DP 2004 – 05
Price Competition amongst Low Cost Carriers
Cristina Barbot
November 2004
CETE − Centro de Estudos de Economia Industrial, do Trabalho e da Empresa
Research Center on Industrial, Labour and Managerial Economics
Research Center supported by Fundação para a Ciência e a Tecnologia, Programa de Financiamento
Plurianual through the Programa Operacional Ciência, Tecnologia e Inovação (POCTI) of the Quadro
Comunitário de Apoio III, which is financed by FEDER and Portuguese funds.
Faculdade de Economia, Universidade do Porto
http://www.fep.up.pt /investigacao/cete/papers/index.html
Price competition amongst Low Cost Carriers
Cristina Barbot1
CETE/ Faculty of Economics of Porto
Abstract
This paper addresses two questions on low cost carriers (LCC’s) price
competition: 1) if they reply to price changes of a full service carrier (FSC) and 2) if
they compete amongst themselves and how that competition works. We present a
theoretical model combining vertical and horizontal differentiation. Then, an empirical
test is performed for the Paris-Milan route, where three LCC’s and one FSC operate.
The empirical study confirms some of the results of the theoretical model.
Namely, that there is a strong price competition amongst LCC’s and almost always in
the way predicted by the model. In contrast, results suggest that their competition with
FSC’s is limited to an entry adjustment, and they ignore post entry price changes by the
flag carrier.
JEL Codes: L93, L13
Keywords: low cost carriers, competition, horizontal differentiation, vertical
differentiation
1
Postal Address: Professor Cristina Barbot Faculdade de Economia Rua Dr. Roberto Frias
4200 PORTO PORTUGAL. Email Address: [email protected]. Phone number:
+351225571218. Fax number. +351225505050.
1
Price competition amongst Low Cost Carriers
Cristina Barbot
1. Introductory notes on airlines competition
One of the most striking features of aviation industry in the beginning of this
century was the availability for consumers of a new concept of flying. Low cost carriers
(LCC’s) opened a totally new product: no frills, no food, no drinks, no spacious seats,
no travel agencies bookings, but a very low price. Southwestern Airlines was a pioneer
of this concept in the United States. In Europe, the first one was Ryanair, who started
operating flights in the route Dublin-London in 1986. In their own words, “we go after
the big guys for a slice of the action and end up smashing the Aer Lingus / British
Airways high-fare cartel on the Dublin-London route.” (Ryanair, 2004). LCC’s were
convinced not only that flying can be less costly, but also that an effective competition
in the sector would result in an advantage for them.
EasyJet followed in 1995 with two routes from London Luton to Scotland, and
since then there has been a proliferation of LCC’s, not only in Europe, but also in Asia
and Latin America.
As Ryanair forecasted in 1986, LCC’s have shaken the aviation market. Full
Service Carriers (FSC’s) reacted by cutting costs and changing networks, mainly
directing them to segments where LCC’s are absent, like transcontinental routes. But
they also had to adjust their fares on European routes.
2
There is a considerable literature on FSC’s reaction to LCC’s entry, and already
a few empirical works that account for this reaction. Dresner et al. (1996) found that
FSC’s reacted to the entry of Southwestern Airlines with a fall in their prices. Windle
and Dresder (1999) also found that an LCC entry made Delta lower its fares on the
same routes. Alderighi et al. (2004) studied the FSC’s prices reactions to the presence of
LCC’s in the market, for several European major airlines and routes. They found that
the presence of one LCC has a negative and significant impact on FSC’s fares, and this
happens for all the existing classes. Oliveira and Huse (ATRS 2004) also found, for
Brasil, that FSC’s lower their price after the entry of one LCC.
The fast growing number of LCC’s means that there has been a market for all of
them. Their demand is either originated by former customers of other airlines or by new
passengers that otherwise would not fly. In most routes between two points in Europe,
FSC’s compete with LCC’s. If much has already been written on competition between
these two types of carriers, most of these studies cover, however, the reaction of FSC’s
to LCC’s entry, and not the way they adjust their fares once all of them are well
established in a particular market. But do LCC’s react to FSC’s prices, and vice versa,
once they are already competing in the market? If they do, it means that customers
watch both firms’ prices, and interchange flights between both kinds of companies. If
they do not, then passengers chose to fly or not LCC once they enter the market, and
keep loyal to the type of carrier they have chosen. This is the first question to which this
paper tries to provide an answer.
