A Model to Identify Airport Hubs and Their Importance to Tourism In Brazil

Air transportation in Brazil has been recently liberalized and one of the consequences of this process is the concentration of flights in a few hubs. During the years 2006–2007 two fatal accidents created unprecedented chaos in both land and air
of 9
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
  Author's personal copy A model to identify airport hubs and their importance to tourism in Brazil Tiago F.G. Costa a , Gui Lohmann b , * , Alessandro V.M. Oliveira a a Latin American Center for Transportation Economics, Aeronautics Institute of Technology, Praça Marechal Eduardo Gomes, 50, Vila das Aca´cias. CEP 12228-900, Sa˜o Jose´ dosCampos, SP, Brazil b School of Travel Industry Management, University of Hawai’i at Maˆ noa, 2560 Campus Road, George Hall 346, Honolulu, HI 96822, USAKeywords: Hub-and-spoke networksTourismBrazilAir transport crisisHerfindahl–Hirschman Index (HHI)Federal Aviation Administration a b s t r a c t Air transportation in Brazil has been recently liberalized and one of the consequences of this process isthe concentration of flights in a few hubs. During the years 2006–2007 two fatal accidents createdunprecedented chaos in both land and air sides of the system with harmful consequences to tourism inBrazil. The consequences were more airport congestion and many episodes of flight delays and cancel-lations that lasted for several months. We argue that, among other factors, this state of blackout wasa result of the increase in the degree of concentration in few airports, particularly Congonhas (in Sa˜ oPaulo) and Brası´lia. Using data obtained from a survey with Brazilian experts, a comparison was madewith two existing methods (the one used by the US Federal Aviation Administration and the usualHerfindahl–Hirschman method) in order to calculate the number of hubs in Brazil. Due to the hugediscrepancy obtained between data from the survey and the other two methods considered, a newmathematical method based on the Herfindahl–Hirschman Index was proposed to identify the numberof hubs in a given network. Drawing from the examples of what happened to tourist destinations duringand after the air transport crisis in Brazil, the article concludes discussing the need for a more accuratetool to identifyand to monitor the concentration of flights at the Brazilian air transportation network andits importance to tourism.   2009 Elsevier Ltd. All rights reserved. 1. Introduction Hub-and-spoke (HS) networks are found in several areas of modern society, including transportation, telecommunications andlogistics (Horner & O’Kelly, 2001). In air transportation, HSnetworks appeared for the first time in 1955, when Delta Air Linesused Atlanta as its hub, in an effort to compete with Eastern AirLinesinthesouth-easternpartoftheUnitedStates(Babcock,2002).After the US deregulation in 1978 and the European deregulationprocess that took place between 1987 and 1997, HS networks wereadopted by most of the full-service airlines that operated inderegulated markets (Alderighi, Cento, Nijkamp, & Rietveld, 2005;Martı´n & Voltes-Dorta, 2009).The major advantage of the HS network is that it allows airlinesto reduce the costs of travel and increase their connectivity (Pels,2001).Theairline’scostof travelisreduced,asgroupingpassengerswith the same travel srcin but different destinations are gatheredin feeder flights, which are then distributed in connecting flightsfrom the hub to their final destinations. In addition, the connec-tivity is increased within the hub by concentrating landings andtakeoffs at the hub, commonly called hub waves (Alderighi et al.,2005). Although it may enhance carriers’ production efficiency, onthe other hand a hub airport typically provides airlines with somemonopoly power to control scarce airport facilities (Nero,1999). Inaddition, they swell the operations in these airports used as hubspotentiallyincreasingairsideandlandsidedelays,entailabarriertonew airline entry, increase detour level in the network and theairspace congestion, overworking controllers and threateningsafety (Button, 2002; Hoffman & Voss, 2000; Rodrigue, Comtois, &Slack, 2006; Oliveira & Salgado, 2006).In spite of the huge importance of hubs, there is little consensusamong scholars regarding a precise definition for a hub. As anexample, Burghouwt (2007) provides a list of fifteen definitionsfrom different scholars. Nevertheless, it is possible to identify somecongruence among these definitions. One of them is the word‘concentration’. The concentration of traffic in both space and timemeans that airlines consolidate their operations so that traffic froma diverse range of srcins can be distributed to a diverse range of final destinations. In this sense, O’Kelly (1998, p. 171) summarizesthat ‘‘hubs [ . ] are special nodes that are part of a network, locatedin such a way as to facilitate connectivity between interacting *  Corresponding author. Tel.:  þ 1 808 956 0485; fax:  þ 1 808 956 5378. E-mail addresses: (T.F.G. Costa), (G. Lohmann), (A.V.M. Oliveira). Contents lists available at ScienceDirect Research in Transportation Economics journal homepage: 0739-8859/$ – see front matter    2009 Elsevier Ltd. All rights reserved.doi:10.1016/j.retrec.2009.10.002 Research in Transportation Economics 26 (2010) 3–11  Author's personal copy Fig. 1.  