You are on page 1of 6

ARTICLE IN PRESS

Journal of Air Transport Management 12 (2006) 4752 www.elsevier.com/locate/jairtraman

Leisure travel in Europe and airline business models: A study of regional airports in Great Britain
Andreas Papatheodoroua,, Zheng Leib
a

Department of Business Administration, University of the Aegean, GR-821 00 Chios, Greece b School of Management, University of Surrey, Guildford, GU2 7XH, UK

Abstract Tourism and air transport are explicitly linked especially in the context of leisure trafc. This paper highlights this relationship by focusing on the impact of the three main airline business models (traditional scheduled, charter and low-cost) on regional airports using Britain as a case study. The panel data econometric results show that despite the current perception, low cost carriers are not the only ones to contribute signicantly to airport aeronautical and non-aeronautical revenue. This observation has important policy implications and calls for transparency in airport subsidies as argued in the conclusion. r 2005 Published by Elsevier Ltd.
Keywords: Tourism; Regional airports; Airline business models; Panel data analysis

1. Introduction Tourism is a global phenomenon of major economic importance. It comprises the activities of persons travelling to and staying in places outside their usual environment for not more than one consecutive year for leisure, business and other purposes (World Tourism Organization, 1994, p.5). The latter include activities such as visiting friends and relatives (VFR), health treatment and pilgrimage. The World Tourism Organization estimates that international tourism arrivals reached 760 million in 2004, an increase of about 10% compared to 2003 (World Tourism Organization, 2005). This is a very encouraging sign given the stagnation that prevailed over the previous years due to the 9/11 terrorist attacks, the war in Iraq and SARS. International tourism receipts were about h460 billion in 2003 (World Tourism Organization, 2004) and are expected to exhibit signicant growth in the following years especially as China is gradually becoming a dominant player in terms of both inbound and outbound tourism. The Asian tsunami disaster is likely to have a negative impact in the short run, but the longer-term still looks very
Corresponding author.

E-mail address: academia@trioptron.org (A. Papatheodorou). 0969-6997/$ - see front matter r 2005 Published by Elsevier Ltd. doi:10.1016/j.jairtraman.2005.09.005

promising especially when domestic tourism is also taken into consideration. Air travel is intrinsically related to tourism mainly in terms of international ows, but also for domestic movements in larger countries. The tourism ratio (which shows the percentage of demand attributed to tourism over total industrial supply) of air transportation should exceed 90% (Smith, 1998). Even if we consider only leisure and VFR trafc, many destinations in the Mediterranean and the Caribbean region are still highly dependent on air transport. Leisure travellers have been traditionally regarded as time-rich, highly price sensitive people with preferences exhibiting a strong seasonal pattern peaking during holidays and festive periods of the year. Trips may last for one or two weeks usually with the family and involve both short- and long-haul destinations with inight service considered more important on the latter. The gradual emergence of the short-break trend, however, has created new peaks during weekends and is essentially associated with people who are better off but time constrained. Singles and people belonging to the double income no kids (DINK) family category are the typical representatives in this case. VFR tourists have similar attitudes to leisure travellers but are more likely to make last minute travel arrangements with limited exibility,

ARTICLE IN PRESS
48 A. Papatheodorou, Z. Lei / Journal of Air Transport Management 12 (2006) 4752