- insert Table 1-
Often there is more than one LCC in a particular route, which brings
competition between them. As far as we know, there are no theoretical or empirical
3
studies on competition amongst LCC’s. That this kind of competition is a fact, and that
they charge different prices is another fact. Table 1 shows some examples of different
fares charges by LCC’s for the same routes, and sometimes using the same airports. The
second aim of this essay is to investigate the sources and nature of competition amongst
LCC’s. If fares differ, and if information is almost costless2, then this competition must
involve some kind of product differentiation. Section 2 investigates on the nature of this
differentiation, using evidence on these firms’ strategies and product characteristics.
This provides a basis for the model presented in section 3, which tries to establish a
theoretical framework combining competition between FSC’s and LCC’s and amongst
these latter. Finally, in section 4 an empirical study tests the theoretical model and
section 5 follows with a few concluding remarks.
2. The nature of competition among LCC’s
Flights performed by different airlines are not homogeneous products. It is
obvious that there is a noticeable quality differentiation between the services provided
by LCC’s and FSC’s. This vertical differentiation element is well known and needs no
further comments. But even amongst LCC’s flights are differentiated. Each airline has
its own characteristics and some possible sources of differentiation are listed below. The
question now is to know what kind of differentiation exists, if a vertical or a horizontal
one.
Dudden (2004) divides LCC’s in two segments: lowest cost carriers, like
Ryanair, that use secondary airports and offer nothing (every extra item is purchased
during the flight), and lower cost carriers, that use large airports, EasyJet and AirBerlin
2
Internet sites provide tickets’ prices and quick flight reservation. So this is not the case for imperfect
information.
4
being included among the latter. These two segments seem to be vertically
differentiated. As justified below, we do not follow with this view, and consider only
horizontal differentiation.
There are five main sources of differentiation:
1. Airports of departure/arrival. Some airlines use secondary airports, while others
land and take off at main airports. Table 2 illustrates some characteristics of hub and
secondary airports. The latter usually involve a longer distance to city centres, and
have a much smaller number of shops, bars and restaurants than large hubs. They
are no frills, functional airports. This may account for vertical differentiation.
2. Aircrafts. LCC’s use different aircrafts, some relying exclusively on Boeing 737,
others in Airbus 319, or on the two of them, or in other equipments.
3. Free services during the flights, such as food and drinks.
3. Differentiation in the supply of services related to the flight, like providing
accommodation or car hire.
4. Passenger loyalty. Customers may prefer a particular airline, and keep flying with
it. This may happen on account of experience, reliability, or just because the airline
is national and the crew speaks the passenger’s language.
- insert Table 2-
These sources of differentiation were checked for some airlines, precisely for
those chosen for the empirical study: Ryanair, EasyJet and Volare. An inspection of
airlines main characteristics displayed in their websites shows that:
1. Ryanair uses mainly secondary and regional airports. The other two use either
hub, regional or secondary ones. Then, there is not a clear vertical differentiation,
5
though Ryanair’s flights could be considered as having a lower quality, but
depending on the route. Even between two large airports, like Malpensa and Linate,
a consumer may have a preference for a particular one, and this may be a source of
horizontal differentiation.
2. Aircrafts used by the three LCC’s for scheduled flights are different, but
approximately of the same kind and with the same capacity. There seems to be no
reason to classify an aircraft as better than another one. In what regards seat density,
Ryanair and EasyJet display the same patterns for similar aircrafts, and Volare does
not provide that information.
3. All the three airlines do not supply free meals during the flights, but sell food and
drinks, which, of course, cannot be identical.
4. They all offer the same extra services. They all organise accommodation, car hire,
car parking at the airport and travel insurance, but with agreements with different
companies.
5. Loyalty is a possible element of horizontal differentiation. When introducing their
company, they all stress the element “low fare”. But, for instance, Volare says it has
“something of Italian”, which is probably addressed to Italian customers.
Only the first source of differentiation (airports) could mean a difference in
quality, and this does not always happen and depends on the route. The other
sources show that there are clear elements of horizontal differentiation and this is
the option of this paper.
