Aviation Strategy
Analysis What is the future of the European flag-carrier model?
uropean aviation is beginning a process of trolled both carriers, had undertaken a disastrous Emajor restructuring. The competitive land- conglomerate strategy (code-named "Hunter"), scape of 2005 is likely to look quite different than withdrawing assets from Swissair and investing Aviation Strategy that of 1995. Serious flaws have developed in the them in independent airlines in other countries or is published 11 traditional business model followed by the major in separate service businesses. Swissair, Sabena times a year by carriers, and the future viability of that business and Crossair were separate operating companies Aviation model and those airlines has been openly ques- and were not directly exposed to the losses or lia- Economics tioned. bilities of the other airlines and service compa- at the beginning of the month This article is designed to provide a frame- nies. work for the better understanding of several ques- Network Management for Swissair and Editor: tions: Sabena (senior management for the two airlines Keith McMullan • What drives competition between the major had been combined in 1999 but remained sepa- European carriers? rate from the holding company) knew that both Subscription • Why some European hubs have been more carriers were in a highly vulnerable position, given enquiries: profitable than others? the industry-wide profit declines and the obvious Julian Longin Tel: +44 (0) 20 • Why the profitability of these carriers (as a failure of the outside investments. Airline man- 7490 5215 group) collapsed long before September 11? agement also knew that fleet decisions that SAir • What business models might provide a basis for Group had imposed on the airlines would reduce Copyright: profitable operation in the future? future profitability by hundreds of millions of dol- Aviation • What path is industry consolidation/restructuring lars. The fleet decisions had been driven by con- Economics All rights reserved likely to follow? glomerate objectives (including the development Discussions of airline business models, of an aircraft leasing company) and without any Aviation demand segmentation and detailed traffic flows real reference to whether the aircraft could be Economics can sometimes seem a bit dry and academic. The operated profitably within the Swissair or Sabena Registered No: data and analysis presented here was originally networks 2967706 (England) developed in 2000, in the decidedly non-academ- Assuming (heroically, as it turned out) that the Registered Office: ic context of two airlines whose survival was high- airlines could somehow be reorganised indepen- James House, LG ly uncertain. dently of the SAir Group conglomerate invest- 22/24 Corsham St The challenge facing Swissair ments, Network Management undertook a major London N1 6DR internal study in 2000 to address two questions: VAT No: 701780947 and Sabena in 2000 • Could either Swissair or Sabena survive long ISSN 1463-9254 In 1999, Swissair had a minus 4% profit mar- term, given competitive changes across Europe? The opinions expressed in gin while Sabena had a minus 6% margin, a bit • What future business model and short-term this publication do not nec- essarily reflect the opinions below AEA averages, after having earned small changes would give the greatest chance of sur- of the editors, publisher or contributors. Every effort is profits the previous two years. These declines vival? made to ensure that the mirrored downward profit trends among airlines This article will outline one view of the com- information contained in this publication is accurate, but across Europe. Both airlines were financially petitive and profitability issues facing all of no legal reponsibility is accepted for any errors or healthy, in the sense of having strong positive Europe's large airlines, as seen from the per- omissions. cash flow, easily meeting all current obligations spective of these two struggling mid-sized carri-
The contents of this publica- and having much of their networks earning fully- ers, based on data available at the time (1999- tion, either in whole or in part, may not be copied, allocated profits. Although no national airline in 2000) of the study. It is not intended to provide a stored or reproduced in any Western Europe had ever failed before, both car- complete discussion of the events leading to the format, printed or electronic, without the written consent riers were destroyed and liquidated within eigh- of the publisher. teen months. By Hubert Horan; Questions and comments to SAir Group, the holding company that con- [email protected]
September 2002 Aviation Strategy
Analysis
demise of Swissair or Sabena. While the specific recommenda- EUROPEAN BUSINESS MODELS MID 90s tions developed for the two carriers two years ago are of no more than Charter historical interest at this point, the question of what drives hub prof- • Narrow focus on one itability in Europe, and the viability O&D demand segment Flag 1st of competing business models • Limited fleet/marketing Div remains highly relevant. • Network ubiquity • Home market dom- The classic "flag-carrier" inance business model Flag 3rd • Multiple market Div segments The two classic European air- • Mixed fleet/prod- line business models were the Flag uct/marketing for "flag-carrier" model, which was 2nd Div diverse markets designed to operate at a large scale and serve a very broad range of potential customers, and Note: Market shares based on ASKs the "Charter-carrier" model, which was designed to only serve a specific, narrow Lufthansa followed in Frankfurt. demand segment. The flag-carrier model, best European aviation will always represented by Lufthansa's Frankfurt hub-based network, was adapted by almost every scheduled be highly fragmented airline from Portugal to Finland, and has five key European air travel demand has always been features: extremely fragmented due to heterogeneous • Domination of travel demand from the carrier's national markets, huge disparities in disposable home market; income levels and market sizes, strong distinction • Service to multiple, diverse demand segments between leisure and business destinations, and (business/leisure, domestic/intra-Europe/ inter- wide disparities in transport alternatives. It is continental, home market/sixth freedom) to max- unnatural for any one business model to become imise total travel volumes; the overwhelming standard across such a hetero- • Large US-style hub operations in order to aggre- geneous marketplace. The central position of the gate demand from dispersed markets; flag-carrier model was heavily influenced by reg- • A mixture of different aircraft sizes in order to ulatory and aeropolitical constraints and has maximise the frequencies offered; and already begun to break down. Airlines such as • Significant marketing infrastructure (such as Ryanair are attempting to develop new leisure ori- worldwide