especially in urgency (e.g. attendance of a funeral). Expatriate communities established as a result of previous (South to North in the 1950s and 1960s) and current (East to West) immigration waves in Europe have enhanced VFR trafc; foreign students provide a further boost to such travelling. Charter carriers were the primary air service provider of leisure travel in Europe until the early 1990s. The seasonal and/or occasional character of such ights matched well the pattern of leisure travellers. These airlines acted as the original low-cost ones by emphasising cost reduction: dense seat conguration and high passenger load factors allowed economies of density and low unit costs; irregular and sometimes inconvenient schedules plus the choice of somewhat secondary airports reduced airport fees; basic in-ight servicing cut operation expenses; and the explicit linkage with the travel distribution system in the context of an integrated tourist package led to very low publicity and distribution outlays. As a result of these low costs, charter carriers were able to offer low prices and satisfy the conscious leisure travellers. On the other hand, VFR trafc was a more active user of traditional scheduled carriers, as these y throughout the year directly to major cities, where immigrants and students are likely to be based. Still, the relatively high cost of such air services discouraged frequent VFR travelling; in this context, charter carriers could offer a seasonal alternative. The completion of the European air transport liberalisation in 1997 and the creation of a single aviation market had major implications for leisure and VFR travel (Papatheodorou, 2002). The abolition of any legal distinction between scheduled and charter carriers and the gradual emergence of low-cost carriers (LCC) as a third, robust player in the market were among the most important developments. Within a short period of time, LCC became leaders in cost reduction: they provided a basic quality service in a single class cabin of dense conguration; they ew to secondary airports and outsourced many activities to lower staff; and they instigated disintermediation by favouring the direct contact between the airline and the customer over the Internet. LCC also innovated by offering one-way tickets priced according to efcient yield management techniques, based on simplicity and very restricted conditions. By the late 1990s, the differences in the airline business models have become clear: traditional scheduled carriers focused on service delivery offering a network-based product, which could serve well the business passenger and the afuent leisure and VFR clientele; LCC provided point-to-point ights of basic quality but priced so low that attracted many leisure and VFR travellers; whereas, charter carriers were sitting somewhere in-between with a rather uncertain future. This product differentiation framework, however, did not last for long as the dynamics of competition gave rise to osmotic phenomena: currently, it is becoming increasingly difcult to distinguish between the three types of carriers. For example, many European

traditional scheduled airlines such as British Airways, Lufthansa, SAS, Iberia and Aer Lingus engaged in a drastic cost-cutting exercise aiming to replicate many of the features of the LCC model, at least in short-haul ights. On the other hand, some LCC started upgrading their services; easyJet, for example, serves also primary airports (e.g. Paris Orly) and offers exible tickets, while Jet Blue in the USA provides leather seats and a live satellite television programme. Charter carriers decided to radicalise their product according to the LCC prototype albeit in occasionally longer haul destinations. In some cases, charter carriers setup their own LCC afliates as shown by the examples of My Travel Airways and My Travel Lite in Britain and Hapag Lloyd and Hapag Lloyd Express in Germany. As a result of this osmotic process and the acquisition of a new air travel culture based on knowledge and bargaining, a clear-cut customer segmentation is also difcult: illustratively, the results of the latest Barclaycard Business Travel Survey reveal that 70% of business travellers used LCC with satisfaction rates over 95% (Davies, 2005). Conversely, a recent report by the British Air Transport Users Council stressed the importance of considering the time of booking a ight as well as the hidden monetary and time transfer costs associated with secondary airports; interestingly, traditional scheduled carriers can occasionally offer a better-value-for-money than their low-cost counterparts (AUC, 2003). Still, the various airline business models retain some differences regarding network structure and airport choice. In particular, the traditional scheduled airlines aim at adding value to the passenger by offering an extensive network based on own hub-and-spoke services and interlining agreements; its appeal is enhanced by airline participation in one of the three major strategic alliances (oneworld, Star Alliance and Sky Team) and the existence of sophisticated frequent yer programmes. To deliver these network services, traditional carriers use primary airports with all necessary facilities. Regional airports play only a minor role in their business model as the majority of related services act as feeders to major hubs. On the other hand, LCC focus exclusively on basic point-to-point ights. This does not mean that LCC abstain completely from connecting trafc: the latter amounts to 30% of Southwests market, while in London Stansted Airport 14% of passengers engaged in do-it-yourself connections in 2003 (Mosner, 2005). However, LCC are not prepared to offer these extra services themselves to avoid complication and an increase in their cost base. They choose primarily secondary regional airports (Orly is one of the few exceptions) and benet substantially from low airport charges and station costs. For example, Frankfurt Hahn costs Ryanair h4.25 per departing passenger and there is no landing fee; in contrast a B737 operator at Frankfurt Main pays h13 per departing passenger and a landing fee of about h1.75 (Button et al., 2002). In other words, the sustainable competitive advantage of LCC is mainly derived from the adoption of point-to-point services