6
3. Theoretical model
The model is designed to suit the case of the empirical study, but to cover the
generality of LCC’s competition situations as well. Outputs are direct flights in a
particular route between two large European cities. Flights are operated by FSC’s and
LCC’s, with vertical differentiation between the two types of carriers.
Only one FSC operates in the high quality segment of the market. In Europe, it
is usual that two flag airlines perform these flights in code share, or with some price
agreement, so it seems adequate to reduce their number to one. In the low quality
segment a small number of LCC’s operates. Evidence suggests that this number is never
very high, and that only about two or three LCC’s serve two pairs of large cities. Of
course that sometimes there is only one carrier operating, mainly in regional routes, but
this case is not under the purpose of the paper. We then consider three airlines operating
in the low quality segment of the market, though this number could be extended to n
carriers without difficulty.
The FSC has a quality of q2, while all the three LCC’s have the same quality, q1.
To avoid excessive parameters, we set q1 =1 and q2 = a, so that a measures the quality
differential. Qualities are previously established and cannot be changed. This is
consistent because once that a carrier opts for a particular quality, many factors are
fixed, such as aircrafts, seat density, marketing and booking and changes involve a long
run decision.
There are two versions of the model. The first one is designed to analyse price
competition both between FSC’s and the LCC’s and amongst the latter. It uses a
combination of vertical and horizontal differentiation framework. Demand for flights is
derived from vertical differentiation theory, as proposed first by Gabszewicz and Thisse
7
(1979) and Shaked and Sutton (1982) and from Hotelling’s horizontal differentiation
model, in the version presented by Tirole (1988). The main theoretical issue here is to
combine both approaches.
Let the FSC’s variables be denoted by the subscript 4, and the LCC’s by the
subscripts 1,2 and 3. The letters p and y denote, respectively, the tickets prices and
demands. It is supposed that p3>p2>p1.
Initially all consumers are uniformly distributed according their willingness to
pay for quality parameter, v, along a line of length equal to the unit. Every consumer
buys one unit of the good, or one ticket. First consumers chose to fly low cost or high
quality. The consumer who is indifferent between the two has a value of v equal to vt,
such that: vtq1 –p3 = vt q2 – p4, or vt –p3 = vt a – p4.
The value vt divides the line in two segments. On the right side are passengers
who prefer high quality flights, who are the demand of the FSC, equal to 1- vt = 1-(p4p3)/(a-1). On the left side, are passengers flying low cost, whose quantity is of vt-0 =
(p4-p3)/(a-1).
Then we find the demand of each LCC. This may be done as follows. Take the
left side of the line, a segment of length vt, which contains passengers that have decided
to fly only in LCC’s but have not yet chosen which one. Re-arrange these consumers
according to their horizontal preferences in a Hotelling/Salop fashion. These
preferences are a combination of the items listed above, which can be sources of
horizontal differentiation. While re-arranging consumers, those who prefer the
characteristics of the airline with the highest price, p3, are kept on the right side of the
line. This can be done by a selection of preferences in such a way that the most
expensive airline’s demand keeps close to the FSC’s.
8
Consider, as usual, t as the unitary transportation cost, or the unitary difference
of utility between the ideal specification of each consumer’s product and the most
similar available product, and x the distance between the two.
Within this segment, a consumer characterised by the distance xm, is indifferent
between airline 2 and airline 3 if: p2-txm = p3+t (vt-xm), while the consumer indifferent
between airline 1 and 2 has a distance xk such that: p1-txk = p2+t(vt--xm -xk). This way
demands for the three LCC’s flights are computed, after substituting vt by its value as
written above:
y1 = xk – 0 = (2(1-a)p1+3(a-1)p2+(1-a-t)p3+tp4)/(4t(a-1))
y2 = xm - xk = (5(1-a)p2+2(a-1)p1+3(a-1-t)p3+tp4)/(4t(a-1))
y3 = vt -xm - xk = ( (a-1)p2+(1-a-t)p3+tp4)/(2t(a-1))
y4 = 1- vt = 1-(p4-p3)/(a-1)
In order to make the model as simple as possible, marginal costs are constant.
The FSC has a marginal cost equal to C, and the LCC’s have (lower) marginal costs of
bC, with b<1. Airlines compete in prices.
