passengers choosing air travel. Passengers as
consumers have now more choices between different service levels and corresponding
prices. LCCs appeared first in US market and later in Europe and the rest of the world.
Today most countries and regions are served by LCCs together with traditional carriers,
altering the industry structure in a global scale. In order to achieve low cost, LCCs have
formed strategies that either reduce any cost that can be trimmed down or completely
remove parts of their services.
For LCCs, airports were initially a substantial obstacle; costs were too high to suit their
business model and operation from many airports was very limiting, especially in saturated
markets. Airports located in capitals and other maj or cities could not offer capacity at
reasonable pricing, usage period and hours. On the other hand, secondary and regional
airports had both the unused capacity and the willingness to negotiate and offer competitive
low fees. Any cost reduction achieved through better contracts with the airports allows LCCs
to offer lower ticket fares. Lower fares combined with the use of additional airports and the
enlargement of catchment areas has resulted in increasing of the passenger volume. Of
course traditional carriers have also taken measures to remain competitive in the evolving
market conditions. This paper examines the emergence and expansion of LCCs, their effect
on traditional carriers’ strategies, on consumer habits, on airport strategies, and finally their
entrance and presence in the Greek market.
2. AI R TRANSPORT DEREGULATI ON
Since 1944 air transport is regulated by the Chicago Convention. The deregulation started at
national level, first from the USA, followed from Canada, Australia, Japan, Taiwan, South
Korea, and UK. Gradually most countries allowed the operation of new airlines along with
their flag carriers. Deregulation was introduced in order to create “more competitive
aviation services” (Iatrou and Oretti 2007). The main idea was to establish a global aviation
market without entry restrictions. This would allow free and open competition, leading to
more efficient airlines and improved consumer choices (Iatrou and Oretti 2007).

The aviation market has changed radically in the last two decades. The renegotiation of
previous bilateral agreements has placed the previously protected national carriers into a
competitive and turbulent deregulated market. At the same time new strong competitors are
appearing, pressuring airports for even more operational freedom (Delfmann 2005). As the
deregulation allows for more choices and options, it also increases the uncertainty and
reduces the predictability of the environment. In the aviation industry, the airlines were the
first ones to adapt new strategies better suited to a competitive environment, while airports
took much slower steps to meet the new conditions (Delfmann 2005).
3. THE EMERGENCE AND THE CONCEPT OF LOW COST CARRI ERS

During the 1990s LCCs entered and occupied a firm position in many markets. Previously
dominant oligopolies were replaced by open competition (Lawton 2004). Deregulation
encouraged many LCCs to set up extensive networks with scheduled flights (Lei g
Papatheodorou 2010). The expansion of the LCCs is often considered as one of the most
important recent advances in the European aviation (Pels et al 2009). The presence of the
LCCs forced traditional flag carriers to lower their prices and restructure their business (Lei
g Papatheodorou 2010). These actions led to more attractive prices in the whole industry
which in turn increased passenger volumes (Barrett 2004).
The cost strategy adopted by LCCs is based in a simpler service model. Any service that can
be avoided or reduced is not included in the base price of the ticket. Typical examples are
flight with only one seating class (economy), dense seating pitch, limited additional services
during flights, and abandonment of the transfer concept (Pels et al 2009). The choice of
favorite seat, the free newspapers, the baggage handling between carriers, the frequent
flyer rewards, and the dedicated airport lounges are all sacrificed in order to keep cost as
low as possible. Wherever some additional services are still offered they are charged as
extras. The distribution and sales cost is also kept at a minimum by the use of internet
sales, proprietary boarding control, and limited marketing budget. Fleet is typically based on
a single aircraft type allowing for more efficient maintenance and lower operational costs.
The intensive negotiations between LCCs and airports for fees and itineraries are in contrast
with the previously nonexistent competition between airlines and airports (Barrett 2004).

In the 1970s, Southwest became the first LCC in North America and the LCC concept was
developed as an attractive strategy for short haul connections. In the US market where the
competition is increasingly open, LCCs meet favorable conditions to expand. As long as they
keep their operational model simple, they have certain efficiency and cost advantages over
their competitors. While traditional carriers spend money and resources to organize multi-
segment flights, flight seating flexibility, and baggage handling, all LCCs have to do is to
board their customers into a single flight where everyone travels in the same class and has
the same destination.
4. GLOBAL EXPANSI ON OF LOW COST CARRI ERS

In Dublin Ryanair serves 25% of passengers by using only 11% of check-in desks and the
aim is to completely abolish the remaining desks and replace them with self-service
procedures. Each desk can serve annually 130,000 passengers in comparison to only
48.000 served by the desks of traditional carriers. At Stansted airport, where all airlines
operate under the same principles, Ryanair serves over 110,000 passengers at each desk
while the competition only reaches 70 (Barrett 2004). Another indication of Ryanair’s
effectiveness is the number of passengers served by each employee. At Ryanair each front
office worker can check-in 8,000 passengers while the same worker at a national carrier can
only reach 873 passengers on average (Barrett 2004).
Increased levels of competition have led to very low airfares especially for destinations that
are simultaneously served by LCCs and traditional carriers. This observation is also true for
adjacent airports (Lian and Rønnevik 2010). Many flag carriers were not prepared to
compete in a deregulated market and soon they were facing serious problems. Sabena and
Swissair declared bankruptcy in 2001, followed by other national carriers, with most recent
example that of Malev in 2012. Other traditional airlines adjusted their strategies and
concentrated in cost reductions as an answer to the LCCs (Barrett 2004). Charter airlines
were also affected and in many cases they are facing direct competition by LCCs. In many
popular vacation destinations, LCCs offer frequent and flexible itineraries allowing shorter
vacations with smaller budget. In areas such as coastal Spain, LCCs are the preferred
method of air travel, further limiting the market share of both flag carriers and charter
airlines (Martinez-Garcia and Royo-Vela 2010). Affordable prices and frequent connections
have contributed to the popularity of weekend travel in Europe and have influenced

positively real estate and timesharing activities. According to a survey in UK over 800.000
residents had a second home abroad, an increase of 45% compared to the figures only
three years earlier (Lei and Papatheodorou 2010). It is widely accepted that LCCs do not
follow the typical hub and spoke network scheme used by other carriers. Instead they favor
point to point connections based on secondary or regional airports. Table 1 shows the
strategies used by LCCs to lower their costs.
Table 1: Cost reduction strategies adapted by LCCs
Areas Goal and result of strategy
Limited or no use of intermediaries

Sales
Direct sales through internet
Reduction of additional services at airports and during flights
Passengers
Single cabin layout
Aircraft manufacturers Negotiation for big discounts
Intensive utilization of aircraft and crews
Personnel and aircraft Use of single aircraft type and interchangeable crews with
common type ratings
Procedures for restructuring

Negotiation for low fees and pressure for indirect subsidiaries
Airports Use of secondary airports with excess capacity
Creation of competition between airports
For many decades European flag carriers enjoyed several privileges, including the de facto
control of major airports. Since they had been operating on marginal profitability, it has
often been argued that high salaries, benefits, and pensions combined with governmental
protection resulted in very low productivity. At the same time, possibilities for entrance of
competitors and introduction of cost strategies were practically nonexistent (Barrett 2004).
According to data from ACI (2010) the LCCs’ market share increased from approximately
10% to over 30% in 2006. In regions such as Asia and Australia the trend remains
significant. During 2001 and 2009 LCCs had a steady increase of 38% on average annually,
compared to the total increase in the region that did not exceed 6%. During the same
period, the number of cities connected by LCCs increased from 48 to 576. LCC expansion is

not uniform in every continent. Combined with recent economic crisis, several regions have
witnessed a sudden halt in growth. According to the European Low Fare Airline Association
(EFLAA 2010a), during 2009 the members of the organization carried almost 9% more
passengers compared to 2008. The activities of LCCs showed a slight decline during 2009 in
most major markets, including Germany, Italy, Spain, and UK. Domestic flights seem to be
more resistant to the effects of economic turbulence. For example, in UK, during the first
years after 9/ 11 LCC capacity grew enormously. In 2007 it reached a peak and after a
couple of years of decline, in 2009 it had dropped back to the levels of 2006 (Centre for Asia
Pacific Aviation 2010).
In any case, LCCs seem to have acquired a reasonable share that is steadily around 30% of
the total intra-European capacity (EFLAA 2010b). LCCs managed to seize most of the
capacity growth in Europe between 2000 and 2009. Focus has now moved to the promising
markets of Eastern Europe. According to Boeing Corporation, the global expansion of LCCs is
one of the main reasons for the predicted growth of aviation (Boeing 2010). Growth rates
are expected to be much higher for LCCs compared to traditional carriers and charter
companies, based on recent analysis released by Boeing and other stakeholders. The
following table (2) shows the airline market status before and after the deregulation.
Table 2: Airlines Before and After the Deregulation
Before Deregulation After Deregulation

Government support and privileges for flag
Removal of state intervention
carriers
Obstacles for entry of new airlines I ncrease of competition
Price regulation and limited competition Bankruptcy of traditional carriers
Appearance of LCCs
Significant reduction in average airfares, up to
80% in Europe
Increase in passengers
High costs for human resources
Further adoption of point -to-point model
Traditional carriers either turned into LCCs or
restructured their business model closer to
that of LCCs