Aircraft movements to top airports in Brazil during the period July 2005–June 2006 – only routes with six or more average daily flights are presented. T.F.G. Costa et al. / Research in Transportation Economics 26 (2010) 3–11 4  Author's personal copy places’’. Apart from the concentration-distribution characteristic,centrality is also another key term usually associated with hubs(Shaw, 1993). Lohmann, Albers, Koch, and Pavlovich (2009), for example, describe the advantages that Singapore and Dubai ach-ieved in terms of their location and central position by developinga hub network to improve not only air traffic but also tourists totheir destinations.Another common aspect identified among some of the defini-tions presented by Burghouwt, it is the fact that an HS networkshould have a limited number of hubs. For example, Button, Hay-nes, and Stough (1998, p. 20) affirm that ‘‘in hub-and-spoke oper-ations, [ . ] carriers generally use one or more large airports’’.Kanafani and Ghobrial (1985) state that the concentration of flightsoccurs at fewairports, while Oum, Zhang, and Zhang (1995, p. 837)affirm that an airline’s operation using hub-and-spoke networkswill occur at one, or very few, hub cities.Considering the lack of a single definition for a hub and the factthat HS networks should have few hubs, the question to be posedthen is: how to identify or measure the number of hubs in a givennetwork? A considerable number of studies have tried to oper-ationalize the definition of a hub and, overall, their focus has beenon the airline network, rather than the airport in itself. Calculationsare heavily based on standard economic concentrations measures,such as Theil/Entropy measures, Coefficient of Variation, the Her-findahl indices and the Gini index (Alderighi et al., 2005; Bur-ghouwt, Hakfoort, & Ritsemavan Eck, 2003; Martı´n & Voltes-Dorta,2009; Reynolds-Feighan, 2001).As in many other parts of the world, air transport in Brazil wentthrough a deregulation process, aiming to improve competitionamong airlines. Indeed, according to Oliveira and Salgado (2006)the deregulation process reduced fares and increased operationalefficiency and competitiveness, but airports and air traffic controlwere not deregulated or privatized. They stayed under thegovernment management. These factors combined to builda peculiar network design (see Fig.1 for the concentration of trafficinthecitiesofSa˜ oPaulo–insetmap-andBrası´lia–largemap)thatended up contributing for two of the worst air transport accidentsin the country (with 354 fatalities), and repercussions to the wholedomestic air transport system that lasted between 2006 and 2007.In the case of the first accident (in 2006), initially air trafficcontrollers in Brası´lia were blamed, but not willing to assume fullresponsibility for this fatality and also blaming the lack of invest-ments in infrastructure (TCU, 2006), they decided to adopt a slow-down work procedure of following exactly what prescribed by theregulation, particularly in relation to the number of aircrafts theycanhandleatatime(Endres,2007).Astheairtrafficcontrolsystemin Brazil was underfunded, with fewer and not well-qualifiedcontrollers, several flights had to be systematically canceled orwere delayed and many scary potential travelers changed theirplans fordomestic trips (G1, 2006). This includes notonlychangingthe choice of destinations visited, but also the modes of transportused. Consequently, domestic tourism was heavily impacted. Thesecond accident (in 2007) happened in Congonhas, the busiestairportinthecountry.Thelackofconfidenceinthesystem,notonlyfromthepointofviewofsafety,butalsointermsofpunctualityandthe stress travelers were under wondering about whether theirflights would be available, made the authorities finally to takeaction. The first decision was to restrict connections and long haulflights from Congonhas Airport, with direct flights only to desti-nations located within a distance of 1000 km. This ultimatelytransformed its characteristic from a national hub into a regionalairport (Knibb, 2007). Unfortunately, these measures did not lastlong.Hence, the aims of this article are three-fold. Firstly, we discussthe recent air transport crisis that took place in Brazil. There areseveral factors to explain what the media has labeled as ‘airtransport blackout’. We argue that one of the key reasons for thetwo fatal accidents and the repercussions