ARTICLE IN PRESS
A. Papatheodorou, Z. Lei / Journal of Air Transport Management 12 (2006) 4752 49

operating from regional airports (Lei et al., 2004). Charter carriers have always offered point-to-point services and they increasingly y also from secondary airports. These different airport choices have important implications for the airlineairport relationship. Traditionally, this has been characterised as love-and-hate; despite their common future and the complementarity of their operations, airlines have often accused airports of abusing their market power, while the latter justied any charge increase on the need to improve and expand a risky and sunk piece of infrastructure. The Ryanair experience has revealed that secondary airports are prepared not only to charge less but even to offer subsidies to attract trafc in the context of wider regional economic development (Papatheodorou, 2003; Starkie, 2002). Demand for air transport services is essentially derived from tourism activities. When all potential destinations are considered, tourism choice is discrete, i.e. go to place X and not to Y; lower airport charges and direct airline services signify a substantial improvement in accessibilityif the destination is nally chosen, prots from consumption of local hospitality and other tourism services may more than compensate any losses at the airport level. Similar results hold when an airport acts as an origin rather than a destination gateway. Enhanced tourism consumption in this case is substituted by new employment opportunities at the airport and the wider local economy and the subsequent generation of additional expenditure and income through a multiplieraccelerator process. As expected, this subsidisation strategy raised adverse reaction from the part of various airlines such as Air France, which accused Ryanair of beneting from unfair competition. The issue has reached the European Commission, which in its recent verdict on Charleroi Airport, espoused partly the rationale behind cross-subsidisation stressing, however, the need for transparency and fair trade; an airport should not offer secret concessions to specic airlines but adopt a tender-like procedure open to all carriers. The private investor principle should also be met; in other words, the longer term nancial sustainability and protability of an airport as a stand alone investment should be guaranteed (European Commission, 2004). In this context, it is essential for an airport to know what type of airlines and passengers are likely to be more benecial. This is admittedly a very difcult exercise, as the benets or losses of the airport as such should be compared with the impact on the business turnover of the wider local tourism activities. Nonetheless, exploring differences in consumption patterns according to a market segmentation based solely on the choice of airline may be somewhat bizarre; what matters primarily is demographic and vacation habit criteria (Holloway, 2004). Moreover, the private investor principle necessitates focusing on the airport side by providing a breakdown in terms of aeronautical and nonaeronautical revenue. The empirical part of this paper undertakes such an econometric exercise aiming at proving useful to airport planners, regional economists and tourism destination policymakers.

2. Empirical study The study focuses on the British experience essentially due to data availability and the need to focus on airports served by all three-airline business models for comparison reasons. Twenty-one regional airports are considered over an eight-year period ranging from 19951996 to 20032004. The airports exhibit considerable differences in size; London Gatwick, which is the largest in the sample, recorded 32.1 million passengers in 20002001 while Blackpool, which is the smallest, handled only 46,820 passengers in 20022003. For this reason, the data set is further subdivided into two groups, large and small airports, using three million passengers per year as the cutting benchmark. Ten large-size airports are included in the sample namely Belfast, Birmingham, Bristol, Edinburgh, Glasgow, London Gatwick, London Luton, London Stansted, Manchester and Newcastle (in alphabetic order). The remaining eleven are classied as small-size airports, namely, Aberdeen, Blackpool, Bournemouth, Cardiff, East Midlands, Exeter, Humberside, Leeds and Bradford, Liverpool, Southampton and Teesside. Airport revenue data are collected from the Annual Airport Statistics published by the Centre for the Studies of Regulated Industries and deated using the British Consumer Price Index (CPI). Airline ight and passenger data are obtained from the UK Civil Aviation Authority. All three airline business models are considered; data are adjusted to the British nancial year, ranging from 1st April to 31st March. Table 1 shows relevant descriptive statistics. Aeronautical charges (AR) in Britain are levied according to aircraft size and passenger numbers. As it has been difcult to obtain aircraft size data, air transport movements, i.e. number of ights is used as a proxy for aircraft size. Non-aeronautical revenue (NAR) is related to the number of passengers. Although meeters, greeters and airport personnel can also affect the magnitude of NAR, the primary drive is passengers (Gillen and Hinsch, 2001). Two empirical models are estimated in log-linear form to nd out the rate of growth with respect to ight and passenger numbers. Passengers are highly correlated with ights and thus dropped from the AR model. On the other hand, the number of passengers carried by LCC, charter and full-service carriers are used as explanatory variables in the NAR model. log ARit ai b1 Lccflightit b2 Charterflightit b3 Fullserviceflightit eit ; Model 1