With these costs and demands, it is easy to compute profits3. Each firm
maximises its profit resulting best reply functions, which are:
p1 = (2Cb(a-1)+3(a-1)p2+(1-a-t)p3+tp4))/4(a-1)
p2= (2(a-1)p1+3(a-1-t)p3+tp4+5bC(a-1))/10(a-1)
p3=((a-1)p2+tp4+bC(a+t-1))/2(a+t-1)
p4= (C+p3)/2
3
Their expressions are omitted in the text. All calculation was developed in SW4 and may be requested
to the author.
9
As the aim of this paper is the analysis and empirical test of these best reply
functions, it is not necessary to go further. It is supposed that a Nash equilibrium exists
and is unique.
Conclusions may be summarised as follows:
1.
If t is very small, or near zero, none of the LCC’s respond to changes
in p4. If t is positive, the higher its value, the stronger their reaction.
However, the FSC always replies to changes in p3. The theoretical
answer to the first question posed above is, then, that FSC’s adjust
their prices to changes in LCC’s fares, and the inverse happens if there
is a sufficient level of horizontal differentiation. The intensification of
horizontal competition strengthens price competition with the FSC,
precisely by relaxing price competition amongst LCC’s. Notice that
∂p1/∂
∂p4 = t/(4a-1), ∂p1/∂
∂p4 = t/(10(a-1)), and ∂p1/∂
∂p4 = t/(2(a+t-1)).
The higher the degree of vertical differentiation the less LCC’s prices
react to the FSC fare.
2.
Airline 1’s price depends on all the others’. It replies positively to a
change in p2, the price of its nearest airline, and negatively to a change
in p3, as 1-a-t <0.
3.
Airline 2 reacts positively to p1 and also positively to p3 if 3(a-1-t)>0 ,
or if a-1>t. This condition means that if the degree of vertical
differentiation is higher than that of horizontal differentiation.
4.
Airline 3’s price does not respond to changes of the lowest price, p1,
but it reacts positively to changes in p2.
5.
The FSC only reacts to p3.
10
Thus if t is very small LCC’s fares do not react to the FSC’s price. A small
value of t means that it does not make much difference, for a passenger, to fly in any
other LCC instead of flying in her favourite one, but she only wants to fly low cost. In
this case the second hypothesis of section 1 is confirmed. Only the entry of a LCC has
impact on the FSC price, but there are no price adjustments afterwards. For the LCC,
when it enters the market, it watches the FSC price, but does not react to changes in this
price.
This situation can be depicted by a second version of the model. In this version
vt is kept constant, as the choice between flying LCC or FSC was already done when
the first LCC entered the market. There is a segment of length equal to vt containing all
passengers who definitively fly low cost. In this segment, LCC’s compete with
horizontal differentiation, and the hypothesis that p3>p2>p1 is kept. Along the line,
airlines are placed in the sequence: 1, 2, 3.
Considering only this segment, consumer xk is indifferent between airline 1 and
2, as p1-txk = p2+t(vt--xm -xk). Consumer xm, is indifferent between airline 2 and airline
3, as p2-txm = p3+t (vt-xm). Demands are, as before: y1 = xk – 0, y2 = xm - xk and y3 = vt
-xm - xk. Substituting xk and xm by its values there result the demands of each airline.
Constant marginal costs are now equal to c, as there is no need to differentiate them.
Maximizing profits, the LCC’s best reply functions are:
p1 = ½ c+3/4p2-1/4p3+1/4tvt
p2 = ½ c+1/5 p1+3/10 p3+1/10 tvt
p3 = ½ c+ ½ p2+ ½ tvt
These best reply functions show that:
11
1.
Airline 1’s price depends negatively on p3, and positively on p2. It
changes its price in the same direction than its neighbour, but
oppositely to the most distant carrier.
2.
Airline 2 replies positively to both of its neighbours, as it is placed in
the middle of them.
3.
Airline 3 only accounts for p2‘s changes, but changes in p1 do not
affect it.
4.
The magnitude of prices depends on t, but not the reply of any airline
to others’ changes in prices.
There is an equilibrium for these prices, that shows that effectively p3>p2>p1:
p1= c + 21/58 tvt
p2 = c + 11/29 tvt
p3 = c + 20/29 tvt
Price cost margins grow with t. This is a usual result of general theory on
horizontal differentiation and LCC’s are interested in increasing product differentiation.
