What hurts the dominant at hub airports?

Barcelona, May 25th 2015

Xavier Fageda Associate Professor ([email protected]) operate average location • airports • airports KLM • airlines • Introduction A While In But loss 2002 ……… - Transavia in have … inter that choices air - . 2013 . the ) Airport More in fares Europe, have typically - hub competitiveness , , network traffic, . dominance traditionally Bankrupcy competition Also, this benefited more downward airlines’ is changing and intercontinental of of been from may the share high Malev dominated pricing (examples dominant weak be route has , intense, Ryanair competition pressure fallen concentration ; network destinations by Difficulties at in former and their 17 of with airlines of LCCS Norwegian flag own may of 22 : LCCs Impact Iberia carriers on large hubs lead . routes & European to on in network . , higher firm’s main they

GiM - Departament de Política Econòmica | Facultat d’Economia i Empresa Av. Diagonal, 690 08034 | T. (+34) 93 402 19 43 | www.ub.edu/gim 2 pair are • their frequencies • they • competitive But NOT • Introduction Questions Aims Hence, I identify ……… competing levels to are hub examine . able to I airports, ? the estimate determine Do offered position Does to with competitive network provide welfare competition using low by of an which - dominant cost European on data implications airlines equation a position airlines? given dimensions for takes the reduce airlines network route in period of of place which LCCs each . frequencies of at 2002 airlines at competition entry hub the the - airports 2013 in dependent route on hub by at . routes the their , might airports airport flight hubs variable departing undermine and/or frequencies when is from they city the the -

GiM - Departament de Política Econòmica | Facultat d’Economia i Empresa Av. Diagonal, 690 08034 Barcelona | T. (+34) 93 402 19 43 | www.ub.edu/gim 3 network An Frequencies equation I • airports • Luo, 2002 • Contributions Literature add airports Literature Syverson Literature important 2013 ; Richard, to ) airlines : this , Merger on ( HHI secondary_airport 2008 Bilotkach Focus Dummies 2003 on literature on airline difference at ; the the Bettini (route, ; their on Pai Delta , effects Fageda , impact frequency route 2010 by for and hub & with airport including Northwest ; Oliveira, of competition multi & airports Bilotkach of Flores mergers previous competition LCCs - ), airport - 2008 . Fillol , the Share_LCCs Fageda on on ) . ( , following HHI_route studies 2013 capacity cities, the & at ) distribution Flores the presence is variables (route, competition , - the route D Fillol LCC , focus . D , level 2010 hub of of airport in ) traffic on ; the LCCs ( Schipper ( Brueckner Goolsbee choices ), frequency between in D merge et both and and al of . r , ,

GiM - Departament de Política Econòmica | Facultat d’Economia i Empresa Av. Diagonal, 690 08034 Barcelona | T. (+34) 93 402 19 43 | www.ub.edu/gim 4 • • • • • Sample and SAS Finnair (Vienna), Air I restricted The MAN, one A period I from Route have include number (Stockholm Tarom exclusion dominant 2002 - BCN) complete (Helsinki), level of British 22 () study sample (Paris of to or data large - large 2013 of ARN, because Airways airline - and CDG, KLM data important from EU covers . Copenhagen, EU . at (Amsterdam), airports Paris for operating which large (London airports they 936 most - Orly European EU are that routes at ), - of LHR, Oslo are airports dominated Air which out hubbing LOT dominant - not London Lingus OSL), and of (Warsaw), airports the included them to the SN operations - (Dublin), same LGW), European by airline Brussels sample for from LCCs Lufthansa airline because Czech the was Alitalia (PMI, our (Brussels), contains in and whole Airlines not was EU sample (Frankfurt, TXL, we non (Rome . a dominant low 11 period Swiss were - LTN (), European , is - 232 - FCO), cost , a (Zurich), Munich, STN unable limitation observations (BUD, carrier throughout Austrian ) Iberia destinations TAP Dusseldorf), to MXP, . (Madrid), identify but (Lisbon) Airlines ATH, . our the

GiM - Departament de Política Econòmica | Facultat d’Economia i Empresa Av. Diagonal, 690 08034 Barcelona | T. (+34) 93 402 19 43 | www.ub.edu/gim 5 • • • • • Descriptivedata former flag airlines Overall,it of non number data analyzedintercontinental on haveflights)airlines dominantbut the reduced reductiontotalOnlyin of non number growth Evolution excludes airportsthat of sample our that of those are base home the substantially17 in reduced share Mean 14 reducedin been has Traffic Mean (HEL, LIS, OSL, OTP) OSL, LIS, othersand (HEL,recording losses (ARN,BRU, MAD). at European hub in totaltraffic quite isdiverse airportsrecordingsome with substantial is not clear the aggregatednot clearis effect increasing ofthe presence of carriers. of networkof shareairlinedominant the of - stop22 airports. destinations 8 ofthe in of airlines > 50% with some some withfew50% > airlines exceptions. been has It airportsbut theymay 22 airports. of the 22 airports.is Thisquite the 22 remarkableof weif consider - stopZRH(no destinationsFRA and LHR,in alwayscases in60%. It some > and > 30% have weakenedthe of dominance the LCCs LCCs