log NARit di g1 Lccpaxit g2 Charterpaxit g3 Fullservicepaxit uit ; Model 2 where the subscripts i and t indicate the airport and time period (year) respectively; a and d are coefcients that vary across airports but are constant over time; the b and g are coefcients which are constant across airports and time; e,

ARTICLE IN PRESS
50 Table 1 Descriptive statistics Variables All airports (N 21) Mean Total pax Non aeronautical revenue ( 103 ) LCC pax Charter pax Full-service pax Aeronautical revenue ( 103 ) LCC ights Charter ights Full-service ights 4,821,583 25,355 858,651 1,697,978 2,264,954 26,478 8,096 15,073 36,458 Std. dev. 6,879,189 43,827 2,152,222 2,776,110 3,841,623 32,227 18,892 14,674 42,270 Large airports (N 10) Mean 8,943,011 47,707 1,566,028 3,110,572 4,266,412 46,766 14,572 20,421 62,027 Std. dev. 8,137,319 55,510 2,919,122 3,497,314 4,811,512 36,878 25,414 15,535 48,617 Small airports (N 11) Mean 1,074,830 5,034 215,582 413,802 445,446 8,035 2,210 10,211 13,214 Std. dev. 931,398 3,227 519,606 421,964 513,425 5,976 4,962 11,989 11,616 A. Papatheodorou, Z. Lei / Journal of Air Transport Management 12 (2006) 4752

Table 2 Impact on regional airports aeronautical revenue Dependent variable: log AR Constant a LCC ights b1 Charter ights b2 Full-service ights b3 No. of observations R2 F test of group effects LM test Hausman test All airports 8.8232 (47.41)*** 0.0125 (12.73)*** 0.0144 (2.39)** 0.0109 (7.39)*** 189 0.68 F(20,165) 170.08 655.68 6.84 Large airports 9.5020 (66.67)*** 0.0108 (13.56)*** 0.0135 (2.34)** 0.0090 (7.47)*** 90 0.80 F(9,77) 46.31 185.41 5.01 Small airports 7.8499 (32.11)*** 0.0379 (8.20)*** 0.0280 (3.37)*** 0.0343 (4.12)*** 99 0.50 F(10, 85) 123.02 309.97 1.64

Notes: Figures in parentheses are t values; One, two and three stars indicate signicance at the 10%, 5% and 1% levels, respectively.

u are the error terms, which are identical and independently distributed for all i and t; AR and NAR are aeronautical and non-aeronautical revenue in airport i in year t; Lccight, Charteright and Fullserviceight show the total number of ights of the low-cost, charter and full-service airlines, respectively, in airport i and year t; Lccpax, Charterpax and Fullservicepax are the total number of terminal passengers carried by the above three types of airlines, respectively, in airport i and year t. All b and g are expected to have positive signs. Models 1 and 2 have been estimated using panel data techniques to account for differences across airports and time. In principle, estimation can be done in three ways depending on whether the individual cross-section effects are considered to be constant, xed (parametric shifts of the regression function) or random. Following Greene (2003), Breusch and Pagan Lagrange multiplier (LM) and Hausman tests were applied to choose the appropriate model formulation. The LM test is used rst to compare the random effects specication against the simple linear model. Acceptance of the null hypothesis means that the classical regression model with a single constant term is appropriate, i.e. the model can be estimated by pooled ordinary least squares (POLS). Conversely, rejection of the null hypothesis is in favour of the random effects specication. The Hausman test has been subsequently