This point is interesting as it shows that the existence of competition amongst LCC’s
provides an incentive for advertisement or to the implementation of other forms of
differentiation, like frequent flyer programs.
Obviously that as t approaches zero, products become homogeneous and
prices equal marginal costs. That’s the usual Bertrand result.
These margins also depend positively on the value of vt, which expresses the
total demand of LCC’s, or the market split between them and the FSC.
12
4. Empirical test
In this section the best reply functions derived in the precedent section, for
the two versions of the model, will be empirically tested.
For this purpose, the route Paris-Milan was chosen. This route is operated by
two FSC’s (Airfrance and Alitalia) but in code share, so they can be reduced to one.
There are three LCC’s on the same route: Ryanair, Volare and EasyJet. Among these
carriers there are some elements of differentiation. First, they fly from and to different
airports. Ryanair uses only secondary airports, Paris Beauvais and Milan Bergamo.
Both Volare and EasyJet fly from Orly, but the former arrives at Malpensa, while the
latter arrives at Linate.
This can introduce some vertical differentiation between
Ryanair and the other two, as it flies from and to airports that are situated far from the
cities centers. However, the model drops this element and establishes no quality
differentiation amongst the three LCC’s.
There are some elements of horizontal differentiation. These airlines use
different aircrafts4 and try to get some loyalty from customers. Volare considers itself as
a carrier “sharing the principles of other European low-cost airlines, but offering
something more: “Italianness”, that special added value resulting in comfort, efficient
personnel, first-class service, and the highest aircraft reliability.” (Volarewebsite).
EasyJet stresses safety and efficiency, stating as its mission: ”to provide our customers
with safe, good value, point to point air services; to effect and to offer a consistent and
reliable product and fares appealing to leisure and business markets on a range of
European routes.” (EasyJet website). In announcing its attributes, Ryanair is more
4
Volare’s fleet is composed of Boeing 767 and Airbus 320 and 330. EasyJet uses Boeing 737-700 and
Airbus 319. Ryanair uses Boeing 737-300 , 737-200, and 737-800
13
oriented towards the efficiency of its service, being number 1 in Europe for punctuality,
flight completions and fewest complains (Ryanair website).
There is no evidence of differentiation in other services, as they all offer the
same items. They organise accommodation, car hire, car parking at the airport and travel
insurance. Inflight service is much about the same. None of them provides free meals
but they all sell snacks and drinks during flights.
Prices for the Paris Milan route were taken from airlines’ websites. These fares
may not be the effective ones, but they are the announced prices, which seem to be the
most adequate ones to estimate best reply functions. Data on prices was collected daily
during 42 days for a flight on the 27th of October 2004. LCC’s fares are for one way
flights. As Airfrance did not change its one way fare, the prices for this company are for
return flights, with return one week afterwards.
For all the observations, EasyJet always charges the highest price and Volare’s
fare is higher than Ryanair’s. Airfrance’s tickets always cost more than the double of
EasyJet’s. This allows us to place EasyJet as airline 3, Volare as airline 2, and Ryanair
as airline 1, the cheapest one.
- insert Figure 1-
Usually, airlines increase prices as the date of booking approaches the date of
the flight. Figure 1 depicts the price growth for the three LCC’s. It is interesting to
verify that the condition p3>p2>p1 always holds, and that the difference between daily
prices does not change significantly.
14
Results for the first version of the model are presented in Table 3. The variables
FR, VA, EZ and AF are, respectively, the fares set by Ryanair, Volare, EasyJet and
Airfrance.
-insert Table 3-
Ryanair’s best reply function has the form:
p1 = Cb/2+3/4p2+(1-a-t)/ 4(a-1)p3+t/4(a-1)p4
The test shows that its prices only depend, and positively, on Volare’s.
Surprisingly, the coefficient for EasyJet’s price is not significant, suggesting that 1-a-t =
0. As 1-a<0, this goes against the basic hypothesis of the model, namely, the existence
of vertical differentiation.