that followed is theconcentration of flights at two major hubs, i.e. Brası´lia and Con-gonhas. The impact of this crisis on tourism in Brazil is then pre-sented. Secondly, we compare data obtained from a survey withBrazilian experts with two existing methods to calculate thenumber of hubs in Brazil: (1) the method used by the US FederalAviationAdministration–FAA;and(2)theNEPmethod.Thirdly,dueto the huge discrepancy obtained between data from the surveyand the other two methods considered, a new mathematicalmethod based on the Herfindahl–Hirschman Index is proposed toidentify the number of hubs in a given network. We concludestating that having a better method to identify and monitor thenumber of hubs can serve as a tool for policy makers, air transportorganizations and other stakeholders involved with it, includingtourism enterprises and agencies, to evaluate the degree of concentration of a given network. We assume that hubbing isa choice by airlines and is clearly driven by economic incentives.Our claim here is that by permitting a better identification of hubsourmethodallowsauthoritiesto betterplan andinducethepathof infrastructure investments. In the end, concentration driven byeconomies of hubbing will occur without causing unwantedcongestion and externalities that more than compensate thoseeconomies. Undoubtedly, in a deregulated market, this identifica-tionbyauthoritiesdoesnotmodifyhubbingdecisionsandcertainlywillnotdirectlyinfluenceairlinenetworks.However,withacleareridentificationofairlines’ strategiestowardshubbingauthorities areable to design specific policies regarding major airports. Forexample, Ramsey pricing combined with slot allocation may bedesigned in order to induce airlines to behave in such way thatcongestion is avoided and secondary airports are benefited. 2. The 2006–2007 air transport crisis in Brazil and its impactson tourism With the introduction of the Air Transport Liberalization Policyin the 1990s, the rigid structure existing in Brazil diminished theeconomic regulation that used to control the sector. New smallairlines started to have access to air transport markets that becamemore competitive. In 2003, after a certain period of economic Fig. 2.  Domestic enplanements (in million passenger boardings per year) and theshare of top-two largest airports in Brazil. T.F.G. Costa et al. / Research in Transportation Economics 26 (2010) 3–11  5  Author's personal copy freedom, regulation was imposed again due to what was consid-ered as an ‘‘excess in capacity’’ and ‘‘predatory competition’’ in themarket(Bettini&Oliveira,2008).Nevertheless,thereareclearsignsthat the Liberalization Policy brought many benefits, such as lowerairfares, higheroperational efficiencyand competitiveness throughmarketexpansion. However, the same economic liberalismwas notappliedtothesector’slagginginfrastructure,includingairportsandair traffic control.One of the negative aspects of the liberalizationprocess was theconcentration of aircraft movement into key airports, particularlyin two major airports located in the city of Sa˜ o Paulo (Congonhasand Guarulhos) and Brası´lia’s airport. Fig. 1 illustrates all city-to-city routes in the country with pairs of routes with six or moreaverage daily flights (three return trips) and provides a visualdemonstration of the importance of the city of Sa˜ o Paulo, whereCongonhas and Guarulhos airports are located, and Brası´lia, thecapital city, as the key hubs and the bottlenecks of air transportsystem in Brazil. These two cities also became the epicenter of theworst air transport accidents in the country.The first accident happened in September 2006, when a mid-aircollision occurred between a Boeing 737 from Gol Airlines and anEmbraerLegacy600executivejet.Theexecutivejetlinerwasabletoland safely, but all 154 passengers and crew onboard Gol’s Flight1907 were killed. At the time of the accident, both airplanes werebeing handled by the air traffic control in Brası´lia. After that, inOctober 2006, a breakdown in one of the country’s radar systems(Cindacta 2 – Integrated Center for Air Defense and Traffic Control)located in Curitiba caused delays in at least 146 commercial flights.Inthesamemonth,theinvestigationsabouttheaircollisionstartedand the air traffic controllers were feeling pressured. On November2nd, of the same year, a group of air traffic controllers went onstrike in the Brası´lia air traffic control center (Cindacta 1), whichhandles the Center-West and Southeast regions of Brazil,accounting for 75% of Brazilian air traffic. Flight delays were spreadthroughout the country, leaving a crowd of passengers in airportloungesofBrazilianmajorcities.Asaresult,inDecember2006,55%of all flights were delayed and for several months air transportsystem in Brazil was not considered reliable anymore (Folha de S.Paulo, 2006b). The first half of 2007 was flagged by several