GiM - Departament de Política Econòmica | Facultat d’Economia i Empresa Av. Diagonal, 690 08034 Barcelona | T. (+34) 93 402 19 43 | www.ub.edu/gim 6 • • • Network trend Currently TUIFly NIKI, FlyNordic Air, LCCs Network bmi : Norwegian, : Air , Air Virgin baby, , airlines Arabia some Berlin, Germanwings vs Express, Cai Low cost airlines hybrid Maroc : Germanwings OLT, Former Second, Ryanair Volare , LCCs , Air Hapagfly flag Finland, (fare , Centralwings , Vueling carriers SkyEurope , Norwegian, bundling , Jet Air , and/or 2 Wind , One, Jet , , Smart and/or Condor, 4 Vueling Jet, you, airlines Air Wizzair Wings, Transat, LTU, connecting Corsairlfy integrated Monarch, , Sterling, Zoom Air , Berlin, flights) Air dba, in Thomson, MyAir alliances easyjet Alpi but , MyTravelLite Eagles, very Transavia , Fly recent Blue Me, , ,

GiM - Departament de Política Econòmica | Facultat d’Economia i Empresa Av. Diagonal, 690 08034 Barcelona | T. (+34) 93 402 19 43 | www.ub.edu/gim 7 • • • Datalimitations competition of schedule, Enter No help) Some origin connecting on Under the data pricing in and potential hub amount on the 4 destination - ) and passengers prices strategic Prices variables market - of Similar Airline spoke omitted Aggregated Results traffic (most or . decisions data structures, explanatory behavior not of related factors flexible the ; of 2 supply ) passengers analysis Once : with can of variable) Airport the variables airlines direct entered, data be service should viewed . . congestion ; Focus and I for levels cannot can routes be connecting price as on complemented be of a previous and and (airport distinguish airlines multistage indirectly frequency aircraft passengers (and and in the with between considered decision technology, this) year equations route studies study dummies will process direct is 3 that : through depend ) on impact Flight may 3 and : use ) 1 )

GiM - Departament de Política Econòmica | Facultat d’Economia i Empresa Av. Diagonal, 690 08034 Barcelona | T. (+34) 93 402 19 43 | www.ub.edu/gim 8 Variables Frequencies Population Income D Distance Variable US_openskies D EU destination dominant_airline destination Dummy Dummy Metropolitan after Total annual of number Number flightsoffered the by Index_country NUTS 3(Europe), NUTS 3(Europe), dominant airlinedominant the Description for for OS of US intra kms agreement areas destinations - EU (Rest) flown (Rest) routes Intermodal Expected More restrictions Dependent variable Dependent larger integration restrictions No No Demand Demand competition regulatory regulatory planes ( planes relationship competition , stronger (+) (+) (+) (+) - ) ( - ) ,

GiM - Departament de Política Econòmica | Facultat d’Economia i Empresa Av. Diagonal, 690 08034 Barcelona | T. (+34) 93 402 19 43 | www.ub.edu/gim 9 Variables D Hub_competition D interhub_same_alliance routes_with_multiple D origin_airports Variable origin_airport D D merger year Dummyfor routes airservices which in hubs of airlineshubsintegrated of the same in Number of airports in the sample thatsample airportsthe in of Number origin airports (LHR, LGW;(LHR,airportsorigin ORY).CDG, the dominant airline wasdominant airlinethe acquiredby are offeredsimultaneously from two have directflightsto thedestination Dummyvariable forroutes which in Dummyfor routestwo connecting Dummies for Dummies of origin the airports another larger another airline. Dummies Description alliance for years airports larger of airline ( Controlfor time Controlfor common the airport Reorganizationof route Demand Expected Proxy for code networkfavorin of trend on allroutestrend on competition agreements ( Demand - variables specific omittedspecific (+), Inter(+), relationship (+) - invariant - share ( - - - ) ) hub

GiM - Departament de Política Econòmica- | Facultat d’Economia i Empresa )

10 Av. Diagonal, 690 08034 Barcelona | T. (+34) 93 402 19 43 | www.ub.edu/gim Frequencies HHI Variables HHI cost route Share_low cost origin_airport Variable origin_airport (+), route secondary_airport Share_low ( - ) - (+), ( - ) - airport market LCCS Number Share terms Herfindahl Share terms Herfindahl of of in from of of low low of - - Description the Hirschman Hirschman - - flight flight a cost cost flights nearby same airlines airlines frequencies, frequencies, offered secondary index index city - pair by in in - - connection passengers Less affect Less passengers - flights incumbent - More Competition A Less frequencies Expected demand possible other to demand boost demand times : routes Cut ( due relationship from demand - response (+) ) in fares : from a connecting route increased and of direct may add