In Canada, during the last decade, competition from LCCs has led into a series of mergers
and a complete re-structuring of the market (Lawton 2004). In Asia, in competitive markets
such as Southeast Asia and Japan, new carriers have appeared for the first time after many
years. The same is true for many European countries as well. In Spain, world’s second most
popular tourist destination, more than one third of tourists are carried by LCCs. In some
regions such as Catalunia, LCCs have larger market share than traditional airlines. Girona in
Spain is a typical example of low cost airport that expanded rapidly and assisted significantly
the growth of tourism in the region (Martinez-Garcia and Royo-Vela 2010). According to
Fageda and Fernandez-Villadangos (2009), the competition has not affected most of the
passengers travelling to and from the major airports, where traditional carriers are based.
The benefits are more clearly visible at the airports used mainly by LCCs. Based on data
from the Spanish market, Fageda and Fernandez-Villadangos argue that the presence of
LCCs results in lower airfares for any carrier that serves the same destinations as the LCCs.
Flag carriers such as Alitalia in Italy have concentrated their operations around major cities
like Rome and Milan, allowing LCCs to create new direct connections between secondary
airports. A new generation of previously unimportant and nowadays rapidly growing airports
has emerged: Bologna, Venetia, Pisa, Torino, and Genoa (Barrett 2004).
5. RESPONSE OF TRADI TI ONAL CARRI ERS TO LOW COST RI VALS

Traditional airlines tried to maintain their market position by various strategies. One of them
was the establishment of their own LCCs while another one was the spontaneous cost
reduction. Many carriers in North America and Europe tried to create their own LCC brands
as a direct answer to the aggressive LCCs. While holders of AOC, these LCCs were actually
business units or “airlines within airline” of their parent company. To counter the rise of
LCCs, Continental airlines established Continental Lite in 1993. First class was removed from
the aircraft, no meal was served and flights were typically less than two and a half hours.
Even though Continental Lite operated with no less than 100 airplanes, the attempt proved
to be both short-lived and extremely costly. Delta Air Lines made two similar attempts.
Delta Express was created in 1996 to compete with Southwest, Air Tran and JetBlue. It was
replaced by Song in 2003; Song was also 3 years later absorbed back to Delta. US Airways
created Metrojet in 1998 to compete with Southwest and Delta Express. Metrojet ceased to
exist shortly after 9/11 (Vasigh et al 2008).

Following the examples from the other side of the Atlantic, British Airways and KLM bouth
introduced their own LCCs. In 1997 British Airways created Go Fly operating from London
Stansted. The company advertised its ties to British Airways and was profitable in 2000.
However the new management of British Airways blamed Go as one of the reasons for the
main company’s declining passenger volumes and decided it did not suit their revised
business model. Go was bought and subsequently merged by EasyJet. In 2000 KLM created
Buzz in order to compete with LCCs such as EasyJet, Ryanair and Go in the British market.
Not following one of the main operational rules of LCCs, Buzz maintained two separate small
fleets of BAe 146s and Boeing 737-300s. Without economies of scale, the operational costs
were quite high and Buzz was soon to follow the fate of Go. Ryanair bought Buzz, kept it
under the brand for a year and finally absorbed the short-lived rival (Vasigh et al 2008).
The above examples show that in both sides of the Atlantic the experiments of the
traditional carriers with their own LCCs were disappointingly unsuccessful. Part of the result
can be explained by the fact that operation of these LCCs was newer low enough, especially
in terms of labor cost. Since this strategy did not bring the expected results, the next
approach was to limit their service contents. However, providing a stripped product would
bring their services closer to the ones offered by LCCs. Since LCCs had a much lower
structural and operational cost, this would have been a very risky strategy. As most
reactions towards LCCs proved to be partly or completely futile, many traditional carriers
tried to compete them by actually avoiding competition; concentrating on long haul flights
and international routes where LCCs were in disadvantage due to legal restrictions. (Vasigh
et al 2008).
6. EUROPEAN AI RPORTS
6.1 Current trends and overview

In 2010 more than 1,600 airports in all continents were members of the Airports Council
International (ACI). Over 98% of global air passengers travel through ACI members. The
4.9 billion passengers travelling in 77 million flights are expected to double in next 15 years.
According to data available, the global financial crisis starting from 2009 and the increasing
oil prices have limited the recent growth rate. Half of the airports witnessed increase in
terms of passengers served. Several maj or airports showed a decline, while smaller ones

strengthened their position. At the same time, the increase in passenger volume is higher
than the increase in aircraft movements (6.6% and 2.4% respectively in 2007), which
indicates both a preference for larger capacity aircrafts and better utilization of fleets (ACI
2010).
In ACI’s statistics for 2010, five European airports are among the 15 largest ones. The sizes
of the airports seem to be directly related to long distance flights. Recent research (Gillen
2007), argues that European airports as a whole have three distinct characteristics. First,
there are a large number of airports with scheduled flights, disproportionate to the size or
population of the countries. Countries like Greece, Norway, or Sweden have 38, 51 and 44
airports respectively, while France and Germany have 68 and 48. Second, the density of the
airports results in low utilization. In Ireland, two thirds of the airports serve less than
100,000 passengers annually and this is also true for most French airports. Third, the major
central airports depend on an effective and extensive rail network that expands their
catchment area and allows for combined air and high speed ground travel. The ownership of
the airports varies; Spain, Portugal, Sweden or Greece have publicly funded and operated
airports, while UK has privatized them. In Germany and France airports are in the
responsibility of the local governments.
6.2 Financial aspects of the airport operations

During the new millennium, European airports are facing two main challenges: pressure for
cost reduction in terms of ground handling and fees, and adoption of new strategies to
reduce delays. Additional problems are related to pollution, land use and other
environmental factors. Although most airports are still under state control and are often
used as instruments for national and regional development, the new trend adopted by most
stakeholders dictates the sustainable operation of all airports. Even during periods when all
airlines recorded high loses (e.g. after 9/ 11) all major European airports managed to remain
profitable. Airports have two sources of income: aeronautical from flights and commercial
from other activities. Commercial revenues have grown significantly during last decades and
today contribute by over 50% in total income. While at the same time labor cost has
decreased, investment depreciation has increased steadily, reaching nowadays over one
quarter of total cost.

6.3 Airports and destinations
Between 1994 and 2003 passenger traffic increased globally by 5% annually. During the
same period many European airports had very high growth rates. Typical examples are
London Stansted (46% ), Antalya (26%), Prague (18% ), Vilnius (13% ), Warsaw (12%),
Barcelona (11%), Madrid (11%), Paphos (9%), and Budapest (8%). An average annual
growth of 10% means that these airports effectively doubled their customers within the
decade. Besides the increase in passenger volumes, the airports also expanded their
connections: Stansted served 28% more destinations, Bratislava 20%, Palma de Majorca
7%, Munich 6%, Ljubljana 6%, Prague 5%, and Budapest 3% (ACI 2010). From a statistical
point of view, there seem to be a positive correlation between increase in number of
destinations and number of passengers. In most cases, new destinations were the result of
new routes established by LCCs.
6.4 Dynamics of Point- to- Point netw orks

Traditional carriers expanded their networks based on hub and spoke models. However, as
Chang and lee point out, the establishment of those networks was mostly based in
experience and intuition with reasonable cost being the main target (Chang g Lee 2010).
In the past, point-to-point networks were the choice for regional airlines serving small and
medium distance connections. LCCs adopted the same model as one of their main strategic
tools. Point-to-point flights have typically higher operational costs in comparison with
services based on a hub model. On the other hand, they have the advantages of higher
reliability and more convenient schedules. Since LCCs do not need to worry about
connecting flights, they have greater flexibility in the selection of suitable airports, including
secondary and regional airports with additional advantages. According to a research
conducted in 2005, for the management of LCCs there are three main factors considered for
the selection of an airport. First, the air travel demand must be high enough, second, the
facilities must allow for a short turnaround time, and third, there must be availability in slots
(Chang & Lee, 2010). In general, LCCs are not willing to share an airport with many
competitors, although they prefer airports with good land connections. Table 3 shows the
status of airports before and after the deregulation.

Table 3: Airports Before and After Deregulation
Airports Before Deregulation Airports After Deregulation

Lack of price competition. Higher prices for Airport restructuring into a more dynamic
airfares. environment
Very limited incentives for productivity and Airports are transformed from public facilities to
efficiency modern business units
Seasonal use by charter flights Airports help the expansion of LCCs
Limited vertical integration between airports
Competition between airports intensifies
and airlines
Commercial revenues increase, especially at the
Limited commercial revenues
airports used by LCCs
At regional airports the low revenues are not Airports start t o see passengers as their own
enough t o cover operational expenses. customers as well
Regional airports act as feeders to major Regional airports support their own networks.
airports. Their location is turned into an asset.
7. I NTERACTI ON BETWEEN LCCs AND AI RPORTS

