You are on page 1of 13

Some Speculations and Empirical Evidence on the Oligopolistic Behaviour of Competing LowCost Airlines Author(s): D. E.

Pitfield Reviewed work(s): Source: Journal of Transport Economics and Policy, Vol. 39, No. 3, The Economics of Cost Recovery in Transport (Sep., 2005), pp. 379-390 Published by: University of Bath and The London School of Economics and Political Science Stable URL: . Accessed: 12/03/2012 07:40
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact

The London School of Economics and Political Science and University of Bath are collaborating with JSTOR to digitize, preserve and extend access to Journal of Transport Economics and Policy.


of Transport


and Policy,


39, Part

3, September





Speculations Competing

and Empirical Low-Cost


on the Oligopolistic

Behaviour of

D. E. Pitfield

Address for correspondence: Dr D. E. Pitfield, Transport Studies Group, Department of Civil and Building Engineering, Loughborough University, Loughborough, Leicester shire LEU 3TU, UK;

Abstract This paper reviews the theory of cost in which that recovery European applies. and with oligopoly low-cost airlines with more a view manage than one to advancing yield, some

as to the way judgements the market upon morphology their this will draws that


depending in direct operator

(at the same airport) competition

manage where paper seem between yield be in a manner affected

or indirect (at an adjacent airport) competition will

on there

that is traditional sector for their of the industry, but of the competitors. This by the yield management process of airline evidence there is direct It would price setting when competition. is evidence of price and more of a strong correlation leadership generally of the differentiated product that the airlines offer in competition.

the fares


of receipt

of final






of Transport


and Policy


39, Part

This paper is concerned with

the full cost recovery by airlines through environments. Airlines competitive the that

of yield management no direct, or have

in differing

can in unregulated markets indirect, competition cost to contribute, in part or in above average variable generate surpluses barriers total costs as they exploit monopoly power. Where full, to average or the market to entry break down, is in certain the contestable, respects in its pricing is and if the competition airline has less freedom strategy, actual, then the market

well be diminished for all participating may Full cost recovery may not their differentiated carriers, despite products.1 sources to other of revenue, and the recourse be possible such as cross from monopoly routes, may well be important. subsidy The cient recent revenues are financial around to ent markets in a variety of differ of many airlines performance to generate that their abilities the world suffi indicate recover have been diminished. The main full costs

exceptions US market

in the such as Southwest of low-cost carriers, in Europe, that have consistently been profitable. and Ryanair at the interactions between There has been a succession of studies looking at the and in particular and the established this type of operator airlines, a number 'Southwest low cost In the advantage Effect' carriers short of (Vowles, compete term numerous the low-cost in the US, 2001) with each other. studies have carriers but little at situations where

to the comparative pointed The in current market conditions.

of good manage this is purely a matter questions longer-term on the part of these carriers in the prevailing market ment practices context, or whether to the is superior of the low-cost carrier the generic model a market if the and from whether, stability perspective, 'legacy' carriers, are whether lost-cost It but

carrier model to

is difficult some



to prove in the future. robust is stable, it is going a full empirical of these analysis subjects, set within the context of the theory of the firm, can be

2.0 Pricing and Cost Recovery

The low-cost airline model becoming the is not new but has dominant evolved over is rapidly
1 Schnell been

to market


forty years and for short-haul

(2004) offers

a study of some of the barriers


that still exist


that have






on Oligopolistic


of Competing




services. of




adopted by if price first principles, however, costs are decreasing then full cost run, on trading. all fixed This (1946), the assumption that revenues may Although costs can be covered. scenario, is more on which

the impact on the profitability and Fagan (2005) discuss to the principles from their differing adherence first Airlines. from the US domestic Southwest carrier, Starting cost when is set at marginal average will not occur and in the long recovery are rational, the firm will cease investors be sufficient to cover variable costs, not


yield management cost than marginal a monopoly perhaps oligopoly revenues to cover

in its modern form was first presented by Coase to apply to a carrier when there is some form of likely a particular a monopoly is held, route.2 When position some consumers more to charge the airline will allow for their flight, so that all costs can be recovered and

a duopoly or (Hanlon, 1994). When profit gained the market is fragmented; is the situation, demand and average to each participating airline are reduced and price is less likely

costs are reduced. The ability of the airline in such unless so that some consumers more to manage than pay competition yield cost is lessened. In this case, revenues may be augmented marginal by routes and from other income sources from other monopoly cross-subsidy costs such as the sale of on-board