sales and distribution) and systems ented markets (Stansted to Rimini or Biarritz) out- complexity (yield management, airport opera- side of the traditional charter model, while tions) to efficiently serve the diverse markets. easyJet and others are developing more busi- Charter carriers aggregated demand via spe- ness-oriented O&Ds while avoiding the compre- cialised pricing, packaging and distribution, and hensive scope and infrastructure intensity of the organised operations around larger single-class traditional flag-carrier model. Where demand is aircraft with lower unit costs, and only served highly fragmented, it is normal for companies to O&D markets that fit into this approach. experiment with new or modified models, and it As late as 1995 the central strategy question should be possible for multiple, overlapping busi- for European airlines was scheduled versus char- ness models to successfully serve different seg- ter. Once a airline chose the "scheduled" path, it ments. then pursued every logical source of demand in Airline business models are demand driven, order to maximise traffic volumes and scale. Smaller markets produced smaller airlines, but not cost driven they all followed the same business model that Because of the fragmented demand base, the
September 2002 Aviation Strategy
Analysis
The First Division market size advantage most critical strategic issue for any airline is net- work ubiquity versus a narrow focus-the decision There is a marked difference in the size of the to serve multiple, diverse traffic flows or to con- local revenue base between the four First centrate on one specific market segment. Cost Division hubs and the ten Second Division hubs. structures must be then carefully tailored to the The CDG market is three times larger than target market, but it is dangerous to segment air- Zurich, Brussels or Munich, while Heathrow is six lines on the basis of concepts such as "low cost". times larger, and these gaps would be even larg- There is no such thing as a "high cost" business er if one considered total London/Paris demand model. Classic charter carriers avoid many of the instead of the airport level demand. This size branding, CRS, and hub airport costs that British advantage of the ASK capacity operated by the Airways and KLM must bear, but as a result they First Division hubs mirrors the differences in the cannot efficiently serve more diverse scheduled underlying revenue bases. This is in marked con- markets or scale their operations to a large net- trast to the US hub environment where origin work size. Ryanair's approach achieves low costs market size gaps between the top tier hubs on Stansted-Ireland routes but would be uncom- (Atlanta, Dallas, Chicago) and second tier hubs petitive on Heathrow-Austria routes. Airlines (Houston, Denver, Pittsburgh, Philadelphia) are under any business model will fail if they add too much smaller. This also explains why new much capacity relative to their target markets, or entrants using 140-seat aircraft have had suc- cannot keep costs in line with what those markets cess developing networks of large O&D markets will pay for. ex-London, and much less success at Brussels, Three segments within the "flag-carrier" Munich or similar cities. The First Division offers two to three times the business model level of Intercontinental departures than Second In 2000, the 14 largest hubs accounted for Division hubs, even though levels of intra-Europe 91% of all capacity operated within the entire service are broadly similar. SAS at Copenhagen scheduled European industry. The industry's had roughly the same number of short-haul flights "First Division" -the top four hubs (British Airways as British Airways at Heathrow. Second Division at Heathrow, Lufthansa at Frankfurt, Air France at carriers would have to quadruple their ASK Charles DeGaulle and KLM at Schipol) alone capacity in order to match the size and operating accounted for 55% of scheduled industry capaci- scale of the First Division. ty. Flag-carrier hub profitability fell by These 14 First and Second Division hubs strictly follow(ed) the classic "flag-carrier" busi- over $1bn in 1999 ness model, with service to a very broad range of One cannot evaluate competitive perfor- European and intercontinental destinations, and mance or strategic issues without reference to dominant home market network and distribution relative profitability. The table below includes positions. Many of the 9% of scheduled ASKs in Swissair estimates of the operating profitability of the Third Division are also tied to the flag-carrier the 14 hubs in its competitive set. These reflect model, including smaller national carriers with educated guesswork based on public financial much more limited networks (Finnair, Aer Lingus, information and should be used with the appro- Malev, TAP), and secondary hubs of larger carri- priate grains of salt. European airlines do not ers (BA at Manchester, Iberia at Barcelona, Air have detailed public yield and cost data available France at Orly, Crossair at Basel), designed to and thus cannot easily estimate their competitor's maximise Home Country coverage. But this Third route and hub profitability as US airlines can. group also includes new entrants following differ- In 1999 the profitability of these 14 hubs fell ent models, including domestic-focused hubs by over US$1bn (a 45% decline) from 1997-98 (AOM-Air Liberte at Orly, Deutsche BA), and the levels, a downward trend that continued in 2000 satellite London operations (easyjet at Luton, and 2001. More importantly, there is a structural Ryanair at Stansted). In 2000, these new entrants profit gap between the First and Second division were still a tiny percentage of scheduled industry hubs. In the "good years" of 1997-98, the top four capacity. hubs operated 61% of the capacity of this group
September 2002 Aviation Strategy
Analysis
FLAG CARRIER HUB PROFITABILITY While capacity grew roughly Hub ASK Estimated Estimated Estimated Capacity 97-98 hub 97-98 hub 99 hub 35% in this peri- Rank/ Index Op Profit Op Margin Op Profit od, industry-wide (FRA=100) revenues (adjust- LHR—BA 1—125 $750 m 5-7% $150m ed for inflation) First FRA—LH 2—100 $550 m 7-9% $500m Division hubs CDG—AF 3—92 $300 m 3-5% $500m barely grew at all. 55% ASKs AMS—KL 4—82 $400 m 3-5% $100m While there were Subtotal $2.0 bn ~4% $1.3 bn isolated cases of