GiM - Departament de Política Econòmica | Facultat d’Economia i Empresa an

11 Av. Diagonal, 690 08034 Barcelona | T. (+34) 93 402 19 43 | www.ub.edu/gim • variation • 2012 and within and and • • • • Econometricissues Wooldridgetest for Potential Lagged toModels considerunobserved routeheterogeneity: Route Levinlin Breusch RE imprecise do ; Troeger . variation not dependent panel data panel - PaganCook / endogeneity change , estimations 2008 Hausman RE FE but : : over Captures ) ……… C unit unit root test autocorrelation variable ontrol - of time Weisberg testfor . test potential when route for both but as not omitted competition low … a : NO PROBLEM: NO . between regressor effect useful bias within in panel data: panel inPROBLEM variables from of heteroskedasticity when - variation time and ? variables correlation Annual - within within that invariant correlate versus : use variation variation between variables of : PROBLEM random quarterly lags with while low explanatory cannot variables vs (Clark data, FE fixed only be low and effects of identified variables captures interest within Linzer, -

GiM - Departament de Política Econòmica | Facultat d’Economia i Empresa

12 Av. Diagonal, 690 08034 Barcelona | T. (+34) 93 402 19 43 | www.ub.edu/gim Results(RE) D routes_with_multiple Explanatory variables Population D Hub_competition Income interhub_same_alliance D Distance US_openskies D D merger EU destination destination origin_airports 3462.99*** 239.82*** 376.54*** 0.028*** - - 0.14*** 4.18** - 19.77* All(1) - 24.05 1.68 Dependent variable:Frequencies airlineDependent of dominant 3451.57*** Intra - - 7.09*** 0.10*** 29.28** 0.89*** 5.27** - 15.53 - - - EU (2) Intra 3766.04*** 11.55*** - 0.12*** kms 1.03*** - - - 32.53 32.79 5.27 EU - - (3) < 900 Intra > 900 1049.10*** - 9.07*** 6.94*** 0.10*** 0.47*** - - - 20.85 EU 0.07 kms - - routes (4)

GiM - Departament de Política Econòmica | Facultat d’Economia i Empresa

13 Av. Diagonal, 690 08034 Barcelona | T. (+34) 93 402 19 43 | www.ub.edu/gim Results(RE) Share_low Frequencies Explanatory variables Share_low HHI Intercept HHI origin_airport - R cost N secondary_airport route 2 - cost origin_airport route - - 294.48*** 206.36*** 162.33*** 10296 Dependent variable:Frequencies airlineDependent of dominant All(1) 18.14 - 0.44 0.04 2.20 - 1442.1*** 422.95*** - 196.19*** 205.47** Intra 15.32 6985 0.46 0.03 (2) - EU Intra 1143.40*** - - 492.96*** 533.54*** 313.15*** 0.13*** kms - 4114 - 14.94 0.45 EU (3) < 900 Intra > 900 1155.89*** 297.39*** 48.10** 145.49 - - - 2871 36.39 0.60 EU 0.04 kms routes (4)

GiM - Departament de Política Econòmica | Facultat d’Economia i Empresa

14 Av. Diagonal, 690 08034 Barcelona | T. (+34) 93 402 19 43 | www.ub.edu/gim Results(FE) • route • Frequencies • • R No Same Stronger 2 : 0 statistical . 04 results statistical secondary_airport effect for significance of HHI HHI airport route of in , the regressions Share_lowcost merger variable for all route sample , Share_low_cost and long - haul originairport intra - EU ,

GiM - Departament de Política Econòmica | Facultat d’Economia i Empresa

15 Av. Diagonal, 690 08034 Barcelona | T. (+34) 93 402 19 43 | www.ub.edu/gim Caveatson results airport week carrier • flights achieve An route about In than • F I t requencies terms airline seems the compared implies . a in . LCCs 2 a impact the of % needs 10 that decrease elasticities % route in about offered increase the secondary of to to ( their the provide 3 impact a weekly in 1 at : dominant % frequencies, in a share secondary decrease 10 that airports of 18422 % flights) in the increase airport the carrier annual share in provide to airports route while frequencies achieve while frequency in of flights ……………… a the low an 10 an are % such average - share cost airline in . too increase the of increase carriers small 4 of airport needs flights frequency LCCs in to the in in to per ( in impact 355 the the share provide the day of weekly route airport airport two in the of the . 175 LCCs flights dominant flights) is primary implies annual higher in per the to

GiM - Departament de Política Econòmica | Facultat d’Economia i Empresa

16 Av. Diagonal, 690 08034 Barcelona | T. (+34) 93 402 19 43 | www.ub.edu/gim Conclusions could network achieving subject • reduction routes their • Implications Dominant hub ensure on to and airports the in which airlines restrictions The An demand low that Negative best for increase introduction - they cost they possible airport may from airlines compete effects go . be A to of connecting management use worried reasonable the . airport of suffered of directly, market airlines the by current charges passengers the by - : able based objective but Capacity dominant increased also to capacity mechanisms may make on of expansions have airlines other presence the airport by airlines best a routes in are differential use managers the of at not operating low of large allocation that only them - cost may airports impact felt should . airlines in suffer of hubs on slots the are on be at . a

GiM - Departament de Política Econòmica | Facultat d’Economia i Empresa

17 Av. Diagonal, 690 08034 Barcelona | T. (+34) 93 402 19 43 | www.ub.edu/gim