newstrikes,massivedelaysandawidespreaddiscussionintheBraziliansociety, including the Congress, which installed two investigativecommissions to deal with the airports problems. However, in July2007, another fatality happened when an Airbus A320 from TAMAirlines slipped off the airstrip at Congonhas airport, in Sa˜ o Paulo,and crashed into a building causing 200 deaths.These events highlighted the main problems in the Brazilianairport network and brought the discussion about the high growthof air transportation traffic and the negative impacts it imposed onthelackinginfrastructureanditsconcentrationonfewairports.Thevarious specialists interviewed by the media during the air trans-portation crisis suggested that while the system was clearlyaffected by operational issues including an overall lack of invest-ments in the airside (runways, control towers and air trafficcontrol), lack of transparency and managerial issues related to themilitary control of the air traffic control system and the concen-tration of traffic on very few hubs (Endres, 2007; Knibb, 2007). Airtransport authorities also acknowledged these issues as the firstdecisions made after the second accident was to restrict long haulflights in Congonhas airport and the discussion to expand Gua-rulhos airport or to build a third airport around the city of Sa˜ oPaulo.The repercussions of the first accident were felt in many areas,including tourism, one of the most affected sectors (Brancatelli,2007). As a result, some potential tourists cancelled their trips,while actual travelers shifted their mode of transportation from airto road (cars and busses) or even decided to take a cruise vacation.At the end of 2006, the highway traffic intensified with someestimations of a 25% increase in comparison to the previous year,particularly in the states of Sa˜ o Paulo, Minas Gerais and Rio de Janeiro, which are the largest domestic-tourist generating stateswithin the country (G1, 2006).The impact of the crisis was particularly severe to the coastalresort destinations in the Northeast region of Brazil, such as For-taleza,Natal,RecifeandSalvador(seeFig.1).Theydistancebetweentwoandthreehoursbyplanefromthekeygeneratingmarketswithvisitors from these states opting to travel to nearby destinations.Package sales, for example, were down by 8% in December 2006,affecting particularly hotel cancelations in destinations located inthe Northeastern and Northern parts of Brazil. Overall tourismdemand in the Northeast region of the country were down 35%,with over 15,000 hospitality employees been dismissed. A hugeincrease in the demand for cruise tourism was also identified,presumably influenced by the fact that most of the ships departfromtheportsofSantos(lessthan60kmfromthecityofSa˜ oPaulo)andRio de Janeiro,two of the largest touristgenerating cities in thecountry. Approximately 70,000 extra tourists took a cruise in Brazilin the 2006/2007 season, in comparison to the 2005/2006 summermonths (Folha de S. Paulo, 2006a). 3. Assessing airport concentration in Brazil The deregulation of the US domestic passenger aviation in 1978was followed by a notable concentration of traffic around a smallnumber of central airports or ‘hubs’. This phenomenon has beenobserved in recently liberalized markets all over the world, withmajor airlines aiming at consolidating their networks to enhanceefficiency and market power. In Brazil, the same trend wasobserved in recent years. Fig. 2 presents the evolution of domesticenplanements from 1998 to 2006. It is possible to note the fastgrowthof the industry, with a production of 16.9 (43.–26.2) millionadditional enplanements within the period. This represents a 64%increase in nine years. Additionally, the industry has also experi-enced a notable increase in airport concentration. In fact, theconcentration ratio of the two major airports (hereafter to be Fig. 3.  Airports that are hubs and that should be hubs according to the responsesobtained from a survey with air transport experts ( n  ¼  79). T.F.G. Costa et al. / Research in Transportation Economics 26 (2010) 3–11 6  Author's personal copy mentionedas C-2 ratio), Sa˜ o Paulo/Congonhasand Brası´lia airports,has apparently increased from 22% in 1998 to 29% in 2006, witha peak of 32% in 2004.Our aim here is to verify whether the increase in the C-2 ratiowas really an indication of a higher concentration of networkswithin the domestic air transportation system in Brazil. Toaccomplish that, we make use of a proposed methodology forassessing network concentration in air transportation. Indeed,many authors such as Alderighi et al. (2005), Huber (2008), Martı´nand Voltes-Dorta (2009) and Reynolds-Feighan (2001) aimed at assessing major impacts caused by liberalization and competitionon the air transportation networks by directly measuring concen-tration. The concentration measures found in the literature areusually related to the Gini or