LGW—BA 5—46 $75 m 2-4% ($25m) profitable growth ZRH—SR 6—41 $75m 2-4% 0 (Air France's hub Second MAD—IB 7—36 $100 m 3-5% $50m development at Division hubs MXP—AZ 8—34 $75 m 2-4% ($25m) 36% ASKs BRU—SN 9—29 $50 m 0-2% ($25m) Charles CPH--SK 10—22 $75 m 3-5% $25m DeGaulle), in the FCO—AZ 11—18 $50 m 2-4% 0 vast majority of VIE—OS 12—17 $50 m 2-4% 0 MUC—LH 13—17 $50 m 1-3% $25m cases this expan- ARN - SK 14-14 $0 m (1)-1% 0 sion destroyed Subtotal $0.6 bn ~2% $0.1 bn corporate value. Within the but earned 77% of the profits, a full two margin AEA averages, the smaller, weaker carriers were points better on average. As overall industry actually growing faster than the larger, more prof- financial performance declined, Second Division itable ones. The First Division airlines grew at an profits declined much more rapidly, falling 85% in overall average of 7%, adding a huge number of 1999 versus an estimated 35% decline for the new seats to the market, although BA and KLM First Division. Size matters. There is no evidence slammed the brakes on growth in 1999 once they in this time frame of any smaller flag-carrier hub realised how overexpansion was harming their earning more profits than any significantly larger shareholders. However, the Second Division car- one. riers grew 12% and AEA carriers in the Third While an obvious point, it is worth noting that Division group (such as TAP, Aer Lingus and 2-4% operating margins in the "good years" of a Finnair but not including Ryanair or easyJet) grew business cycle usually indicates that an industry at 13%. It is unclear whether any managers at has extremely serious structural problems and these carriers actually believed that their core could not support new capital inflows without business revenue base was likely to grow at 12- major restructuring. 13% type rates. But in their rush to emulate the Overcapacity and the industry profit collapse Frankfurt hub (and to narrow Frankfurt's size and network scope advantages) the flood of new of the late 90s capacity destroyed their corporate earnings. The tendency of airlines to expand capacity Flag-carrier revenue and competition much faster than the growth in revenue base can be observed throughout the industry's history, is driven by Intercon markets and was the primary driver of the overall decline 64% of all revenue originating or terminating in flag-carrier profitability in the late 90s. in Europe in 2000 was long haul, while only 25% The 14 AEA carriers grew capacity roughly 7- 1996 1997 1998 1999 8% per year in this period. As seen in the table European GDP growth 1.5% 2.4% 2.6% 2.0% below, Intercon traffic grew in line with seat AEA INTERCON ASK growth 8.7% 10.5% 6.4% 8.0% growth, while intra-European traffic growth RPK growth 8.1% 7.5% 7.6% 8.7% lagged slightly behind. However, Intercon traffic Real RASK growth (8.2%) (5.5%) (6.7%) (8.5%) growth was almost exactly offset by 7-8% real AEA INTRA-EUROPE yield declines -the added seats were only filled by ASK growth 6.4% 7.7% 10.0% 5.5% RPK growth 7.3% 5.3% 7.9% 5.8% cutting prices. The real yield declines were almost Real RASK growth (6.8%) (6.8%) (3.8%) (4.6%) (but not quite) as bad on the short haul network.
September 2002 Aviation Strategy
Analysis
% of total Revenue All markets O&Ds with Connect-only lengths. Yield penalties on sixth freedom markets base by O&D nonstop O&Ds category service averaged 15-20% for Swissair and 20-30% for Total 100% 65% 35% Sabena. Hub profitability requires a strong mix of Intercontinental 64% 32% 32% home market traffic relative to lower-yielding sixth Europe cross-border 25% 21% 3% freedom traffic. Domestic 11% 11% 0% Excess capacity has a disproportionately came from cross-border short-haul markets. The greater impact on sixth freedom markets. Carriers remaining 11% was from domestic, intra- can avoid price wars in Home Country business Scandinavia or UK-Ireland markets which the top markets such as Paris-Frankfurt, by filling empty 14 hubs only serve on a very limited basis. seats with Miami-Frankfurt or Paris-Bangkok pas- Competition between the top 14 hubs is dom- sengers instead. inated by the dynamics of Intercon markets. 32% Three hubs Sixth Freedom Traffic as % of the total flag-carrier revenue base comes from which had the mis- of Total Hub Traffic (1999) Intercon markets that have no nonstop service. fortune of being LHR—BA 18% Many of these markets (Berlin-Los Angeles, located in smaller Lyon-Tokyo, Sao Paulo-Copenhagen) have good national markets - FRA—LH 27% single carrier one-stop service from five or more Amsterdam, Zurich CDG—AF 20% competing European hubs plus alliance or inter- and Brussels - had AMS—KL 44% line service via non-European carriers. Long haul extremely high per- LGW—BA 15% markets with nonstop service (another 32% of the centages of lower- ZRH—SR 34% total revenue base) will still face significant com- yielding sixth free- petition from hubs offering connecting service, dom traffic, and are MAD—IB 7% given the price sensitivity of these markets, and especially exposed MXP—AZ 13% the relatively small elapsed time penalty. to revenue down- BRU—SN 36% Given the power of hubs, intra-Europe non- turns due to indus- CPH- SK 9% stop markets tend towards stable one or two car- try-wide overcapac- rier markets. With typical 1hour 30minute narrow- ity. Other Second FCO—AZ 2% body block-times, connecting alternatives cannot Division hubs had VIE—OS 24% compete without very deep price cutting. Intra- lower percentages MUC—LH 11% Europe O&Ds that do not have nonstop service, of sixth freedom Average of where passengers must connect (Bari- traffic but also had 13 hubs 21% Amsterdam, Stavanger-Rome) tend to be weaker Intercontinental networks. extremely low-demand markets, and are insignif- Second Division hubs cannot compete icant to the larger competitive picture. In the US, narrowbody must-connect markets account for for Intercon traffic over a third of the domestic market, while in Second Division hubs with long haul networks Europe the thousands of O&Ds in this category face a daunting challenge. Intercontinental wide- only account for 3% of the total revenue base. bodies can feed huge volumes of traffic on their Hub profitability depends on a limited mix existing short haul flights, rapidly increasing the scale economies of their hubs, and significantly of low-yield connect traffic improving the scope of their network to the levels 79% of all traffic served by the top 13 hubs business customers demand. But by competing originated or terminated in the hub carrier's home directly with Frankfurt, Paris and Amsterdam for market, 68% at the hub city, and 11% on con- this traffic, they end up with an unsustainable mix necting domestic services such as Hamburg- of sixth freedom connecting traffic, and the Frankfurt on Lufthansa or Toulouse-Charles expansion needed to maximise network competi- DeGaulle on Air France. Average yields on home tiveness disproportionately trashes their own rev- market connect flows are generally similar to hub enue base. None of the Second Division hubs nonstop levels, Swissair yields on connecting operating in 2000 could match the market share Geneva/Basel traffic was actually higher than of the First Division hubs in the Intercon market, nonstop Zurich yields at comparable stage and none of the Second Division Intercon opera-
September 2002 Aviation Strategy
Analysis