According to studies, airports had very high initial capital cost and low marginal cost for
each additional flight and passenger. Based on calculated economies of scale, the marginal
cost decreased sharply for the first one million annual passengers, continues to decrease
until three million passengers and remains relatively stable after that. The 25 largest airports
– which represent 2% of the 1192 airports with international flights – serve more than 32%
of total air traffic. The global uneven distribution of passengers is one of the biggest
challenges for any airport. Since an airport needs a critical mass of passengers before it can
become economically viable, the target is to cover initial costs and sustain expected
damages over a period of growth leading to a next stage of profitable operation (Francis et
al 2003). Traditionally, airlines were the customers of airports. However, as the commercial
revenues have started to form a significant source of income, airports are gradually treating
passengers as if they were their own customers. At the same time, airlines consider
passengers as their exclusive customers, brought to the airports by them. These views
create a complicated and specialized relationship between three elements (Gillen and
Morrison 2003). For the regional and the smaller airports, limited number of flights is
translated into equally limited aeronautical and commercial revenues. The possibility to

attract LCCs is becoming an important solution and while there is extensive research around
LCCs, there is not enough knowledge for their exact effects on airports (Francis et al 2003).
After deregulation, airports have started transforming themselves from state controlled and
financed facilities into competitive business units. Flag carriers that enjoyed a dominant
position in major airports and monopolies in regional ones are now forced to share their
former back yard with other carriers, including LCCs (Fageda and Fernandez-Villadangos
2009). This interaction is often accompanied by tension and disagreement (Barrett 2004).
Although LCCs are attractive for airports, they do not have the stability associated with flag
carriers. Airports have to develop scenarios and assess the possibilities of LCcs withdrawing
from destinations or from the market altogether (Gillen and Morrison 2003).
Whatever they may chose airports have no other option than to adjust into the new highly
competitive environment of deregulation and LCCs (Barrett 2004). The value and importance
of an airport for a LCC is based on its location and catchment area. When two or more
airports share the same area, they directly compete with each other. During 2002 Southwest
was invited by more than 140 airports and only very few “lucky ones” were included in the
company’s network (Fageda and Fernandez-Villadangos, 2009). In Europe, LCCs such as
Ryanair are in continuous negotiations with airports. Large airports in the vicinity of
metropolitan areas often sign up attractive contracts with favorable terms and acceptable
collectable fees. On the other hand, abandonment of Rimini in favor of Ancona in Italy by
Ryanair demonstrates the power LCCs exercise over smaller airports. (Fageda and
Fernandez-Villadangos, 2009). Table 4 shows what LCCs demand and what they offer to an
airport in order to establish cooperation.
Table 4: What LCCs Ask and w hat they Offer to Airports
What LCCs ask from airports What LCCs offer to airports

Excess capacity I ncreased traffic
Fast and effective ground services Increased market share
Short turn-around time of 25 minutes Increase in aeronautical revenues
Good local transportation Increase in commercial revenues
Low airport fees Enlargement of catchment area
Suitable slots Above average increase in vehicle rentals
Possibilities to increase the catchment area Reduction in the capital costs

The presence of an LCC leads to a significant increase in terms of passengers. Even if the
airport agrees to lower fees in order to attract an LCC, the passenger volume increase could
alone result in an overall positive situation, due to the associated commercial revenues.
Furthermore, it can be argued that the bare services model used by LCCs creates
opportunities for airports to increase their ground sales and services. Since commercial
revenues at smaller airports are usually less than 35% of total revenues, this appears to be
a realistic expectation (Graham 2001). From a certain point of view, this could be explained
as a strategic choice between aeronautical and commercial revenues, where airports choose
a different mix and balance between their main sources of income. According to Barrett
(2004) such trends are visible during last two decades.
Traditionally, secondary airports have a limited role either as passenger feeders to central
airports, or as points serving seasonal charter flights (Fageda and Fernandez-Villadangos
2009). LCCs offer the possibility of a more autonomous development with increased
passenger volume and larger catchment area (Lei g Papatheodorou 2010). Airport
managing companies favor the use of secondary airports as supplemental to their main
hubs. This can be observed in cases such as Stansted for Heathrow and Hahn for Frankfurt
(Barrett 2004).
From the passengers’ point of view, the selection of remote airports by LCCs in conjunction
with the other existing airport and airline management strategies has both positive and
negative outcomes. First, it has made lower airfares a reality. Second, it has moved
passengers from congested central hubs to smaller and friendlier facilities. And third, in
many cases it has increased the land travel distances and time (Barrett 2004). Passengers,
including both leisure and business travelers have in general accepted the inconvenient
locations even though many secondary airports are very far away from the metropolitan
areas they are supposed to serve (Lawton 2004). Additionally, the expansion of the
catchment area can lead to overlapping between airports, causing intense competition not
only between main and secondary airports but between regional airports as well. Although
the situation may not be desirable for the airports, it does offer more choices to the
passengers (Francis et al 2003).

8. CAPACI TY AND SEASONALI TY OF GREEK AI RPORTS
The analysis of 11 Greek airports shows that there is high seasonality in most of them. An
interesting observation is the very low utilization of the available apron capacity.
Figure 1: Apron Utilization by Airport
Source: Katarelos, E. and Lagoudis, I. (2011)
The seasonality can also be observed in figure 2, especially in the island airports of
Zakynthos, Santorini, Corfu, and Kos.
Figure 2: Seasonality of Greek Airports

During high season all Greek airports, with the exception of Heraklion, utilize less than 50%
of their apron capacity; actually most of their capacity remains completely unused form
extended periods of time. Figure 3 shows the predictions of Greek airport utilization up to
year 2030, based in three scenarios and historical data of 18 past years. It can be argued
that current capacity is enough to sustain a steady increase for the next twenty years.
Capacity issues would become a problem only in the most optimistic scenario of 7.5%
annual increase and even then, it would need at least a decade before full capacity is
reached in most airports (Katarelos and Lagoudis 2011).
9. RECENT TRENDS I N GREECE

Although the use of secondary airports is one of the main strategies for LCCs, they have not
adopted that particular rule in the case of Greece. Almost all non seasonal flights connect
the main hub of Athens and a few others the secondary hub of Northern Greece,
Thessaloniki. All other LCC flights to Greece are seasonal. These include some promising
links to regional airports that could potentially serve large catchment areas, such as the
airport of Volos, located in the mid-distance between Athens and Thessaloniki. The lack of
suitable airports in combination with a non-existent national regional development strategy,
are possibly two of the main reasons for this unusual choice of LCCs, a “paradox” as
described by Papatheodorou and Arvanitis (2009).
Figure 3: Greek Airports and Future Capacity Scenarios

The recent inauguration of flights to Volos from Ryanair in May 2010 and the presence of Air
Berlin at the same airport are currently seasonal. It is argued that although the potential
exists, Volos fails to extend the catchment area both towards Athens and towards
Thessaloniki, the two main metropolitan areas of Greece with a combined population of over
6 million people. The land travel distance exceeds two hours in either direction and the
airport which has a mixed military and civilian usage, lacks necessary passenger
infrastructure. Responding to these shortcomings, the authorities have initiated the
construction of a new terminal building in order to improve passenger services. Additionally,
the airport is located outside the 100 km radius exclusive zone of Athens International
Airport where current legislation prohibits the operation of any public airports with
commercial flights and activities. Although it is not forbidden for private sector to build and
operate a private airport inside this zone, under current conditions it seems highly unrealistic
that any entrepreneurs would be willing to make investments of the necessary scale.
Besides, the main idea around secondary airports is the use of existing ones and the
utilization of their idle capacity and not the creation of new airports. As Papatheodorou and
Arvanitis (2009) observe, the area surrounding Volos has the potential to support scheduled
flights that would not be limited to seasonal and recreational demand. It is also interesting
to note the announcements accompanying the launch of the new connections to Volos;
Ryanair stresses the importance and the direct and indirect benefits of the flights for the
local economy, which would “create 200 new jobs” and “boost the Greek economy” (Ryanair
2010).
It could be argued that even if Ryanair or any other LCC manages to pay very low fees for
the use of the airport of Volos, or if the construction of specific infrastructure becomes an
indirect form of subsidies, the investment could still have significant long term benefits for
the region. This is true for other continental Greek airports as well. Most of them are in
parallel use by civil aviation and military with limited commercial infrastructure. The cost to
further develop these airports is not prohibiting and the excess capacity can be utilized with
relatively limited efforts. Since they are not attractive for traditional carriers and some of
them are not near popular tourist destinations, they could be a good choice for LCCs for two
reasons. First, especially in Greece, even less favorable regions have great potential for
development and are near various interesting sites. Both conditions are met by the airport of
Volos and obviously Ryanair and Air Berlin have taken them into account. Second, charter
operators and mostly the big tour operators do not offer any guarantees or stability

regarding their presence and therefore their activities are not only seasonal, but also have
high risks as instruments of long term development.
The above example of Volos describes the local perspective and the effects of the local
airport. On the other side of the same issue are the European airports and their prospective
to remain competitive. Any new destination linked to them, is measured in additional
aeronautical and commercial revenues as explained earlier. The strategy of expansion to
new destinations is more critical for secondary airports that have committed to LCCs in order
to remain viable. According to the point-to-point model, the second European member of
the link will be another secondary airport in another country. For each flight from Volos to
Frankfurt Hahn or Milano Bergamo, all three airports have their share in benefits and
revenues. The main items of the negotiations between LCCs and airports and possible
outcomes for each issue are summarized in Table 5.
Table 5: Greek Airports and Cooperation w ith Low Cost Carriers
What LCCs ask from Greek Airports Estimate of Greek airports’ potential
Excess capacity to accommodate increased
Exists or may increase
demand
Fast and effective ground facilities Exist or can be created
Exist due to seasonality and excess
Suitable time slots
capacity
Good local connections I n some cases needs improvement
Limitations in pricing policies due to
Lower airport fees current legislation and ownership of the
airports
Capital assets Can be raised
Enhanced facilities for ground transportation Possible to develop
As a general conclusion from the above table, Greek airports seem to have the potential for
cooperation with LCCs. During the recent past, one of the reasons limiting the ability or the
willingness of the local airports to negotiate openly with LCCs was the legal actions of many
traditional airlines against any contract between LCCs and airports. Their main argument
was that the low fees were in fact disguised public subsidies, forbidden by European aviation

framework. Recent decisions of the European Commission and the European Court (ECFI -
European Court of First Instance), in December 2008 regarding the airport of Charleroi in
Belgium, and in January 2010 regarding the Bratislava airport in Slovakia, rule that in both
cases the agreements between the airports and the LCCs are in accordance with the
European market and competition principles (EU Market Economy Investor Principle) (EU
2010). However, for one LCC, Ryanair, there are still several open cases with the question of
illegal public subsidies in the agreements between Ryanair and the airports of Alghero, Pau,
Lübeck, Frankfurt Hahn, Berlin Schönefeld, Aarhus, and Tampere (EU 2010). According to
the view for the side of the LCCs, the very low fees offered by some airports are part of
perfectly fair and legal commercial agreements that reflect the current market conditions
and trends and are balanced by the benefits of increased traffic and the creation of new
jobs (Ryanair 2010). In reality, until today, 2012, there has not been any pre-mature
termination of any agreement as a response to exposure of anti-competitive or other unfair
practices.
Currently, the initiative is in the hands of the airlines. LCCs evaluate and select routes and
airport pairs based on their own cost and efficiency targets. The authors suggests that it is
in the best interest of the airports to become actively involved in this process and interact
dynamically with airlines and local communities in order to promote or support the
expansion of suitable connections with other cities.
10. CONCLUSI ONS

Increase in number of destinations is a strategy mostly used by smaller European airports.
LCCs seem to prefer secondary airports for the deployment of their point-to-point network
model. Airports are interested in utilizing their excess capacity, while LCCs aim to minimize
their overall costs. When everything turns out as intended airports, LCCs, passengers, and
local economies can all benefit. To minimize the risks associated with the preferred form of
long term cooperation between airports and LCCs market conditions must be carefully
assessed. LCCs typically negotiate significant airport and landing fee discounts; airports
expect positive results from the increase in traffic and commercial revenues.