food and no

and drink,

car parking, cost

and hotel curve

booking it is is



the usual where

traditional consumer costs

U-shaped surplus

analysis, if the airline

worth following Hanlon (1994)who supposes thatmarginal cost ishorizontal

up to full capacity vertical is available

practising perfect price discrimination

becomes to reflect airline

(or something

like it) and then

to put on an extra flight. This implies cost are also horizontal that average cost and average variable and pricing cost would below marginal raise similar issues to those discussed above. some passengers Hanlon how yield management to enables demonstrates pay less than the profit-maximising cost. and marginal It seems probable that this cost scheduled same where carriers flights of low-cost costs with issue of not covering airline and price, while others pay between this to the but the

is more analysis appropriate than the traditional analysis, arises. competition

The nature of these fixed and variable costs is shown inDoganis

it appears that of pricing perspective


the arguments from the supports practice cost shows that recovery theory. Doganis

2 et al. (1988) discuss economic Baumol when there is contestability. sustainability 3 Issues may be raised as to whether the resulting profitability is a useful guide to the allocation of resources and, if price is to be set equal to marginal of measurement cost, whether may arise. problems



of Transport


and Policy


39, Part 3

variable crew used


costs, such as handling and

operating and airport fees paid


such as fuel costs, en-route and charges,

being out engineering of averaged and main cancelled, along with some portion tenance costs that in aggregate amount to something like 30-45 per cent of costs. Fixed direct operating costs cover depreciation total operating and for leased aircraft and aircraft insurance charges sum to about 25-30 and cabin crew costs. These and are not avoided in the short term as well as per cent of if a flight is cancelled. of these are route specific, per air

the route

flight and cabin service costs, passenger on the type of aircraft to concessionaires, depend be avoided flown. These would if a flight was


any rental fixed flight total costs

costs. Some there are indirect operating Finally, are escapable which (about 5-10 per cent), and others are not (25-35 costs include costs at the destination station and ground cent). These

costs at that same destination. The allocation port and sales and advertising in a variety of ways as shown by of some of these costs can be approached Doganis. The flight from doubt. apparent practice covers its variable revenue. Revenues costs. If it does of some costs and then this, airlines route is to assess whether a particular costs operating in to

not, in excess of

specific its continued



be should operation can make a contribution

fixed operating costs (as defined by Doganis)

operating minimum In the short of some

and non-specific


a flight has to contribute a term, therefore, cent of total operating costs to cover all marginal 50 per In the longer term the fixed costs costs, leaving 50 per cent as fixed costs. can be changed route networks and fleet size. It seems from by adjusting a practical of view that yield management is vital and that the fixed point costs of operation are not at all trivial.

3.0 Competition Analysis

3.1 Oligopoly
A have been advanced, of oligopoly all of of economic models variety are well known, those of Cournot, and Stackel which Bertrand, including firm In addition, of price leadership there are models by a dominant berg. or a lower-cost firm. The correspondence firm or a barometric of actual

price setting with these models

establish different The empirically. market conditions issue

of oligopoly
is really modify that not

is of interest but difficult to


to cover costs. The games attempt them and with whom, is a secondary

in but how this, however, in an their yield management and how they play, they play



Empirical Evidence on Oligopolistic Behaviour of Competing Low-Cost Airlines


3.2 Yield Yield seats


is aimed at maximising total flight revenues management by making to go unsold at a lower price to passengers that are expected available not travel, while at the same time ensuring who would otherwise that those to pay more. lower fares are not purchased Techni by passengers willing

Its advent came in the cally, it is a form of dynamic price discrimination. Act US after the 1978 Airline removed fare regulation and Deregulation a commercial to seek ways of recovering full costs in forced airlines envir onment airline on and Rose, information 1987; Borenstein 1994). Direct (Levine, net yield per seat is not for reasons of commercial in confidentiality, Fare data are, however, the public domain. Only average yield is available. on airline web sites. Fares are a direct indicator of yield as they are available

to the demand in response for seats.4 broadly changed in Europe since the mid-1990s Low-cost have carriers that have emerged a yield management that sees them offer seats introduced strategy generally at low fares well approaches, to departure strategy of one in advance if there common of are the departure date. As so the available fare fares for increase, reaching seats still available. the whole cabin the departure date a peak immediately This simple pricing make revenue



management has recently

simpler and the way the fares have a reverse taper over time carriers. The question been copied by more traditional is, is when there a modification of this price behaviour demand falls setting on an incumbent for the seats airline because of the advent of competition? This can be examined observation and analysis. by