tions were profitable in the late 90s. Traditional flag-carriers such as Austrian and SAS who had Share of Intercon Share of Intercon recognised this dilemma and backed away from Departures Sixth Freedom direct competition with Frankfurt and Paris (Winter 99-00) Connect Traffic remained frustrated by the inability to exploit the (MIDT 99) growth and scale economies of the major wide- LHR—BA 18% 17% FRA—LH 15% 18% body operators. CDG—AF 17% 14% As the table below indicates, the ability to AMS—KL 14% 23% attract sixth freedom Intercon traffic is a direct LGW—BA 7% 6% function of the number of Intercon departures ZRH—SR 8% 7% offered. The only hub to capture a disproportion- MAD—IB 4% 3% ate share of this traffic is Amsterdam, which has MXP—AZ 4% 2% BRU—SN 5% 5% long been the industry leader in this segment, CPH--SK 3% 1% and developed the first strong alliance network to FCO—AZ 2% 2% serve interior US points and the first and VIE—OS 2% 1% MUC—LH 1% 0% strongest network to feed regional UK traffic through a large hub. No one in the Second pete without an alliance partner (albeit at a slight Division has been unusually successful in attract- disadvantage) as the larger UK-US market can ing connecting flows to their hub. support more direct service, but no continental Aircraft economics are also key since unit hub could remain in the First Division without an costs climb as total seats decrease. Carriers with immunised US partner. While many observers at smaller home markets must rely on smaller the time were expressing doubts, the study 767/A330 type aircraft, and thus lack both the argued that there was no question about the long- capacity to serve a larger share of sixth freedom, term viability of KLM's Amsterdam position. Its and the low costs needed to serve lower-yield smaller home market would always limit its poten- passengers profitably. If a 200-seat long range tial profitability (relative to Paris or Frankfurt), but aircraft with 747 unit costs existed, Second its decades of experience with hub and alliance Division hubs would be much stronger competi- management and other factors provided offset- tors. ting strengths, and it would clearly benefit from How many Intercon hubs can Europe support? the inevitable shakeout of sixth freedom capacity. A BA-KLM merger would have had a huge The Swissair/Sabena study argued that all 14 impact on European hub competition by allowing hubs were viable and sustainable, but not if they one company over 45% of the sixth freedom followed the traditional flag-carrier business Intercon market, more than double the share of model and developed networks directly competi- Paris or Frankfurt. tive with Frankfurt and Paris. One could operate Observed (2000) long haul sixth freedom long haul from a European hub with a level of connect traffic had been inflated by unsustain- departures competitive with Frankfurt and Paris, able discounting, but there was at least the pos- or maintain a very limited service strictly support- sibility that the market could support a fifth flag- ed by the local market similar to Austrian at carrier type hub, although this was by no means Vienna or Alitalia at Rome, but there was no certain and would depend heavily on the avail- viable "in-between" strategy. ability of a strong US alliance partner and exact The number of viable flag-carrier model path of industry restructuring. 27% of the Intercon hubs was limited by the total pool of traf- observed sixth freedom traffic already used fic from cities without non-stop service and the Second Division hubs; a viable fifth flag-carrier number of strong, immunised alliance with hub would need to achieve a 10-15% share with domestic US carriers. The North Atlantic the same (or better) yields achieved by the First accounts for 54% of the total long haul market ex- Division hubs. Europe, and 20% of the US market can only be Swissair as Europe's fifth Intercon hub? served in conjunction with an alliance partner. British Airways at Heathrow could clearly com- It was readily accepted that Sabena had no
September 2002 Aviation Strategy
Analysis
business expanding its Intercon network, and the Given the current profit squeeze and industry decision to double Sabena's long-haul fleet from overcapacity, any expansion not clearly profitable 10 to 20 aircraft (imposed by SAir Group in 1997 would be cancelled, including the A340-600 air- when effective control was acquired) needed to craft on order which were much too large for be totally reversed. None of the aircraft added Swissair's markets. Profit recovery would depend had operated profitably. It was argued that on a rapid shakeout of other unprofitable Second Sabena could profitably support perhaps 8 to 12 Division Intercon capacity-not only a downsizing long-haul aircraft, (the pre-SAir Group level) limit- at Brussels, but also Gatwick and Malpensa, ed to large destinations such as New York and gauge reduction at Heathrow, and no new expan- Boston, American Airlines hubs, and traditional sion at Copenhagen or Munich. It would also markets in Francophone Africa. As will be dis- require strengthening the alliance with American cussed below Sabena had a clear opportunity to Airlines to the level achieved with the previous abandon the flag-carrier model, and to restructure alliance with Delta. SAir Group had allowed the along a "City Network" business model approach. Delta alliance to collapse as alliances and acqui- Swissair's dilemma was much more difficult, sitions with Third Division European carriers were as both growth and contraction seemed highly deemed more important. Any BA-KLM merger, or unattractive. The size of the Swiss market and the US-UK open skies leading to a fully immunised limitations of Zurich airport precluded any expan- BA-AA alliance would have destroyed the sion towards the size of the First Division net- prospects for this approach. works, but all evidence suggested that competi- The alternative most in line with the changing tiveness would decline rapidly if Intercon service competitive situation was to downsize to an were cut back to a smaller scale. But 55% of all almost exclusively short-haul network following Swissair revenue came from Intercon traffic (ver- the Sabena approach. This would have required sus 30% for Sabena) and Intercon was a huge the elimination of over 25 widebodies and replac- percentage of Swissair's asset and marketing ing at least 20 of Swissair's A320s with smaller base. Swissair operated 36 long haul aircraft and aircraft as the loss of long haul feed would require only seven of these aircraft were operating a complete downsizing of the short haul fleet. It unprofitably in 2000 and all but one was clearly would have eliminated 80% of the jobs under the cash positive. Swissair had a reputation for ser- current Swissair pilot contract. The risks of such vice quality that allowed it to compete successful- radical downsizing were huge and no airline in ly for higher-yield sixth freedom traffic against history had ever gone through a restructuring carriers with larger networks. In fact, the value of remotely similar. All current operations were cash Swissair's brand almost