Greece has a relatively high number of airports, 39 in total. Since most of them have very
low utilization, LCCs appear to be an attractive opportunity. Greece uses almost exclusively
the hub and spoke model and point to point connections only exist in few subsidized PSO
lines. After deregulation, the emergence of private carriers in Greece offered lower airfares,
however the competition currently is quite limited and airfares relatively high. LCCs offer
connections mostly to the main hubs and only recently they have experimented with a
couple of regional airports. The majority of the smaller airports struggle to cover at least
part of the operating costs and theoretically any airport chosen by an LCC would potentially
have enough excess capacity to share.
A central issue that determines the relationship between LCCs and airports is the fact that
LCCs demand a long term contract to be signed. Under current legislation, most Greek
airports cannot sign such contracts which may additionally include special clauses. If and
when these obstacles are removed it would be possible for regional airports to offer
incentives to LCCs in order to attract them, as it has happened in other European countries.
Under current status, both the autonomy of the airports as well as their readiness to enter a
more competitive market is questionable. The centrally organized and applied state
management and development schema is considered to be both restrictive and ineffective.
Each airport should be assessed and as a unique business unit in order to select the most
suitable long term strategy.
REFERENCES
‚ ACI (2010), Airports Council International. Official website, last accessed 15/ 09/ 2010
‚ Barrett, S.D. (2004) How do the demands for airport services differ between full-service
carriers and low-cost carriers? Journal of Air Transport Management 10 (2004) pp 33–39
‚ Boeing (2010). Boeing Current Market Outlook 2010-2029, July 2010
‚ Centre for Asia Pacific Aviation (2010). Official website, last accessed 23/ 08/ 2010
‚ Chang, Y., Lee, N. (2010) A Multi-Objective Goal Programming airport selection model
for low-cost carriers’ networks. Transportation Research Part E (2010) (I n press)
‚ Delfmann, W., Baum, H., Auerbach, S. and Albers, S. (2005) Strategic Management in
the Aviation Industry. Ashagate Publishing Company, pp 1-2
‚ European Low Fare Airline Association - EFLAA (2010a). Official website, last accessed
12/07/2010

‚ European Low Fare Airline Association - EFLAA (2010b), “European LCCs gain
passengers, while network carriers lose passengers, in 2009”, March 2010
‚ EU (2010). Web sites of European Commission and EU Services, last accessed
14/07/2010
‚ Fageda, X., Fernandez-Villadangos, L. (2009) Triggering competition in the Spanish
airline market: The role of airport capacity and low-cost carriers. Journal of Air Transport
Management 15 (2009) 36–40
‚ Francis, G., Fidato, A. and Humphreys, I. (2003) Airport-airline interaction: the impact of
low-cost carriers on two European airports. Journal of Air Transport Management 9
(2003) pp 267-273
‚ Gillen, D., Morisson, W. (2003) Bunding, integration and the delivered price of air travel:
are low cost carriers full service competitors? Journal of Air Transport Management 9
(2003) pp 15-23
‚ Gillen, D., Lall, A. (2004) Competitive advantage of low cost carriers: some implications
for airports. Journal of Air Transport Management 10 (2004) pp 41-50
‚ Gillen, D. (2006) Airport Economics, Policy and Management: The European Union
‚ Graham, B., Shaw, J. (2008) Low-cost airlines in Europe: Reconciling liberalization and
sustainability. Geoforum 39 (2008) pp 1439–1451
‚ Iatrou, K. and Oretti, M. (2007) Airline Choices for the Future. From Alliances to
Mergers. Ashagate Publishing Company, pp 6-7, 14
‚ Katarelos, ., Lagoudis, . . (2011) Greek Airport Capacity Utilization: What about
Airfreight? International Journal of Aviation Management, Vol. 1, Nos. 1/ 2, 2011, pp 124-
139
‚ Lawton, T.C. (2004) Cleared for Take-Off. Structure and strategy in the low fare airline
business. Ashgate Publishing Limited, pp 86-87
‚ Lei, Z., Papatheodorou, A. (2010) Measuring the effect of low-cost carriers on regional
airports’ commercial revenue. Research in Transportation Economics 26 (2010) 37–43
‚ Lian, J.I., Rønnevik, J. (2010) Airport competition – Regional airports losing ground to
main airports. Journal of Transport Geography (In press)
‚ Martinez-Garcia, E. Royo-Vela, M. (2010) Segmentation of low-cost flights’ users at
secondary airports. Journal of Air Transport Management 16 (2010) pp 234–237
‚ Ryanair (2010) Company website, announcements 27.01.10 and 30.03.10, last accessed
18/09/2010

‚ Papatheodorou, A. and Arvanitis, P. (2009), Spatial evolution of airport traffic and air
transport liberalization: the case of Greece, Journal of Transport Geography 17 pp. 409
‚ Pels, E., Njegovan, N., Behrens, C. (2009) Low-cost airlines and airport competition.
Transportation Research Part E 45 (2009) pp 335–344
‚ Vasigh, B., Fleming, K. and Tacker, T. (2008) Introduction to Air Transport Economics.
From Theory to Applications. Ashagate Publishing Company, pp 323- 327
‚ Warnock-Smith, D., Potter, A. (2005) An exploratory study into airport choice factors for
European low-cost airlines. Journal of Air Transport Management 11 (2005) pp 388–392
‚ Zhang, A., Hanaoka, S., Inamura, H. and Ishikura, T. (2008) Low-cost carriers in Asia:
Deregulation, regional liberalization and secondary airports. Research in Transportation
Economics 24 (2008) pp 36–50.

QUALI TATI VE ANALYSI S OF THE RELATI ONSHI P BETWEEN THE PROFI LE OF
DEPARTI NG PASSENGERS AND THEI R PERCEPTI ON OF THE AI RPORT TERMI NAL
Michelle Bandeira 1
Department of Air Transport, Aeronautics Institute of Technology, Sao Jose dos
Campos, SP, Brazil
Anderson Correia 2
Department of Air Transport, Aeronautics Institute of Technology, Sao Jose dos
Campos, SP, Brazil
ABSTRACT
The opinion of passengers on the degree of importance of the components is required

in order to prioritize services. A low service level can, besides causing inconvenience
for terminal users, increase the waste of resources and increase costs if there is no
adequate planning. Hence, outlining passenger profiles at the airport is relevant to
strategic planning of airport activity management. It is believed that individual
characteristics could influence opinion on the degree of importance or about the quality
of airport services. This article shows that the check-in and the departure lounge were
considered the most important areas in the airport terminal by passengers. Finally it
was noted that the age and reason for travel influenced the passengers’ perception
about the check-in area and the frequency of flying influenced the perception of the
departure lounge.
KEYWORDS: Passengers’ opinion; passengers’ profile; AHP; airport terminal;

independence test.
1
Michelle Carvalho Galvão da Silva Pinto Bandeira, MSc, is a PhD student at the Aeronautics
I nstitute of Technology. Cont act Details: mgalvao@ita.br.
2
Anderson Ribeiro Correia, PhD, is Professor of Air Transportation and Airports at the
Aeronautics Institute of Technology. Contact Details: correia@ita.br.
1. I NTRODUCTI ON
In recent years, structural changes such as commercialization, privatization, and

globalization, together with increased competition between airports, have encouraged
airports and aviation authorities to place more emphasis on quality (Graham, 2008). In
this context, establishing measures to evaluate operational performance of airports is
one of the major problems facing airlines and airport operators today (Correia, 2009).
Airport managers have to struggle with the decision of prioritizing resources. Although
they are motivated to offer a reasonable level of service (LOS) to passengers, there is
a growing worldwide tendency for cost reduction. In this scenario, an effort to
determine the importance that various passenger groups attribute to airport
components would provide a useful indication of where airport managers should invest
their limited resources such as funds, employees and their own attention.
The airport terminal may be considered a set of subsystems that interact between
themselves to allow a change from land mode to air and vice versa. Various
components are installed and different services are produced around these movements
– passenger departure and/ or arrival – in order to meet client expectations. Some
services and areas of the terminal on general are used by passengers, following the
flow of departure or arrival. A low level of service can result in, besides inconvenience
to terminal users, the waste of resources and increased costs if there is no adequate
planning. Hence, service level targets are important because they have serious
implications for costs and the airport’s economy, as well as the “image” transmitted to
the clients and to society (Bandeira, 2008; Ashford et. al., 1997).
Besides the operational and financial concerns, outlining the profile of passengers at
the airport contributes to the drawing up of a strategic plan for the management of
airport activities. I t is believed that individual characteristics related to the frequency of
flying, the reason for flying, income, age, and other factors may influence the opinion
of the degree of importance or the quality of the services in an airport. The answers to
these questions make all the difference in airport planning and in this article are
expected to broaden this understanding.
2. LI TERATURE REVI EW
Brink and Madison (1975) presented one of the first studies done in the area of airport
service levels. They considered that passengers’ perceptions of the airport terminal,