used to determine the choice between xed and random effects model. Under the null hypothesis of the Hausman test, both models are consistent but the random effects specication is more efcient; acceptance of the alternative hypothesis suggests that the xed effects model is more appropriate. Moreover, an F test was performed to examine whether there are airport-specic effects across the different groups of airports (large and small). The F statistic for testing the joint signicance of the airportspecic effects was F(20, 165) 170.08 for Model 1; this suggests strong evidence of airport-specic effects in the data. Similar results were derived from Model 2. On these grounds, both xed and random effect models were estimated. Table 2 reports the results of Model 1; all variables are measured in thousands. The signicant LM test statistic suggests that a common slope for the different panel groups in the pooled regression cannot be assumed: a simple OLS regression of a straightforward pooling of all observations without considering heterogenity would report biased results. The insignicance of the Hausman test statistic at the 5% level suggests that the random effects model is more efcient both for the whole sample and two airport subgroups. Therefore, only the results of the random effects models are reported here (the others are available from the authors upon request). A Chow test is

ARTICLE IN PRESS
A. Papatheodorou, Z. Lei / Journal of Air Transport Management 12 (2006) 4752 Table 3 Impact on regional airports non-aeronautical revenue Dependent variable: log NAR Constant d LCC passengers g1 Charter passengers g2 Full-service passengers g3 No. of observations R2 F test of group effects LM test Hausman test All airports 9.0060 (100.05)*** 8.44E05 (6.77)*** 8.46E05 (1.91)* 1.53E05 (0.47) 189 0.76 F(20,165) 50.53 412.84 13.85 Large airports 10.0092 (89.50)*** 6.17E05 (7.27)*** 2.16E05 (0.71) 2.14E05 (0.98) 90 0.69 F(9,77) 33.93 96.10 26.62 Small airports 7.6269 (56.41)*** 5.54E04 (11.34)*** 7.65E04 (5.17)*** 6.26E04 (4.71)*** 99 0.67 F(10, 85) 28.69 182.17 4.67 51

Notes: Figures in parentheses are t values; One, two and three stars indicate signicance at the 10%, 5% and 1% levels, respectively.

performed to see whether the coefcients of the large airports group are equal to those of the small ones. The signicant F(4,181) 51.4 means that the coefcients of the two groups are different and justies the classication of the airports into two separate groups. All explanatory variables are highly signicant across the three samples and have expected signs. Given the log-linear form, the coefcients may be interpreted as the percentage impact of an incremental thousand ights on aeronautical revenue. For example, an extra thousand LCC ights result in a 1.25% AR growth in the all airports group. For the whole sample, charter ights have the largest impact on airport aeronautical revenue, followed by low-cost and full-service. This may be explained by the use of different types of aircraft. Charter carriers often use larger aircraft while full-service carriers operate turbo-props or regional jets; as for low-cost carriers, B737 or A320 is the standard aircraft type. Similar results hold for large airports. On the other hand, LCC ights emerge as the largest contributor in the small airports group with an impact of 3.79% per additional thousand ights. This nding is very interesting and suggests that the small regional airports in Britain should become more oriented towards low-cost trafc to increase their aeronautical revenue. Based on the same econometric procedure, Table 3 reports the results of Model 2 concerning the impact on the airports non-aeronautical revenue; all variables are again measured in thousands and similar interpretation holds for the coefcients. The LM and Hausman test statistics indicate that the xed effect specication is appropriate for the whole sample and the large airports group, while the random effect model is more suitable for the small airports group. Table 3 shows only the benchmark parameter d; the constant terms for individual airports and all the alternative model specications are available upon request. All coefcients have the expected signs. The results show that the LCC passengers contribution is highly signicant at 1% level across all three samples, albeit smaller for the large airports group than for the small one, where an incremental thousand passengers generate an increase of 0.0554% in non-aeronautical revenue. The contribution of

charter and full-service passengers is insignicant for the whole sample and the large airport group but highly signicant in the small one. In fact, their contribution is higher compared to LCC travellers! The results are not surprising as charter passengers usually travel for leisure purposes and have more desire to purchase goods and/or services at airports. Moreover, they usually stay longer at airports, thus, providing more commercial revenue opportunities for airports. Similarly, full-service passengers may belong to higher-expenditure groups than their low-cost counterparts; alternatively, they may travel on business and wish to make some last-minute purchases (for personal use or gifts) from the airport.