exclusively came from its positive yet any restructuring would have immedi- ability to shift revenue share in competitive mar- ately drained cash and required massive new kets outside Europe. Swissair was respected investment. within Europe, but (adjusted for stage length) its While survival as the smallest of Europe's short haul yields were exactly comparable to First Division global hubs was highly uncertain, those of Sabena or Crossair, somewhat less and many critical requirements were outside famous brands. Swissair's control, this was the shareholders' The study concluded that there was no logical least risky near-term option. Of course the viabil- basis for arguing that Swissair could survive long- ity of this approach would need to be regularly term and earn reasonable returns, even if all of revisited as the actual industry shakeout, and the conglomerate financial problems external to SAir Group's financial restructuring proceeded. the airline could somehow be solved. If one was The new SWISS Intercon strategy (2002) starting from a clean sheet of paper, one would never invest in a global hub based in Switzerland. Of course, SAir Group refused to restructure The recommended approach, however, was any of its failed conglomerate investments and to maintain Swissair's 1999 level of Intercon oper- collapsed in 2001. SWISS became the successor ation and continue to try to compete directly with company to Swissair, using the Crossair corpo- the First Division hubs, making maximum use of rate structure, and acquired all of Swissair's route Swissair's brand equity and service reputation. authorities and other network operating assets.
September 2002 Aviation Strategy
Analysis
As SWISS obtained these assets without any Number of destinations with local Brussels demand (passengers per day, each way, MIDT) requirement to compensate the previous owners >150 100-149 75-99 50-74 35-49 20-34 <20 (the SAir Group creditors) they had the unusual 1 5 6 8 14 16 20 freedom to establish whatever fleet or network mix they wanted. seat aircraft would have difficulty developing a The new owners of SWISS invested large network. The high cost of short-haul flights US$2.5bn of new capital in a business plan at Brussels-National Airport would make it difficult whose centrepiece was an "in-between" size to profitably sustain the low fares that would be Intercon strategy. Zurich Intercon capacity was required to significantly stimulate new demand, cut roughly 25 percent (to a level similar to what and Brussels' appeal as a purely leisure destina- Alitalia operated at Malpensa in 2000), and the tion is limited. Virgin Express has been unable to historic Swissair brand name was abandoned. make money with its small route network. While they had the option of using their lower cost Ryanair's limited network is based at Charleroi structure as the basis for strengthening Intercon (where its airport costs are essentially zero) and service, pursuing a predominately short haul net- has not focused on traditional business destina- work based on regional aircraft, or a lower cost tions. approach based on a narrower set of target mar- The recommended "City Network" business kets, they chose instead to invest in a 26 aircraft model, builds a high frequency network for these long haul strategy. business destinations with a mixed, largely Thus the new owners of SWISS rejected all regional jet fleet, targeting a very small average three of the major findings of the 2000 gauge (75-90 seats) that reduces total ASKs. Swissair/Sabena study-that a short haul regional Under 2000-01 market conditions, the study aircraft based strategy was the best "clean sheet argued that Brussels could have supported 240- of paper" approach, that the largest feasible net- 280 flights, depending on the exact competitive work and the strongest brand would be critical if situation. Seat capacity serving sixth freedom one chose to continue to battle First Division hubs would be drastically reduced, along with the mar- for a strong Intercon position, and that an "in- keting and sales infrastructure serving these between" Intercon network was the worst of all diverse but low-yield markets. Longhaul and worlds, and was the least likely to achieve sus- mainline narrowbody aircraft would be limited to tainable profits. Whether this alternative strate- markets that can be profitably operated with gic approach can earn returns for the new strong reliance on local traffic (London, Malaga, investors is currently being tested in the market- New York, francophone Africa). Global connectiv- place. ity would have been provided in conjunction with A new City Network strategy for Brussels alliance partners (Swissair and American). Revenue would still be optimised with a hub The plan developed for Brussels abandoned schedule, but depeaking the existing Intercon-ori- key elements of the flag-carrier model, including ented Sabena schedule would have provided util- the emphasis on rapid Intercon traffic growth and isation gains enough to fund six or seven addi- sixth freedom connecting traffic. Brussels is a tional aircraft worth of flying. large O&D market, similar in size to Rome, Milan, Sabena - no chance to change direction Munich, Zurich and a bit larger than Copenhagen and Vienna. The study argued that Brussels (and While Sabena management accepted the the other Second Division cities) were fully capa- "City Network" recommendations, and worked ble of supporting large levels of airline service, actively to cut back long haul flying and to signifi- just not global hubs. cantly expand regional jet flying, It was unable Of the 70 European cities Sabena and DAT implement the change in strategy. Sabena had served in 2001, 50 had fewer than 50 local grown at an annual rate of 22% between 1997 Brussels passengers per day each way, and only and 2000 - three times faster than the ruinous six markets had more than 100 local passengers. AEA 7% average rate that had destroyed billions Thus an airline pursuing a Southwest or Ryanair in corporate value across Europe. SAir Group type strategy of serving markets that can fill 140- strategy for Sabena was to focus on intra-Europe
September 2002 Aviation Strategy
Analysis