besides being influenced by the technical and operational conditions, also depend on
subjective factors and each person’s individual characteristics. Some criteria and
characteristics proposed by the research include the purpose of the trip, the frequency
of flying, costs of air tickets and airport services. Other authors, such as Omer and
Khan (1988) Müller and Gosling (1991) and Ndoh and Ashford (1994) concerned
themselves mainly with the method used to collect and analyze passengers’ opinions;
that is, using a model than can transfer linguistic judgment into quantitative values.
Lee and Kim (2003) state that passengers may have different perceptions about
services and installations related to departure and arrival processes in an airport
terminal. In other words, the route the passenger takes and the services related to
their objective – departure, connection, arrival – influence the perception of the service
level of the airport. I n another study, Seneviratne and Martel (1991) developed a study
in which they presented a selection of components of greater importance in the
terminal assisted by a passenger opinion poll in some Canadian airports. According to
these authors, passenger needs can change according to the installations. A manual of
service quality in airports developed by Airports Council International (ACI, 2000)
states that the detailing of the types of clients and services enables comprehension of
the different processes in which quality of the services must be acquired.
Despite the important effort made by the researchers and entities cited above, there is
a major lack of studies which research and identify whether there is a significant
relationship between the evaluation of the service level and the social and economic
profile of the users interviewed. This study intends to approach this question, the
development of which will be detailed in the following sections.
3. RESEARCH METHODOLOGY
Field research was carried out through interviews with 270 passengers in departure
lounges at the São Paulo/Guarulhos International Airport between August 2006 and
October 2007. For the size of the sample a 6% error margin was allowed and a
confidence interval of 95%. Initially the degree of importance of the passenger
departure terminal areas at the airport in question was sought, and their respective
indicators. The Analytic Hierarchy Process (AHP) method was employed to get the
degrees of importance for the attributes according to the passengers’ opinion.

The interviews observed passenger characteristics such as income, age, reason for
travel, frequency of travel, and type of trip. Each one was divided into classes, as
shown in Table 1.
Table 1: Characteristics and classes analyzed
CHARACTERI STI CS CLASSES
1. I ncome up to US$ 40,000

HOUSEHOLD INCOME 2. Income from US$ 40,000 to US$ 80,000
3. I ncome above US$ 80,000
1. Aged up to 30
AGE 2. Aged between 30 and 50
3. Aged above 50
1. Business travel;
REASON FOR TRAVEL 2. Leisure
3. Family reasons
1. 1x a year
FREQUENCY OF TRAVEL 2. 3x to 6x a year
3. Over 6x a year
A statistical treatment was applied to the sample (for each variable used) to identify
whether the responses were significant as regards the degree of importance of the
indicators linked to these areas. As of this point, it was possible to compare
passengers’ opinions against their different profiles through the AHP method. In
addition, it was checked whether these qualitative variables influenced or not opinion
as to the degree of importance. In this case, the independence test from the Chi-
squared method was used. There follows a description of the methods used for the
current study.
3.1 APPLI CATI ON OF THE AHP METHOD

This work used the hierarchical structure of the method to get the global weights for
the airport components. A scale of percentage values was used to get the weights,
corresponding to the values from the fundamental Saaty scale so that the passengers
could relate the scale to some kind of linguistic or verbal concept during interviews
(Bandeira, 2008). Table 2 shows the scales cited.

Table 2: Relation between the Percentage Scale and the Fundamental Scale
Percentage Fundamental Scale

Degree of relative importance
Scale ( Saaty)
TPS
Components Weights Definition
A* B*
90% 10% 9 A is extremely more important than B.
80% 20% 7 A is much more important than B.
70% 30% 5 A is more important than B.
60% 40% 3 A is a little more important than B.
50% 50% 1 A and B are of the same importance.
40% 60% 1/ 3 B is a little more important than A.
30% 70% 1/ 5 B is more important than A.
20% 80% 1/ 7 B is much more important than A.
10% 90% 1/ 9 B is extremely more important than A.
*A and B represent airport terminal components.
Source: Bandeira, Correia and Wirasinghe (2007)
The individual values for each passenger were aggregated in this research in a
geometric average. In the case of an arithmetic average, which gives equal weight to
all the averages, the results would be biased, as there would be a tendency to
disproportionately value a set of weights supplied by the passengers. The Equation (1)
shows the geometrical average used to get the final average of the weights given by
the passengers.
w f (C i ) ? R Pd k
s
s (1)
k ?1
In which:
Ci: Component i;
Pd k : Weight given by the passenger dk;
dk: Passenger (1...k)
s: Number of passengers;
As the AHP method is based on peer to peer comparisons, judgments are put in a
squared matrix n x n, where the lines and columns correspond to the n criteria
analyzed for the problem in question.
Considering C ? ]a ij _, with i, j = 1. 2. ..., n, called the “decision matrix”, each line i
supplies the reasons between the weight of the criterion or sub-criterion for the index i
for all the rest. The matrixes are always reciprocal, such that a ? 1 , and positive.
ij
a ji

Hence, the value a ij represents the relative importance of the criterion for the line i
given the criterion for the column j , where only the principal diagonal assumes values
equal to 1. Peer to peer comparisons are made at all levels of the matrix A. Therefore,
if all the judgments are perfect, in all comparisons it would be possible to see that
a ij · a jk ? a ik , for any i, j, k = 1. ..., n, therefore, following this procedure, matrix A,
would be consistent.
Take n as the number of elements to be compared, nmáx the auto-vector of A and w

the correspondent proper vector or vector of priorities. If the judgments made by the
decision maker are perfectly consistent, the result is nmáx ? n e a ij ? x i . However,

xj
almost always some inconsistency is seen in the judgments, which is nevertheless

admitted by the AHP method.
The inconsistency can be measured in the following way: the closer the nmáx value is to
n , the greater the consistency of the judgments. Saaty (1980) showed that A being a
value matrix, the vector that satisfies Equation (2) will be found.
CW ? nmáxxW (2)
In which:
A: Decision matrix;
máx: Maximum autovalue of A;
W: Autovector of A associated to máx.
After the normalization of W, in (2), the auto-value máx is gotten from Equation (3).
n máx ? Â
1 n [ Aw]i
(3)
n i?1 wi
In which:
A: Decision matrix;
máx: Maximum autovalue of A;
W: Autovector of A associated to máx;
n: Order of the decision matrix;
Wi: Normalized Vector W.
It was observed, furthermore, that small variations in aij caused small variations in
nmáx, in which the auto-vector’s deviation in regard to n (the order of the matrix
number) is considered a measurement of consistency. I t can be said that the auto-
vector gives the order of priority and the auto-value is the measurement of consistency

of the judgment. For Gomes et al. (2004) it is possible to state that máx allows an
evaluation of the proximity of the scale developed by Saaty (1980) with the scale of
reasons or quotients that would be used if matrix A were totally consistent. This can be
done by means of a consistency index (CI). Therefore, according to Saaty’s theorem,
“ A is consistent if, and only if, nmáx n .”
So, if “ A is consistent if, and only if máx = n”, the value ( máx – n) is an indicator of the
consistency of judgments after the formation of A and the obtaining of normalized W.
The closer to zero such a difference is, the greater the consistency of judgments will
be. It must be stressed that this value must serve as a warning to the decider and/or
analyst, not only as an excluding situation. Therefore, the magnitude of the
perturbation in matrix A is calculated using the relation of the Equation (4).
n máx / n
CI ?
n /1
(4)
Based on the theorems describes, Saaty (1980) proposed the calculation of the reason
of consistency (CR) for the decision matrix A in Equation (5).
CR ?
CI
(5)
IR
In which:
CR: Consistency ratio;
CI : Consistency Index;
IR: Random Index.
The greater the CR, the greater the inconsistency of the matrix will be. Generally, an
inconsistency considered acceptable for n > 4 is a CR 0.10. The random index has
been calculated for matrixes squared by an order of n by the Oak Ridge National
Laboratory, in the United States (Saaty, 1991; 2005). Table 3 shows the values for IR
for the matrixes of order n x n.
Table 3: I R Values for Matrixes Squared by an Order of n x n
nxn 1 2 3 4 5 6 7 8 9 10
IR 0 0 0.58 0.89 1.11 1.25 1.35 1.40 1.45 1.49
Source: Saaty (1991)

Having done all this analysis of the judgment for matrix A, and given that this matrix is
coherent the results are normalized by the Equation (6). So, the priority vector for sub-
criterion i ( Aij ) in relation to criterion ( Ci ) is presented in Equation (7).
p i ( A1 ) ?
Âa
a ij
n
(6)
i ?1
ij
Âpi( A )
n
p i ( Aj ) ?
j ?1
j
(7)
n
In which:
p:
i: 1...., n;
Vector ;
A: Criterion for the second level (sub-criterion s);
n: Nº of criteria for one and the same level.
The following formulas, Equations (8) and (9), do the evaluations.
x j (C i ) ?
ÂC
C ij
m
(8)
i ?1
ij
x (Ci ) ? Â
m
x i (Ci )
(9)
j ?1 m
In which:
x :
j: 1...., m;
Vector;
C: First level criterion;
m: Nº of criteria for one and the same level.
Finally, a process of aggregation allows the generation of final values for the weights
of the airport components, ordering them through the following additive function of the
Equation (10).
f (Aj ) = Â w(Ci ) · vi ( Aj )
m
(10)
i ?1
In which:
j: 1...., m;
For the purposes of calculation, the areas of the terminal were designated with criteria
for the first level, and their respective indicators in criteria for the second level or sub-
criteria. The modeling indicated the importance and intensity of each one of the airport
terminal components.