3. Conclusions The emergence of low-cost carriers as a viable business model has revolutionised the airline industry since the early 1990s. Leisure and VFR passengers in Europe have now a wider choice and are able to discover the European regions at lower fares than in the past. No matter whether a regional airport operates as an origin or destination gateway, notable improvements in accessibility can play a signicant role in economic and/or tourism development. The results of the empirical study in this paper, however, belie the perception that LCC are the only way forward for regional airports. Full-service carriers and charter airlines can also have a signicant, if not higher, contribution to both aeronautical and non-aeronautical airport revenue. A replication of the present econometric exercise using different data sets is required of course to test the robustness of these conclusions. A potential validation would provide support for the European Commission argument by calling for the removal of obscure airport subsidies and the adoption of a new transparent framework open to all carriers. Despite the existence of victims, LCC are undoubtedly the major fashion in the contemporary airline industry; by no means, however, does this mean that the evolved full-service and charter carrier models are necessarily dead.

ARTICLE IN PRESS
52 A. Papatheodorou, Z. Lei / Journal of Air Transport Management 12 (2006) 4752 Greene, W., 2003. Econometric Analysis, fth ed. Prentice Hall, Harlow. Holloway, J.C., 2004. Marketing for Tourism, fourth ed. Prentice Hall, Harlow. Lei, Z., Papatheodorou, A., Szivas, E., 2004. The impact of low-cost carriers on regional airports nancial performance: evidence from the UK. Presentation at a German Aviation Research Society Workshop, Bremen. Mosner, A., 2005. The Central/Eastern Europe aviation market: a Darwinian view. Presentation given at CEE Aviation Conference, Prague. Papatheodorou, A., 2002. Civil aviation regimes and leisure tourism in Europe. Journal of Air Transport Management 8, 381388. Papatheodorou, A., 2003. Do we need airport regulation? Utilities Journal 6 (10), 3537. Smith, S., 1998. Tourism as an industrydebate and concepts. In: Ioannides, D., Debbage, K.G. (Eds.), The Economic Geography of the Tourism IndustryA Supply Side Analysis. Routledge, London. Starkie, D., 2002. Airport regulation and competition. Journal of Air Transport Management 8, 6372. World Tourism Organization, 1994. Recommendations on Tourism Statistics. United Nations, New York. World Tourism Organization, 2004. Tourism Highlights 2004. World Tourism Organization, Madrid. World Tourism Organization, 2005. World Tourism Barometer, 3.

Acknowledgements The authors are grateful to the UK Civil Aviation Authority and Nenad Njegovan for providing relevant air trafc data. Zheng Lei would also like to thank Edit Szivas for her valuable guidance and support and Chenggang Wang for his insightful comments on the econometric analysis. References
Air Transport Users Council (AUC), 2003. Booking a Flight? Civil Aviation Authority, London. Button, K., Haynes, K., Stough, R., 2002. Towards an Efcient European Air Transport System: A Study for the Association of European Airlines. Association of European Airlines, Brussels. Davies, P., 2005. 70% of Biz Travellers Fly Low Cost. Travel Mole, 16th May, www.travelmole.com European Commission, 2004. The Commissions Decision on Charleroi Airport Promotes the Activities of Low-Cost Airlines and Regional Development. IP/04/157. European Commission, Brussels. Gillen, D., Hinsch, H., 2001. Measuring the economic impact of liberalization of international aviation on Hamburg airport. Journal of Air Transport Management 7, 2534.