connect traffic, even though this was only 3% of the company. industry revenue, and this traffic had no particular While Sabena's short-term profit outlook in reason to choose Brussels over other, larger 2000 was much worse than Swissair's, this was hubs. SAir Group's 1997 decisions to recklessly largely a function of the fleet and wetlease prob- over-expand, and to simultaneously replace lems. SAir Group had made disastrous aircraft Sabena's entire fleet with aircraft much too large investments at both airlines, but they hit the for its markets created impossible financial bur- Sabena P&L two years sooner. If one assumed dens and the company collapsed by the end of these obligations could be restructured, and one 2001. looked out to the European airline environment of In addition to doubling the long haul fleet, SAir 2005 or 2010, the study suggested that it was Group disposed of Sabena's fleet of 32 737s, more likely that one could operate a profitable (most of which were less than ten years old) City Network airline in Brussels than to make replacing them with larger A320s. This would money in Zurich as the number five global hub in have increased the average gauge of Sabena's Europe. overall narrowbody operation from 98 to 116 While the 2000 Swissair/Sabena study seats per departure, comparable to the level Air argued that the Brussels market could support a France operates in a local market three times the large local-service airline, investors have been size of Brussels, where it also has 50 long haul highly reluctant to step forward, and SN Brussels, widebodies a day feeding connecting traffic onto the successor carrier remains under-capitalised. those seats. The A320 is obviously a fine aircraft, As Sabena imploded, and was not reorganised in it was just totally inappropriate for Sabena's mar- an orderly manner, a significant chunk of its pre- kets. They would have increased annual costs by vious revenue base was lost to foreign airlines, over US$100 million (since they are newer and and perhaps that may have fatally undermined larger aircraft) and would have been totally the potential opportunity. Or perhaps investors dependent on incremental sixth freedom traffic simply do not believe that any European airlines (on top of an already bloated base) to cover those except the First Division global hubs and UK- costs. Because Airbus could not deliver this based new entrants can justify new private invest- added capacity fast enough, SAir Group added ment. wetlease capacity from Virgin Express and Five viable European airline business CityBird at rates over US$35m per year higher than Sabena's own costs, under unbreakable models for 2005 multi-year contracts, which accelerated the cash The central strategic question for European drain and subsidised otherwise unsustainable airlines in 2005 is whether to pursue all possible competitors. sources of demand (global, domestic, short-haul It would have been relatively simple to shift European, mixed business and leisure) in order to from Sabena's 1998 fleet and network position to maximise traffic volumes. If one does, one must a "City Network" type strategy, but there was no incur much higher marketing infrastructure and way to quickly reverse the financial burdens of operating complexity costs. These costs can only the SAir Group changes (fleet and wetlease oblig- be offset by the scale economies of 747s and ations, massive pilot retaining, overcapacity and Frankfurt-type hubs, which can only be realised yield declines, etc.). Press comment at the time by carriers with large, sustainable Intercon net- Sabena shut down tended to focus on longstand- works. The limited base of higher-yielding home ing issues such as brand image, or Belgian social market Intercon traffic suggests that only costs and industrial relations, but these factors Heathrow, Charles de Gaulle, Frankfurt and had almost nothing to do with the actual failure of Amsterdam can survive as global hubs, and so AF-CDG 120 only British Airways, Air France, Lufthansa and Average BRU after Airbus plan 116 KLM have the possibility of pursuing a multiple seats per SR-ZRH 113 intra-Europe BA-LGW 110 demand segment strategy. departure OS-VIE 101 Limited long haul service outside these hubs under hub BRU 1998 actual 98 will survive only when the local market can fill at carrier’s code KL-AMS 93 (summer 00) LH-MUC 87 least half of the seats and local business travelers BRU City Network plan 85 contribute a strong share of total revenue.
September 2002 Aviation Strategy
Analysis
Any European network that cannot enjoy Global Hub and City Network must carefully scale economies must target a narrower mar- limit discretionary low-yield traffic to off-peak ket segment, eliminate all costs not directly "fill-up" capacity, while managers at Big O&D serving that target market, and achieve short carriers must build their marketing and capacity haul operating costs much lower than current plans around these markets. BA or KLM levels. The "City Network" approach No successful airline has ever operated divi- outlined earlier targets local intra-European sions following totally different business mod- business demand and drastically reduces els. The management approach of one model capacity and operating costs by downsizing into always undermines the unique cost discipline or smaller gauge, largely regional fleets. The "Big market focus needed to succeed with the sec- O&D" model uses a standardised fleet of larger ond model. US carriers have repeatedly failed (737/A320 type) single class aircraft and then in attempts to either graft isolated pieces of the targets one of three categories of markets Southwest Airlines business model onto the tra- potentially large enough to fill these aircraft (tra- ditional "Big Hub" model, or to create an "airline- ditional business, tour package leisure, or within-an-airline" following the unique scheduled leisure O&Ds). Southwest model. These Business Models cannot "City Networks" target existing traffic but be combined have limited growth potential It is important to emphasise that any sus- The "City Network" model is designed to tainable airline business model must have lower serve already existing demand for (relatively costs than the traditional flag-carriers. No airline high priced) air service at (relatively high cost) can survive with bloated overheads, above- major airports. This minimises market develop- market input costs or an unfocused all-things- ment costs but means that this model offers to-all-people management mentality. But low very limited traffic growth potential. Natural cost is one of the keys to sustainable profits, not growth of the higher-yielding business revenue an end in itself. Cost efficiency is driven by total- base is probably less than 2% per year, with no ly different factors under each model, and in growth potential until the overcapacity of the each case certain costs must be higher in order late 90s is worked off. to generate critical revenue streams. Only large markets, such as the Second Because different costs play different strate- Division hub cities, can support the large, multi- gic roles in each case, airlines cannot follow ple-frequency networks needed to make this more than one model, or mix-and-match ele- approach work. City Networks at smaller cities ments of different models. The discipline and (Geneva, Hamburg, Nice) are easily over- skills needed to ruthlessly eliminate complexity whelmed by large jet capacity from competitor and infrastructure costs in a City Network or Big hubs, and it is much more difficult to build the O&D approach cannot also serve the diverse customer loyalty and competitive presence customer requirements of a sixth freedom hub. needed to maximise revenue