3.2. STATI STI CAL TREATMENT OF THE SAMPLE
Considering that for the AHP method binary correlations may indicate whether an
element is preferable or equivalent in importance in regard to another, there are two
possible situations to be tested.
The first situation was to verify whether the percentage of equivalence found in the
binary comparisons was statistically significant. Hence, a designation was made for
each binary correlation for the number n of interviews in the sample and the
parameters m, P1 e P2. which were calculated according to the frequency f observed for
a certain airport component, if it was equivalent or preferable to the other.
In which:
P1: Population proportion regarding the first element of the binary comparison;
P2: Population proportion regarding the second element of the binary
comparison;
m: Population proportion regarding the equivalence of the binary comparison;
In this case, there is the first test of the hypothesis, in which the nullity hypothesis is
H0: m P1 + P2 and the alternative hypothesis is H1: m > P1 + P2. where P1 and P2 are
popuational proportions from the sample, and m is equal to the proportion of the
sample when in comparison between two airport components. This test evaluated
whether the degree of equivalence (equality) between the components was statistically
significant, considering g = 5%. Therefore, the nullity hypothesis was only rejected if
Z* < - Z5% ., where Z* is the confidence interval.
For the rejected hypothesis H0. the second hypothesis test is applied, with H0’: P1 = P2
e H1’: P1"Œ P2. to verify whether there had been any significant differences between the
proportions isolated for preference in airport components observed in the binary
comparisons. Hence, the nullity hypothesis was rejected if ÍZÍ > Z2.5% , with g = 5%.
Therefore, in this second situation the hypothesis H0’: P1 = P2 is accepted if a
component does not present relative preponderance in a comparison; or H0’ is rejected
if one component is preferable to another one.

3.3. CHI -SQUARED METHOD FOR ANALYSI S OF QUALI TATI VE VARI ABLES
Chi-squared (e2) is a non-parametric method used to test hypotheses in order to verify
a dispersion value for two nominal variables and to evaluate the association between
qualitative variables.
The main principle of this method is to compare proportions; that is, the possible
divergences between the frequencies observed and expected for a certain event.
Hence, it can be said that two groups behave in a similar way if the differences
between these frequencies in each category are very small or close to zero (Spiegel,
1972).
One measurement of the discrepancy between the frequencies observed and those
expected is provided by the statistic e2. expressed by Equation 11. The results obtained
are in the Contingency Table.
*oi / ei + 2
e2sample Â
k
= (11)
i ?1 ei
In which:
oi : Frequency observed;
ei : Frequency expected;
k: 1...., k;
i: 1...., i;
For the application of the method, it is necessary that the sample be relatively large
with sample N > 40 or at least 5 observations in each plot formed by the variable
analyzed. Furthermore, the data analyzed must be independent of each other and the
observations must have frequencies or counts where each observation belongs to one
and only one category.
It is stressed that if the significant value of e2 was gotten from one small sample (N <
40) and/or from a small expected frequency in a plot (typically when less than 5) for
formula for the obtaining of e2 may produce a greater-than-real value (Spiegel, 1972).
In this case the Yates Correction must be applied (or a continuity correction). The
statistics for the test are shown in Equation 12.

*| oi / ei | /0.5+2
e2corrigido = Â
k
(12)
i ?1 ei
To evaluate the condition of independence or dependence of the qualitative variables,

two hypotheses are tested:
H0: e2sample ø ec2
H1: e2sample > ec2
So that ec2 is the critical Chi-squared measurement with degrees of liberty GL given as
in Equation 13:
GL = (l-1)* (c-1) (13)
In which:
l: Number of lines formed by the classes for one variable x;
c: Number of columns formed by classes for a variable y;
That is, for the current research, the hypotheses cited indicated:
The inherent characteristic for the passenger does not influence the
H0:
opinion given to the degree of importance of the airport component.
The inherent characteristic for the passenger does influence the opinion
H1:
given to the degree of importance of the airport component.
So, from the null independence hypothesis, H0 is accepted when the value of e2sample
found is less than or equal to the value of ec2 designated. H0 is rejected when the value
of e2sample is greater than the value of ec2 designated. In the latter, H1 is accepted and
it is assumed that the variables in question present a dependency relationship.
4. GLOBAL TPS RESULTS

The hierarchical structure presented the values of the priority vectors found for the
areas of the terminal and its respective indicators. The chart below in Figure 1 shows
the global values associated to the TPS areas.
Through analysis of these results it was possible to ascertain that the consistency
ration (CR) for the resultant matrixes is within the limit recommended by Saaty (1990;
1991). Hence, the results found through the AHP method are significant.

In decreasing order to evaluate the degree of importance given by the intensity of the
vectors found, the weights are as follows: the departure lounge (0.25), access area
(parking and curb) (0.16), concessions areas (0.13) and lobby (0.13). Among the
indicators listed by area, time spent in the check-in line (0.59) stood out – with a
priority vector of greatest intensity – and the comfort of the departure lounge (0.57).
Figure 1: Hierarchical Structure w ith Global Values associated to the Degree

of I mportance
CR first level = 0.011

CR parking = 0.017
CR curb = 0.000
CR lobby = 0.011
CR check- in = 0.000
CR departure lounge = 0.000
CR concessions= 0.000
The check-in and departure lounge areas were given the highest values; that is, both
areas jointly represent 58% of the degree of global importance for TPS. This is why
the next step is to analyze whether the qualitative characteristics (income, age, travel
frequency, and reason for travel) have an influence in passengers’ decision making as
to the degree of importance.
Firstly, the percentages “greater importance” and “equal importance” among the
indicators were observed in terms of how significant they were. This analysis was
necessary, as the intention was to demonstrate whether there is a dependency
association in the results of these observations and the passenger’s qualitative
characteristics. Finally, the “Chi-Squared Method” was used by means of the
independence test, for the composition of the final results.

5. RESULTS OF THE STATI STI CAL ANALYSI S OF THE DEGREE OF
I MPORTANCE
Figure 2 indicates the percentage importance for the indicators for the check-in area.
The time taken in processing the line was given the greatest importance in regard to
airline service. However, this percentage was very close equivalence for other services
in terms of importance.
Figure 2: Percentage I mportance for I ndicators for the Check- in Area
CHECK IN
41.85%
40.74%
17.41%
time equivalence service
Table 4: First Hypothesis Test – Check- in area

I s Z significant for c =
Number of Passengers 5% ?
A M B n ( A+ B) m ( P1+ P2) Z* Result
CH ECK I N time equivalence service
113 110 47 270 160 0.407 0.592 -6.192 Reject H0
Table 5: Second Hypothesis Test – Check- in area

Number of Passengers
5% ?
A M B n P1 P2 Z* Result
CHECK I N time equivalence service
113 110 47 160 0.706 0.294 11.455 Reject H0
In which:
A: Nº of interviewees that consider the first element of the binary comparison

preferable to the second;
M: Nº of interviewees that consider the two components to be equivalent in
importance;
B: Nº of interviewees that consider the second element of the binary comparison
preferable to the first;
n: Set of interviewees formed by the sum of A and B;

Tables 4 and 5 present the tests for the hypotheses that prove the significant
difference in the percentages found for the check-in area. Although the percentage for
equivalence of importance (40.74%) is close to the percentage for greater importance
in the time spent in the line (41.39%), the results indicate that there is a statistical
difference between them, as shown in Table 4. The results also show that there is a
statistical difference between the importance of the indicators for time of processing
the line and the service provided by the airline, where the former was considered more
important than the latter, as shown in Table 5.
Figure 3 indicates the percentage of importance for the indicators for the departure
lounge area. Most interviewees attributed greater importance to comfort in relation to
the service offered by the airline’s staff.
Figure 3: Percentage of I mportance for the I ndicators for the Departure

Lounge Area
44.81% DEPARTURE LOUNGE
29.26%
25.93%
comfort equivalence service
Table 6: First Hypothesis Test – Departure lounge area
Number of Passengers 5% ?
A M B n ( A+ B) m ( P1+ P2) Z* Results
DEPARTURE
LOUNGE
121 79 70 270 191 0.293 0.707 -14.982 Reject H0
Tables 6 and 7 present the hypotheses tests that proved the significant difference in
the percentages found for the departure lounge area. The results indicate that there is

a statistical difference between the importance of the indicators for overall comfort and
overall service, where the former was considered more important than the latter.
Table 7: Second Hypothesis Test – Departure lounge area

Number of Passengers
5% ?
A M B n ( A+ B) m Z* Results
DEPARTURE
LOUNGE
121 79 70 191 0.634 0.366 7.658 Reject H0
In which:
A: Nº of interviewees that consider the first element of the binary comparison

preferable to the second;
M: Nº of interviewees that consider the two component to be of equivalent
importance ;
B: Nº of interviewees that consider the second element of the binary comparison
preferable to the first;
n: Set of interviewees formed by the sum of A and B.
6. RESULTS OF THE I NDEPENDENCE TEST

This topic covers the following results:
(i) Comparative analysis of the degree of importance given by passengers according to
qualitative characteristics: household income, age, reason for traveling, and
frequency of travel. For this analysis the AHP method was used, grouping
passengers’ opinions into classes in a certain characteristic. These results are
presented in Figures 4 to 7.
(ii) Next, by way of independence test contingency tables a relation between the
qualitative characteristics and opinion given on the degree of importance by the
passengers was looked for. The opinion on the degree of importance was divided
into three classes: passengers who attributed greater importance to any degree of
an indicator x in regard to another, y, passengers who attributed the same
importance to the two indicators, and passengers who attributed less importance
to the indicator x in regard to another, y. These results are presented in Tables 10
to 14.