Herfindahl–Hirschman (HHI)concentration indexes. Burghouwt et al. (2003) propose a correc-tion for the standard Gini index (the ‘‘NC-Gini’’) that makes itpossible to compare the spatial structure of airline networksindependent from network size. Here we make use of the HHI-related procedure with a simple and direct adjustment for thenetwork structure of carriers in order to assess the concentrationpath in the Brazilian airline industry since liberalization. We aimnotonly to measure the concentration but also to identify what arethehubsinthecountry.Webelievethatothermeasuressuchastheproposed by Martı´n and Voltes-Dorta (2009) may be used to assessconcentration in order to permit comparisons with our results.However, we let those analyses for future work. 1  3.1. A survey on the airline hubbing practices in Brazil Major carriers in Brazil do not explicitly adopt a hub-and-spokestructure as in the United States. In fact, no carrier in the countryhas a fortress hub like American Airlines’ Dallas/Fort Worth,Northwest Airlines’ Detroit Metropolitan or Delta Air Lines’ Harts-field-JacksonAtlantaInternationalAirport.Identifyinghubsinsuchmarket is far from trivial. Our starting point for assessing airportconcentration of airlines’ networks in Brazil was done employinga survey. This survey aimed to have a more structured idea of theorganization of airport operations and on the amount of potentialhubs in the carriers’ networks.We conducted an on-line survey with over 300 air trans-portation academics and professionals. As expected from most on-line surveys,the response ratewas notveryhigh as only 79 expertscompleted the questionnaire. The questionnaire consisted of twoparts.Inthefirstpart,weaskedforthebestdefinitionforahub:‘‘anairport for flight connections’’ or ‘‘a big airport’’. Over three-quar-ters of respondents picked up the first choice, with 18% answeringboth definitions, 3% stating the ‘‘big airport’’ option and only 1%answered none of these definitions. In the second part of thequestionnaire, a list of the top 40 busiest Brazilian airports fordomestic passenger traffic(data for May 2008) werepresented andthe respondents were asked to select: (1) those airports that theyconsidered as a hub; (2) and those that, according to their opinion,should be a hub.Fig. 3 presents the results for the second part of the survey,showing that the majority of respondents believe that only threeairports are true hubs, actually: the ‘‘Top-2’’ domestic airports of Brası´lia (BSB) and Sa˜ o Paulo/Congonhas (CGH) – both named inover 90% of responses – and Sa˜ o Paulo/Guarulhos (GRU) – cited inover 70% of responses. While Brası´lia airport is more centrallylocated in geographical terms, Congonhas and Guarulhos serve thecity of Sa˜ o Paulo, one of the largest metropolises in the world, andthe heartbeat of the Brazilian economy. Guarulhos is the country’smost important international airport with 8.21 million interna-tional passengers in 2006, accounting for 67.4% of the 12.18 millioninternational air passengers in Brazil in the same year.The result of three appointed hubs is significant to thisproportion. Even if we considered ‘‘the majority’’ as being repre-sentative of more than up to 44% of respondents, we would stillobtain the same result, as the fourth most frequently namedairport, Tom Jobim/Galea˜o (in Rio de Janeiro) was considered as anactual hub by 43.5% of the participants in the cross section (seeFig. 3).In terms of whether those airports should be or should not behubs, again the majority of respondents stated only three airports:Brası´lia (over 85% of responses), Guarulhos (over 65%) and Tom Jobim/Galea˜ o, the international airport in Rio de Janeiro with 2.15million international passengers in 2006, and named in nearly 70%of responses. It is interesting tonote that78% of respondents statedthat Congonhas, the central business district airport in Sa˜ o Paulo,shouldnotbeahub.Thisclearlyindicatesthatthebusiestairportinthe country (accounting for 18.46 million domestic passengers in2006, which represents 20.5% of the total domestic passengertraffic for that year) should have its importance in the networkreduced.Indeed,thisendeduphappeningafterthefatalaccidentin Fig. 4.  Number of hubs in Brazil according to the FAA methodology. Fig. 5.  Evolution of concentration according to alternative methods. 1 We contrasted our results with the traditional Gini index approach and did notfind significant changes in the main conclusion regarding the increase in airportconcentration in Brazil during the period analyzed. T.F.G. Costa et al. / Research in Transportation Economics 26 (2010) 3–11  7
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks

We need your sign to support Project to invent "SMART AND CONTROLLABLE REFLECTIVE BALLOONS" to cover the Sun and Save Our Earth.

More details...

Sign Now!

We are very appreciated for your Prompt Action!