performance. The focus on scale Driver of Low Costs Costs needed to generate higher revenue Global Hub • • economies Very low costs per ASK (widebodies, mainline Intercon sixth freedom revenue must cover jets, long stage lengths) higher airport, marketing and distribution costs and com- • Scale economies from very large home market • Costs of added services, premium products plex auto- revenue base and large Intercon traffic flows must be covered by higher fares mated tools City Network • Very low aircraft costs per departure (regional • Local customers must pay for quality schedule aircraft) at expensive airport—must have large local that are key • Reduced complexity and infrastructure costs revenue base and must achieve local market to Global (hubs, marketing and distribution) relative to RASK premium Hubs cannot flag-carrier levels • • be readily Big O&D Low aircraft costs per ASK (standardised fleet Must develop new, discretionary revenue via of all-coach mainline jets at high utilisation) combination of clever marketing and lower applied to • Extreme minimisation of complexity and pricing narrow local infrastructure costs • Each city pair must be large enough to fill larger markets. • Very low airport costs gauge aircraft Managers at
September 2002 Aviation Strategy
Analysis
This approach offers none of the glamour of a domestic markets, has been attempted repeated- rapidly growing long haul flag-carrier network, but ly without any success whatsoever. Every that is not one of the options in a market like Continental country with sizeable domestic O&Ds Brussels, Vienna or Rome. The "City Network" - France, Italy, Germany, Greece, Sweden, model offers an opportunity to make money the Norway, and Spain - is littered with downsized or boring, old fashioned way - by keeping costs in failed startups. In each case the small number of line with a more limited and stable revenue base, O&Ds, high airport costs, and other factors pre- and maximising the satisfaction of local business vented the new airlines from achieving large, sus- travelers with a strong, reliable schedule. tainable price and cost advantages. Go and The "Big O&D" model generates new easyJet achieved better results in the UK where domestic routes were a small piece of a broader demand but has difficult limitations network, and with a much larger and more easily "Big O&D" carriers face the marketing chal- segmentable capital city market. AOM, Air lenge of creating new demand that does not need Europa, Deutsche BA and Air One had credible to travel by air (or travel at all) and was not well operations but the same costs at the same air- served by the traditional flag-carriers, and do so ports as the incumbent flag-carrier. Air Europe at volumes large enough to fill large aircraft. actually had higher domestic operating costs than Exploiting any of the three potential target mar- Alitalia. kets requires solving a difficult tradeoff. The biggest challenge facing "Big O&D" carri- "Big O&D-business" carriers focusing on ers is simply finding enough Big O&D markets more traditional O&Ds (easyJet in Geneva- that can fill large, growing fleets of 140-seat air- Amsterdam, Ryanair in Dublin-regional UK mar- craft. There are certainly large markets out there, kets), have the opportunity to exploit existing but outside of London, they appear to be widely demand, which (as in the Ireland-regional UK dispersed, slow to develop, and not always situ- case) may never be profitable or strategic for ated near low-cost airports. Given this scarcity of incumbent flag-carriers. However, they must stim- cross-border "Big O&D" markets, domestic ulate much greater traffic volumes than the flag- opportunities remain tempting. Ryanair and carriers ever experienced, and overcome the easyJet are planning an attack on domestic strong brand/distribution position of the incum- Germany, while easyJet and Buzz are consider- bents. They need to establish a very stong price ing options within France. and cost advantage to deter retaliation, but must Will "Big O&D" carriers overwhelm the either serve higher cost airports or train the mar- ket to use less familiar alternative airports. Second Division flag-carriers? Pure leisure demand to warm weather desti- Although there are some superficial similari- nations will ensure a market for "Big O&D-tour ties, the European "Big O&D" model is NOT the package leisure" carriers, including the historic classic Southwest model. Southwest pursues charter-carriers. However its share of the holiday- mainstream domestic business markets, with typ- maker market will likely fall as markets mature ical domestic aircraft on frequent schedules, so and the airline component of this business has long as those markets are not at 300-plus flights always been financially challenging. Ryanair has per day hubs where the incumbent hub carrier aggressively begun to expand into the "Big O&D- would have overwhelming advantages. European individual-leisure" segment with low-cost unpack- start-ups must develop totally new markets aged alternatives to southern European airports (Luton, Charleroi, Treviso, etc) with much larger with little or no charter service. Under either the gauge narrowbodies than the flag-carriers use. "Leisure-Individual" or "Leisure-Tour Package" The US has many hundreds of non-hub local approaches, carriers must achieve rock-bottom markets with existing local demand that can sup- costs while absorbing either the packaging and port multiple 737 frequencies. With the exception agency distribution requirements or the challenge of London (and possibly Paris) no European city of developing totally new airports and travel pat- appears able to support more than a handful of terns, plus the normal seasonality and volatility of high-demand non-hub routes. Southwest serves leisure-dominated markets. the large traditional airports business travellers A fourth "Big O&D" target, high-demand favour (San Diego, Detroit, St. Louis), while costs
September 2002 Aviation Strategy
Analysis