Volare behaves accordingly to the model’s predictions, in what regards its best
reply to the others’ fares. The function p2= 2p1+3(a-1-t)/ (10(a-1))p3+t/(10(a-1))
p4+bC/2 is confirmed by the test for t>0 and a-1>t, showing that, for consumers, the
vertical differentiation degree is higher than the horizontal differentiation one, which is
quite plausible. R-squared is also higher for Volare, indicating that the company
changes its prices after watching the others’, and with the expected sign. The coefficient
is higher for Ryanair than for EasyJet. Besides, Volare is the only LCC that replies to
Airfrance’s prices. Notice that Airfrance’s flights are operated in code share with
Alitalia, and that’s probably the reason why only Volare watches their prices, as it may
consider another home company as its competitor.
EasyJet’s best reply function - p3=((a-1)/2(a+t-1) p2+t/2(a+t-1)p4+bC/2 also
seems to behave according to theoretical predictions, except in what concerns
Airfrance’s price. Indeed, the coefficient for Volare is positive and significant, while it
does not respond to Ryanair’s price changes.
15
Airfrance’s regression goes opposite to the model’s results. Instead of replying
only to EasyJet’s changes, it seems to respond positively only to Volare. Considerations
about them being both national carriers may be valid here.
None of the LCC’s, except Volare, seem to reply to Airfrance’s price. This may
lead us to reject the first version of the model, and accept the hypothesis that passengers
may interchange LCC’s and FSC’s flights only between national carriers, and that
companies are aware of that fact. This last hypothesis on consumers’ behaviour would
however need further confirmation.
The empirical test suggests that FSC’s and LCC’s adjust prices only initially, or
when the latter enter the market. Thus consumers choose to fly FSC or LCC according
to a certain price standard of both types of companies, but will not change their
decisions on account of price variations. Competition between FSC’s and LCC’s seems
to be characterised by price adjustments when new entrants begin to operate, but
excludes post entry replies.
The second version of the model predicts basically the same type of price
adjustments, but now keeping vt constant, or withdrawing the direct influence of p4 and
the indirect influence of p3 while defining the length of LCC’s whole demand. This
length is now considered as determined by consumers immediately after the entry of the
first LCC and does not change according to post entry price variations.
-insert Table 4-
Results concerning price competition amongst LCC’s are similar to those of the
first version of the model. Recalling the expressions of the best reply functions:
16
p1 = 0.5c+0.75p2-0.25p3+0.25tvt
p2 = 0.5c+0.2p1+0.3p3+0.1 tvt
p3 = 0.5c+ 0.5p2+ 0.5tvt
Again, Ryanair’s fare depends only on Volare’s. This may happen either
because Volare’s price is the one that has a smaller difference from Ryanair’s or
because Volare is Italian, and the arrival point is an Italian city. The coefficient is
accordingly to the predicted one and there is an improved level of significance in this
version. Anyway results suggest that Ryanair has a more independent price strategy
than the other two airlines.
Volare’s price keeps depending positively on its two rivals, and the level of
significance improves as well. Coefficients for both Ryanair’s and EasyJet’s prices are
not very different from those predicted by the model. This airline’s regression has also
the highest R-square.
EasyJet’s regression displays a better adequacy in this version, with higher F
statistic and level of significance of Volare’s price coefficient. However this is much
higher than the model’s prediction.
As a whole, the empirical study results are satisfactory and show clearly that
there is price competition amongst LCC’s, and that this competition does not differ
much from the model’s predictions.
5. Concluding remarks
This paper intends to investigate if there is price competition between LCC’s
and FSC’s, and amongst LCC’s, by means of a theoretical model that is tested for the
17
route Paris-Milan. Results indicate that LCC’s reply to price changes of their low cost
rivals. However, price competition with the FSC seems to happen only as an entry
adjustment from the latter without further reply to post entry price changes.
A clear limitation of the study is that it only tests one particular route in a
particular time. Further empirical investigation may confirm results as well as some of
the insights provided by the results. One of these insights is the important role played by
nationality. The fact that Volare is Italian and that the arrival point is Milan seems to be
quite meaningful. First, because Airfrance only adjusts its fare to Volare’s price
changes. Second, because the EasyJet’s dependence on Volare is much stronger than
that predicted by the model. Third, because amongst LCC’s Volare has no rivals with
the nationality of the departing or arriving town and its regression behaves quite
accordingly to the model.