3.3 Data Cases date


and methodology there is direct few


Airport (EMA)5 to identical destinations the incumbent fares monitor offered before

between low-cost competition in number. At the UK's Nottingham in 2003 different low-cost carriers, however, in direct were on each competition and easyJet of their web with

carriers offered



East Midlands service that time, to

each other. At It was served


tive weekdays frequencies

of service.

departure The leisure

bmibaby.6 sites at midday to destinations that were destinations


for 30 consecu by similar in and Malaga

of Alicante

4 A discussion

of standard

(2002), along with 5This is a medium charter catchment England. 6They have services area

an account

revenue management practices of traditional of how it has evolved over time.


is contained

in Belobaba

sized airport located centrally in England that has been a traditional facility for to Europe and as a base for scheduled services offered by BMA. It has a significant to the east of and is on the main North-South with good road access Motorway joined by Ryanair.

since been



of Transport


and Policy


39, Part 3

times in the case with very close departure Spain had a Saturday departure, of Alicante, whereas the business of Glasgow and Edinburgh, destinations a maximum where flight times were within of 20 minutes of each Scotland, other in the working trip to Scotland day. In addition, services from London Gatwick start-up competing Airport to Prague for 51 weekdays in (LGW) by the same airlines were examined 2004. early allowed To examine the relationship time series models First between so the use these price series involves that the cross correlation function Functions at appro in the morning, a return


can be determined.

of all Autocorrelation

priate lags (ACFs) and Partial Autocorrelation



(PACFs) are

on stationary as a guide to model data and plotted identification. The models that are produced have the characteristics that their residuals are white noise. It is the correlation these residuals or pre-whitened between

series that establishes the CCF and tells us at which lead or lag that it is
strongest. The detailed methodology and full results are found in Pitfield

3.4 Results The Malaga fare for EM A to achieve first differenced services data are shown and the ACF in Figure 1. The data are an and PACF



AR(1) model
produce white

for bmibaby and a MA(1) model for easy Jet. Such models
noise residuals.

Figure 1 Fares from EM A to Malaga


bmibaby easyJet 1.00 5.00 9.00 13.00 17.00 21.00 25.00 29.00

fare fare



Empirical Evidence on Oligopolistic Behaviour of Competing Low-Cost Airlines


Figure 2
Cross-correlation 1.0 function plot, Malaga


0.0 U"







T-1-1-1? -7 -5 Lag Number

-i-r~ -3

?i?i?i?i?r 3 5

Coefficient I 7

If the raw data

are Yt, then

the differenced


are zt ?




the AR(1) model

z/ = 0i*/-i B = at. (2) + at, (1)

and using

the backshift



A MA(1) model

zt = at-Qlat_l (3)


zt = (\-^)at.
For The bmibaby, goodness

(-1.48). Bayesian

= -0.413 and for easyJet Q{ = -0.291 0! (-2.42)7 Schwartz of fit indicators of standard error,

Criterion (SBC) and Akaike Information Criterion (AIC) are all acceptable
in Pitfield series are subject to a cross and are reported (2005). The filtered 2. correlation function and the results of this are shown in Figure analysis This correlation shows that of 0.452. fares lead bmibaby easyJet A price leadership model fares could a by one day with be being followed,

t statistics

are in parentheses.



of Transport


and Policy


39, Part 3

Figure 3
Fares 100 from EM A to Alicante

bmibaby easyJet 1.00 5.00 9.00 13.00 17.00 21.00 25.00 29.00

fare fare

DAY perhaps undercut The constant prior that of a low-cost firm, if the initial fares of easyJet are shown reflect costs and

bmibaby. fare data for Alicante for a large part

to departure,

in Figure 3. easy Jet's fare remains the survey period and then rises in two steps fare has not, however, got the expected bmibaby's of Figure 4
Cross-correlation function plot, Alicante

E -Q O ?