6.1 VARI ABLES: I NCOME AND DEGREE OF I MPORTANCE
Figure 4 presents the differences between the intensity of the priority vectors among
the three classes of income – income up to US$ 40,000/ year, income between US$
40,000 and 80,000/ year, and income above US$ 80,000/ year.
At the check in, the group of people with an income above US$ 80,000/ year gave
greater importance to the indicator for processing time, while the group of people with
an income of up to US$ 40,000/year gave greater importance to service. In the
departure lounge the group of people with an income above US$ 80,000/year gave
greater importance to the indicator comfort; while the group of people with an income
of up to US$ 40,000/ year gave greatest importance to service.
Figure 4: Preferences related to I ncome
Given the preference among passengers according to the classes designated for
household income, the Chi-Squared Method was used to ascertain whether this
variable influenced passengers’ opinion. To such an end, contingency tables were
drawn up with the expected and observed values – Tables 8 and 9 – for the check-in
and departure lounge areas. All the results in both tables accept the hypothesis H0
concluding that passenger income does not influence opinion on the degree of
importance given to indicators for the check-in and departure lounge.

Table 8: Expected values based on the I ndependence Hypothesis
Check-in Area
Time in line vs. Service at the counter – Check-in area
Score Frequency
1
I ncome/ year less important equal 2 more important 3
Total %
Up to US$ 40,000 42 57 22 121 0.45
Expected value 50.64 49.30 21.06
e4 partial 1.47 1.20 0.04
Betw een US$ 40 and 80,000 35 33 14 82 0.30
e4 partial 0.01 0.00 0.01
More than US$ 80,000 36 20 11 67 0.25
e4 partial 2.26 1.95 0.04
e4 Total= 6.99
Total 113 110 47 270
e4 Tabled g= 5% = 9.48 e GL = 4
1
wait time é less important than the service at the counter
2
wait time and service at the counter are of equal importance
3
wait time is more important than the service at the counter

Departure lounge area
Comfort vs. Service – Departure lounge area

Score Frequency
I ncome/ year less important 1 equal 2 more important 3
Total %
Up to US$ 40,000 37 36 48 121 0.45
e4 partial 1.01 0.01 0.71
Betw een US$ 40 and 80,000 22 21 39 82 0.30
e4 partial
0.03 0.37 0.14
More than US$ 80,000 11 22 34 67 0.25
e4 partial
2.34 0.29 0.53
Total 70 79 121 270
e4 Total= 5.43
e4 Tabled g= 5% = 9.48 e GL = 4
1 comfort is less important than the service at the counter
2 comfort and service at the counter are of equal importance
3 comfort is more important than the service at the counter
6.2 VARI ABLES: AGE AND DEGREE OF I MPORTANCE

Figure 5 presents the differences between the intensity of priority vectors among the
age ranges (up to 30, between 30 and 50, and above 50 years old). People over 50
years old attributed greater importance to the indicators for wait time, comfort at the

check-in and departure lounge, respectively, so showing an inverse relation with the
group of people aged up to 30 years old.
Figure 5: Preferences related to Age
Given the preference among passengers according to the classes designated for age,
the Chi-Squared Method was used to ascertain whether this variable influenced
passengers’ opinion. To such an end, contingency tables were drawn up with the
expected and observed values – Tables 10 and 11 – for the check-in a departure
lounge areas.

Check- in area
Wait time vs. service at the counter – Check-in area
Score Frequency – check-in
Age less important 1 equal 2 more important 3 Total %
Up to 30 30 51 17 98 0.42
e4 partial
1.45 1.58 0.04
e4 partial corrected 1.26 1.40 0.01
Between 30 and 50 46 50 20 116 0.49
e4 partial
0.07 0.01 0.06
Over 50 14 2 6 22 0.09
e4 partial
3.75 6.02 0.99
e4 Total= 13.97 e e4 Yates correction = 11.64

Total 90 103 43 236
e4 Tabled g= 5% = 9.48 e GL = 4
1
wait time é less important than the service at the counter
2
3

Comfort vs. service – Departure lounge area

Score Frequency – Departure Lounge
Age less important 1 equal 2 more important 3
Total %
Up to 30 30 28 40 98 0.42
e4 partial
0.17 0.18 0.00
Between 30 and 50 46 50 20 116 0.49
e4 partial 0.07 0.01 0.06
Over 50 4 6 12 22 0.09
e4 partial
0.81 0.10 1.04
e4 Total= 2.77
Total 67 73 96 236
e4 Tabled g= 5% = 9.48 e GL = 4
1
comfort is less important than the service at the counter
2
comfort and service at the counter are of equal importance
3
comfort is more important than the service at the counter
The result presented in Table 10 for the check-in area rejects the nullity hypothesis
and accepts hypothesis H1. That is, it concludes that age interferes in passengers’
opinion about the degree of importance of the indicators wait time and service at the
counter. For the departure lounge area the result in Table 11 accepts the hypothesis
H0, where we can conclude that passenger age does not influence opinion about the
degree of importance given to the indicators for the departure lounge.
6.3 VARI ABLES: REASON FOR TRAVELI NG AND DEGREE OF I MPORTANCE

Figure 6 presents the differences between the intensity of priority vectors for the
reason for travelling (business, pleasure, and family). Business travelers gave greater
importance to the indicator wait time at the check-in, unlike the other two classes, who
gave greater importance to service at the counter.
In the departure lounge area, while the passengers travelling for family reasons gave
greater importance to the indicator service, others, traveling for business and pleasure,
preferred comfort. Given the preference among passengers according to the classes
designated for the reason for travelling, the Chi-Squared Method was used to ascertain
whether this variable influenced passengers’ opinion. To such an end, contingency
tables were drawn up with the expected and observed values – Tables 12 and 13 – for
the check-in and departure lounge areas.

Figure 6: Preferences related to Reason for Travelling

Check- in area
Wait time vs. Service at the counter – Check- in area

Reason for Travelling less important 1 equal 2 more important 3
Total %
Business 53 27 21 101 0.47
e4 partial 2.21 3.93 0.39
e4partial corrected 1.99 3.63 0.26
Pleasure 32 42 14 88 0.41
e4 partial 0.85 1.69 0.24
Family 7 15 4 26 0.12
e4 partial 1.53 2.31 0.11
e4 Total= 13.26 e e4 Yates correction = 11.24

Total 92 84 39 215
e4 Tabled g= 5% = 9.48 e GL = 4
1
wait time is less important than the service at the counter
2
3
The result presented in Table 12 for the check-in area rejects the nullity hypothesis
and accepts hypothesis H1. That is it concludes that the reason for traveling – Business,
Pleasure or Family - interferes in passengers’ opinion about the degree of importance
of the indicators wait time and service at the counter.

For the departure lounge area the result in Table 13 accepts the hypothesis H0, where
we can conclude that the passenger’s reason for traveling does not influence the
opinion about the degree of importance given to indicators for the departure lounge.


Score Frequency – DEPARTURE LOUNGE
Reason for Travelling less important 1 equal 2 more important 3
Total %
Business 28 25 48 101 0.47
e4 partial
0.18 0.85 0.19
Pleasure 18 29 41 88 0.41
e4 partial
0.90 0.30 0.07
Family 7 15 4 26 0.12
e4 partial
1.53 2.31 0.11
e4 Total= 5.82
Total 55 64 96 215
e4 Tabled g= 5% = 9.48 e GL = 4
1
2
comfort and service at the counter are of equal importance
3
6.4 VARI ABLES: FREQUENCY OF TRAVEL AND DEGREE OF I MPORTANCE

Figure 7 presents the differences between the intensity of priority vectors among the
frequency of travel (1 time/ year, 2 to 6 times/year and more than 6 times/ year).
It was noted that in the check-in area the group that travels only 1 time/ year gave
greater importance to service, unlike the other groups. In the departure lounge area it
was noted that people who travelled more than 6 times/ year gave greater importance
to the indicator comfort, unlike those who travelled only 1 time/year. Such a difference
could be explained by the greater demands made by passengers that travel more
frequently, as they spend longer inside the terminal.
Given the preference among passengers according to the classes designated for age,
the Chi-Squared Method was used to ascertain whether this variable influenced
passengers’ opinion. To such an end, contingency tables were drawn up with the
expected and observed values – Tables 14 and 15 – for the check-in and departure
lounge areas.

Figure 7: Preferences related to Frequency of Travel

Check- in area
Wait time vs. Service at the counter – Check- in area

Frequency of travel less important 1 equal 2 more important 3 Total %
1x/ year 18 32 13 63 0.23
e4 partial
2.58 1.51 0.36
From 2 to 6x/ year. 71 63 25 159 0.59
e4 partial
0.35 0.06 0.28
More than 6x/ year 23 15 9 47 0.17
e4 partial
0.60 0.93 0.08
e4 Total = 6.74
Total 112 110 47 269
e4 Tabled g= 5% = 9.48 e GL = 4
1
wait time is less important than the service at the counter
2
wait time tem equal importance than the service at the counter
3
The result indicated in Table 14 for the check-in area accepts hypothesis H0, where we
can conclude that frequency of travel for passengers at São Paulo / Guarulhos
International Airport does not influence opinion on the degree of importance for the
indicators wait time and service at the counter.
For the departure lounge area, the result found in Table 15 rejects the nullity
hypothesis and accepts hypothesis H1. That is, it concludes that frequency of travel at
the airport studied does interfere in passengers’ opinion on the degree of importance
of the indicators wait time and service at the counter.