The other insight is the fact that Ryanair’s price strategy is much more
independent than other LCC’s.
18
References:
Alderighi, M., Cento, A., Nukamp, P. and P.Rietveld, The entry of low cost
airlines: price competition on the European airline market, 2004, the 9th ATRS
conference, Istanbul, July.
Dresner, M., Lin, J. and R. Windle, 1996, The impact of low cost carriers on
airport and route competition, Journal of Transport Economics and Policy 30, 309-328.
Dudden, J., 2004, Hubs under siege- The effects of low cost carriers market
entry on hub and spoke carriers, the 9th ATRS conference, Istanbul, July.
Gabszewicz, J. and J.F. Thisse, 1979, Price competition, quality and income
disparities, Journal of Economic Theory 20, 340-359.
Oliveira, A. and C. Huse, 2004, Localised competitive advantage and price
reactions to low cost carrier entry in the Brazilian airline industry, the 9th ATRS
conference, Istanbul, July.
Shaked, A. and J. Sutton, 1982, Relaxing price competition through product
differentiation, Review of Economic Studies XLIX, 3-13.
Tirole, J. 1988, The theory of industrial organization, (The MIT Press,
Cambridge, Mass.).
Windle, R. and M. Dresder, 1999, Competitive responses to low cost carrier
entry, Transportation Research- Part E 35, 59-75.
19
Figure 1: prices of the three low cost airlines
350
300
250
200
150
100
50
0
Ryanair
Volare
EasyJet
20
Table 1: Prices for some routes operated by low cost carriers
Routes and airlines
Fares
German Wings, Cologne Bonn-London Stanstead
EasyJet, Cologne Bonn-London Gatwick
67
41,99
EasyJet, London Stansted-Glasgow Preswick
Ryanair, London Stansted-Glasgow Preswick
38,78
21,54
Ryanair, Frankfurt Hahn-Venice Treviso
Volare, Frankfurt Hahn-Venice Marco Polo
69,99
9,99
Prices, in euros, collected on October 21st, ar 11:00 am,
for flights operated on October 28th.
Source: Airlines' websites
21
Table 2: Some characteristics of airports differentiation
Airport
Brussels Charleroi
Brussels International
Paris Beauvais
Paris Orly
London Stanstead
London Heathrow
Milan Bergamo
Milan Linate
Milan Malpensa
Frankfurt Hahn
Frankfurt International
distance to
the city, in Km
number of bars
and restaurants
86
12
90
14
48
24
45
7
45
125
12
1
19
1
25
9
60
4
11
74
12
69
Data collected on September 16 from airports' websites
22
Table 3: Regressions results, first version
Dependent variable: FR
Coefficient Probability
VA
0,82
0,009
EZ
-0,054
0,664
AF
-0,709
0,692
C
-324,2
0,376
R-square: 0,204
T Statistic (prob)
DW: 2,17
0,0036
Dependent variable: VA
Coefficient Probability
FR
0,207
0,009
EZ
0,166
0,006
AF
1,9
0,029
221,05
0,228
C
R-square: 0,498
T Statistic (prob)
DW: 1,91
0,0000
Dependent variable: EZ
Coefficient Probability
FR
-0,09
0,664
VA
1,13
0,006
AF
2,56
0,272
C
638,9
0,151
R-square: 0,335
T Statistic (prob)
DW: 1,96
0,0036
Dependent variable: AF
Coefficient Probability
FR
-0,006
0,692
VA
0,064
0,029
EZ
0,013
0,272
C
101,6
0,001
R-square: 0,276
T Statistic (prob)
DW:1,22
0,007
23
Table 4: Regressions results, second version
Dependent variable: FR
Coefficient Probability
VA
0,778
0,007
EZ
-0,064
0,601
C
-397,8
0,204
R-square: 0,200
T Statistic (prob)
DW: 2,17
0,014
Dependent variable: VA
Coefficient Probability
FR
0,223
0,007
EZ
0,217
0,000
C
471,8
0,003
R-square: 0,428
T Statistic (prob)
DW: 1,87
0,0000
Dependent variable: EZ
Coefficient Probability
FR
-0,113
0,601
VA
1,336
0,000
C
975,4
0,016
R-square: 0,312
T Statistic (prob)
DW: 1,92
0,0008
24