on Oligopolistic


of Competing




Figure 5 Fares from EM A to Glasgow


bmibaby easyJet 11 13 15 17 19 21 23 25 27 29 31

fare fare


as it falls pattern before departure. models series each The these fitted in Figure other to

twice before





to previous


Applying ARIMA


to these two fare series results in AR(1)

data correlation series. For bmibaby, 0! = 0.994

= 0.996 (109.177) and for easyJet, 9, (120.495). The CCF of the filtered
4 shows no a 0.808 at zero routes lag with prices following in Figures 5 lead or lag. and Edinburgh fares on the Glasgow but with

are shown indicate

and 6. Pitfield (2005) describes the difficulties encountered

series. Nevertheless, the best models obtained correlation between the filtered fare series, which it could cost from be argued that The these routes should competition. the figures lack of correlation

in modelling
a weak cross that low seen

is surprising given of be a good illustration be because it can be may

that easyJet manages the yield and sets its fares identically for both Scottish destinations for this survey period. are shown to Prague in of LGW fare changes the patterns Finally, are of the AR(1) 7. Models that produce noise white residuals Figure to the non-differenced in form applied CCF shown data. The resulting this sample of routes. analysis out of EMA, that to Malaga, evidence of a price leadership a one day lag in fare setting by the less experienced is provided, with model correlation is evidenced low-cost airline. A strong contemporaneous by the CCF For one route Figure The 8 reveals a strong with no lag. relationship varied results from produces



of Transport


and Policy


39, Part 3

Figure 6 Fares from EM A to Edinburgh ?s

bmibaby easyJet

fare fare

price data relatively Scotland is some airlines' repeated

for Alicante weak form with evidence of from EMA. simultaneous

from EM A as well of correlation there in fare It seems

as from LGW is found on the

to Prague, whereas 'business routes' to that there

is little doubt

for these cases


management a more extensive

from the competing setting arising of yield. It is likely that this would be cases would be and that more analysis Figure 7




to Prague

bmibaby easyJet 13 17 21 25 29 33 37 41 45 49

fare fare



Empirical Evidence on Oligopolistic Behaviour of Competing Low-Cost Airlines


Figure 8
Cross-correlation 1.0 function plot, Prague



-.5 Confidence Limits


-1.0 -7 -5 -3 Lag Number

Coefficient 1

that might hint at the appropriate In addition, price setting model. in all these situations, it is less likely that full cost recovery in the presence of competition. possible found course,

of is

Low-cost but and the carriers are increasing little analysis

their shares of situations of virtually all airline markets where head-on, they compete on their long-term financial the strategic that games of these the future experience of

there has been

viability. they do and will remains

of such competition implications on the broad nature Much depends play The but our limited

given results here are very restrictive, empirical focusing as they do on a few UK-based and they are not conclusive. It markets, seems there is some evidence, of correlation of price setting in though, uncertain. Where this correlation of either the nature

is simultaneous, it is not easy to speculate or how the competitive it is being played. game a traditional Where there is a lag in price setting between the two airlines, can be posited. text book explanation seem Price leadership models would on apposite. It is clear, however, further more detailed that, although empirical work to pinpoint will be needed the nuances of the situation, the price setting




of Transport


and Policy


39, Part 3

behaviour supply


in the duopolistic that are gradually markets in the evolving a threat to the recovery low-cost airline services is of all fixed

Alamdari, Model Belobaba, F. on P. S. Fagan of the Adherence (2005): 'Impact the Profitability of Low-Cost Airlines,' Transport P. (2002): 'Airline Network Revenue Management: and to the Original 25, Recent Low-Cost 377-92. Developments


and State of the Practice,' in D. ed) Aviation Week, New York.

Baum?l, W. J., J. C. Panzar, of Industry Structure, Harcourt

Jenkins (ed.), Handbook

Contestable San Diego.

of Airline Economics
Markets and


and R. D. Willig (1988): Brace Jovanovich,

the Theory

Industry,' R. H. Coase,

S. and N. L. Rose
Journal (1946):

(1994): 'Competition and Price Dispersion

Econ?mica, 13,

in the US Airline

102, 653-83. Economy, of Political 'The Marginal Cost Controversy,'

London. Hanlon,

R. (2002): Flying Off Course: The Economics of International Airlines, Routledge,

P. (1994): 'Discriminatory Fares: Identifying Predatory Behaviour,' Journal of Air

Transport and Public

Management, Yale

1, 89-102.

Levine, M. E. (1987): 'Airline Competition

Policy,' Journal

inDeregulated Markets: Theory, Firm Strategy,

4, 393^94.

on Regulation,

Pitfield. D. E. (2005): 'ATime Series Analysis

titive Schnell, "Low-Cost" M. C. A. Airlines,' International

of the Pricing Behaviour of Directly Compe

Journal the of Transport Effectiveness Economics, of Barriers 32, 15-39. to Entry 10, 413-26. Journal in

Liberalised Vowles, T. M.

'What Determines (2004): Markets?' Journal Airline of Air 'The "Southwest Effect" (2001): 1, 251-58.

Management, Transport in Multi-Airport Regions,'

of Air