Score Frequency – Departure Lounge
Frequency of travel less important 1 equal 2 more important 3
Total %
1x/ year 26 15 22 63 0.23
e4 partial
5.63 0.66 1.33
From 2 to 6x/ year. 35 50 74 159 0.59
e4 partial
0.98 0.23 0.13
More than 6x/ year 9 14 24 47 0.17
e4 partial
0.85 0.00 0.44
e4 Total= 10.26
Total 70 79 120 269
e4 Tabled g= 5% = 9.48 e GL = 4
1
2
comfort is of equal importance to the service at the counter
3
7. CONCLUSI ONS
The passengers’ opinions on the degree of importance of the components are required
in order to be able to prioritize services. Furthermore, it has become necessary to get
information on the quality of the services and/ or map the profile of the passengers
interviewed in order to contribute to the management of the airport as regards
decision making.
There have been many studies that have reported the relationship that exists between
individual characteristics and the perception of passengers about the degree of
importance or about the quality of the services at an airport. However, these studies
have not statistically proven whether this hypothesis is significant in their analyses.
This proof could make a big difference when resources are limited or if a new airport
terminal is being planned. Therefore, knowing that individual characteristics influence
passengers’ perception contributes more precisely to airport planning.
Unlike other studies, this article has presented, in a pioneering form, a qualitative
analysis of the relationship between the passengers’ profiles and their perception of
the airport terminal. The results obtained have made it possible to ascertain whether
there was dependency or independency between the individual characteristics of the

passengers and their perception of the terminal. This made it possible to ascertain that
“Age” and “Reason for Travelling” influence passengers’ perceptions of the check-in
area and that “Frequency of travel” influenced perception of the departure lounge
area. The final results also indicate that the check-in and departure lounge were the
most important areas in the airport terminal in the passengers’ opinion.
Finally, it can be said that this kind of analysis can achieve great results in airport
planning projects which are designed to direct their resources to a certain passenger
audience or to attract potential clients with a certain profile. We suggest that airport
operators develop this kind of analysis periodically, since variations on the competitive
scenario, economic development, and airport passengers´ profile might have an
important influence on the passenger perceptions. However, the methodology provided
in this paper is robust and valid under different scenarios.
REFERENCES
‚ Ashford, N.; Stanton, H. P. M.; Moore, C. A. (1997), Airport Operations. New York:
McGraw-Hill, 2nd. Ed.
‚ Airports Council I nternational (2000), Quality of service at airports: standards &
measurements. Geneva, Switzerland: ACI World Headquarters.
‚ Bandeira, M. C. G. da S. P. (2008) Análise do nível de serviço em terminais de
passageiros aeroportuários. Tese de Mestrado. Instituto Tecnológico de
Aeronáutica, São José dos Campos, pp.134.
‚ Bandeira, M. C. G. da S. P. ; Correia, A.R. ; Wirasinghe, S. C. (2007), Degree of
importance of airport passenger terminal components and their attributes. In: Air
Transport Research Society, Annual World Conference, Berkeley, CA, v.1, pp.1-13.
‚ Brink, M.; Maddison, D. (1975), Identification and measurement of capacity levels
of service of landside elements of the airport. In Airport Landside Capacity.
Transportation Research Board, Washington, D.C., TRB, Special Report 159: pp.92-
111.
‚ Correia, A. R. (2009), Evaluation of level of service at airport passenger terminals.
Koln: Lap Lambert, 1st. Ed.
‚ Gomes. L. F. A. M.; Araya, M. C. G.; Carignano, C. (2004), Tomada de decisões em
cenários complexos: introdução aos métodos discretos do apoio multicritério de
decisão. São Paulo: Thomson.

‚ Graham, A. (2008), Managing Airports: An International Perspective. Oxford:
Elsevier, 3rd. Ed.
‚ Lee, S. C.; Kim, B. J. (2003), Development of passenger service quality index for
inchon international airport. In: Annual Meeting. Proceedings of 82nd Annual
Meeting of the Transportation Research Board, Washington D.C., TRB, USA.
‚ Müller, C.; Gosling, G. D. (1991) A framework for evaluating level of service for
airport terminals. Transportation Planning and Technology, 16: pp.45-61.
‚ Ndoh, N. N.; Ashford, N. J. (1994), Evaluation of transportation level of service
using fuzzy sets. Transportation Research Record 1461, National Research Council,
Washington, D.C., TRB, pp.31-37.
‚ Omer, K. F.; Khan, A. M. (1988), Airport landside level of service estimation: utility
theoretic approach. Transportation Research Record 1199, TRB, National Research
Council, Washington D.C., TRB, pp.33-40.
‚ Saaty, T. L. (1991), Método de análise hierárquica. São Paulo: McGraw-Hill, pp.
367.
‚ Saaty, T.L. (1980), The Analytic Hierarchy Process. New York: McGraw-Hill.
‚ Seneviratne, P. N.; Martel, N. (1991), Variables influencing performance of air
terminal buildings. Transportation Planning and Technology, 16 (1): pp.3-28.
‚ Spiegel, Murray R. (1972), Estatística. São Paulo: McGraw-Hill, pp.580.

Business Relations Between The Low Cost

Uploaded by

Copyright:

Available Formats

Business Relations Between The Low Cost

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Business Relations Between The Low Cost

Uploaded by

Copyright:

Available Formats

The Journal of Air Transport Studies (JATS – ISSN: 1791-6771) is a peer reviewed journal

Journal of Air Transport Studies (JATS)

Volume 3, Number 1, January 2012, ISSN: 1791-6771.

‚ Airbus (2003), Global Market Forecasts 2003-2022, Toulouse: Airbus.

Full Research Papers

2. FLYER'S CORNER: SOME SIMPLE TRUTHS ABOUT SCHEDULED AIR

3. A COMPLEX SYSTEM APPROACH IN MODELING AIRSPACE CONGESTION

4. BUSINESS RELATIONS BETWEEN THE LOW COST CARRIERS AND AIRPORTS AS A

5. QUALITATIVE ANALYSI S OF THE RELATIONSHIP BETWEEN THE PROFILE OF

Dr Andreas Papatheodorou, Editor-in-Chief

Alongside administrative measures to handle increasing capacity, airport expansion

Journal of Air Transport Studies, volume 3, Issue 1, 2012 Page 2

Ü First, governmental ownership induces a situation in which a maj ority of airports do

2. THE EFFECTS OF PEAK-LOAD PRI CI NG ON AI RLI NE COSTS

Table 1: Operating Airline Cost Shares for I nternational Scheduled Operations

‚ fuel and oil

‚ airport and en-route charges

‚ other passenger services

Source: AEA (2007), ATA (2007)

Journal of Air Transport Studies, volume 3, Issue 1, 2012 Page 5

2. the significance of airport charges for the carrier;

3. the level of charges at the airport.

Tokyo NRT 8,649 7,537 16,186 53 : 47

Journal of Air Transport Studies, volume 3, Issue 1, 2012 Page 8

3. AI RLI NE REACTI ONS TO PEAK-LOAD CHARGES

3.1 I NCREASE I N AI RCRAFT SI ZE/ REDUCTI ON OF FLI GHT FREQUENCY

Figure 1: Economies of Size & Fill in Relation to I ncreased Aircraft Size

seat capacity (K)

Source: Own illustration, based on Button (1982), p. 79.

1. Airport schedules: Existing schedules mean frequency changes can only be

3.2 TEMPORAL AND SPATI AL RELOCATI ON EFFECTS

Similarly, in the case of frequency reductions, the ability of rescheduling to be implemented

Another alternative mainly applicable to low-cost carriers is the relocation of operations to

3.3 PRI CI NG STRATEGI ES

Journal of Air Transport Studies, volume 3, Issue 1, 2012 Page 14

250 Price LH - Off-Peak

Variable Arithmetic Average Coefficient of Variation

Journal of Air Transport Studies, volume 3, Issue 1, 2012 Page 16

Table 5: Estimation results for price setting behaviour

Equation (TXL-FRA) AB Off-Peak LH Off-Peak AB Peak LH Peak

Journal of Air Transport Studies, volume 3, Issue 1, 2012 Page 18

Furthermore, it is still an open question as to whether carriers oppose peak-load pricing

Journal of Air Transport Studies, volume 3, Issue 1, 2012 Page 22

Journal of Air Transport Studies, volume 3, Issue 1, 2012 Page 23

Keywords: Airline Customers, Satisfaction, Complaints, Loyalty

A company’s typical response to a crisis is to minimise costs, followed by some efforts to

Journal of Air Transport Studies, volume 3, Issue 1, 2012 Page 25

Journal of Air Transport Studies, volume 3, Issue 1, 2012 Page 26

3.1 TRAVEL BEHAVI OR AND ECONOMI C CRI SI S

Journal of Air Transport Studies, volume 3, Issue 1, 2012 Page 27

Figure 2: I nformation about air tickets

Journal of Air Transport Studies, volume 3, Issue 1, 2012 Page 28

3.3 AI RLI NE ADVERTI SEMENT AND I NFORMATI ON POLI CY

Figure 4: Advertisements' effects

Journal of Air Transport Studies, volume 3, Issue 1, 2012 Page 29

Figure 6: FFP member’s willingness to receive advertisements

3.4 CRI TERI A FOR SELECTI NG AN AI RLI NE TO FLY

Journal of Air Transport Studies, volume 3, Issue 1, 2012 Page 30

Figure 7: Ranking of factors affecting satisfaction