at the comparable European airports (Vienna, decided was uncompetitive. Alitalia has come to Zurich, Brussels) preclude Southwest-type low grips with the inability of Italian airports to com- costs. Investors looking for "Big O&D" new pete as Global Hubs, but has yet to realign its entrants to replicate Southwest's financial record fleet or operating costs with a new strategy. - 25 years of double digit growth at industry lead- The only European hub currently following a ing margins - will probably be disappointed. "City Network" type approach is Lufthansa at The "Big O&D" model seems ill suited to cap- Munich, where it operates only four long haul ture much of the Second Division flag-carrier traf- flights, a heavy mix of regional aircraft, and fic, should these carriers continue to lose money achieves a European (cross-border) average 82 and decline competitively. Big O&D carriers have seats per aircraft, in line with the size of the local failed to make major inroads to date in markets market (average seats on domestic German such as Brussels or Munich, and there is less rea- routes are slightly higher). 89% of the traffic son to think they could succeed in cities such as Lufthansa carried at Munich in 1999 was German Milan, Copenhagen or Vienna. While easyJet had home market traffic. While the Munich hub may clearly reduced the profits Swissair and BA had have been originally conceived with son-of- previously enjoyed on the London-Switzerland Frankfurt global ambitions, Lufthansa has sensi- route, both incumbents remained profitable, and bly avoided network shifts that would reduce the easyJet posed no threat to Swissair's core Zurich competitiveness of its major hub, while keeping hub or Swiss home market position. The total col- Munich focused on profits, not glamorous routes lapse of Sabena and Swissair has not led to and big aircraft. major "Big O&D" expansion in Belgium or While all of the second tier carriers face Switzerland. daunting political and industrial obstacles to seri- The continued erosion of Second Division ous restructuring, none have demonstrated the hubs will more likely benefit the four Global Hub type of willingness British Airways and KLM have airlines, not new entrants. They will not only win shown to take major action to address obvious hugely profitable share shifts on their main hub problems of cost and overcapacity. The core, irre- routes (Brussels-Frankfurt or Zurich-Paris) but placeable asset of each Second Division carrier is their regional subsidiaries will be able to expand its historical domination of home market service, into a handful of secondary O&Ds (Brussels- distribution and airport operations. The longer Hamburg or Zurich-Lyon). these carriers wait to bring capacity and infra- Second Division major restructuring structure costs in line with the revenue potential of that core business, the greater the danger that There is little evidence that the Second the value of that core will be irreparably damaged. Division profit collapse that began in 1998 will be Alliances are not a panacea reversed, yet no carrier in the second tier has really abandoned the flag-carrier thinking of the While alliance membership may offer useful last twenty years. Each carrier has made positive benefits to Second Division carriers, there is no moves, but none has coherently unified fleet, evidence as yet that it addresses their strategic capacity growth, market focus, infrastructure cost and financial problems in any meaningful way. and productivity improvements into a credible Under certain conditions, alliances can strength- strategy for making money in a restructured en an already secure network base, but they do European industry. not work in all markets, and they cannot turn a Austrian long ago abandoned any global pre- weak, marginal network into a profitable one. tensions but still operates at a very high average The only alliances that have been big wins for gauge for the Vienna market (101 in 2000) and both sets of shareholders are the immunised has been unable to shift from a 150-seat mainline North Atlantic pairings, where two airlines with jet to a Tyrolean regional jet based focus. SAS strong, sustainable "home continent" networks has reformed its capital structure and significant- linked their hubs to capture competitive traffic ly strengthened its Scandinavian home market flows they could not otherwise serve. There are position but has also spent heavily on increasing no meaningful intra-European flows that two hubs its already excessive narrowbody gauge and would not be able to serve unless they joined in restoring Intercon capacity that it had previously an alliance with codesharing. Incremental traffic
September 2002 Aviation Strategy
Analysis
acceptable. EUROPEAN BUSINESS MODELS 2005? Lessons from the SAir debacle
At one level the destruction of Swissair and Sabena would seem to be Charter an aberration, and one certainly expects that this level of financial mismanage- • Narrowly • Network ment and willful disregard for the basic targetted Big Global Ubiquity from economics of European airline competi- on one O&D Hubs Home Market tion will never be seen again. But as one O&D or • Multiple demand Demand considers the magnitude of the industry segment Segments restructuring still needed, several lessons may be worth considering: Big O&D • If a European airline has a serious com- Business petitive/financial problem and fails to do City anything about it, it can fail. Network • The fastest way to cripple an airline is a major fleet investment inappropriate for the airline's markets. It is extremely diffi- captured by the Swissair-Sabena alliance was cult to restructure obligations of this magnitude. negligible. Alliances might serve to discourage Yet these investments tend to receive very little competition for traffic between the two home mar- outside scrutiny, and many assume that "fleet kets, but the competition authorities might not renewal" is always a profitable thing to do. accept this as a major benefit. • Most of the real value of any airline is intangible- Full intra-European alliances (including joint brand equity, home market distribution strength, sales and frequent flyer programs) can shift long managements' ability to manage complex net- haul traffic but this tends to be in one direction works, etc. At the point an airline shuts down that only, with the Global Hub operator gaining share value is destroyed forever, and it may be impos- from the junior partner's home market. It is com- sible (or incredibly expensive) for a new company pletely sensible for SAS and Alitalia to abandon to recreate that value. Thus it is in every stake- any Intercon capacity not largely serving local holders' interest to protect the "going concern Scandinavian or Italian demand. But it is extreme- value" through any reforms or financial restructur- ly difficult to design a pooling mechanism that will ing that might take place. ensure that the benefits of diverting local revenue • Private ownership of airlines has many advan- to Frankfurt or Paris will be shared equally by tages, but is not sufficient by itself to ensure an both the junior and senior partners. Air France efficient industry. A private majority owner of can connect every large Italian airport to CDG Sabena destroyed all of the value held by Belgian without Alitalia's help. While it makes sense for Air taxpayers, creditors, suppliers and employees, France to compensate Alitalia for incentivising its but those stakeholders were powerless to protect customers to fly via Paris instead of Frankfurt or their interests. Amsterdam, it is unclear how such payments • If badly run airlines go bust, the service will not could be large enough to cover the major restruc- necessarily be replaced by other better run carri- turing costs Alitalia faces and drive a major finan- ers. Aeropolitical constraints would have blocked cial turnaround. any foreign company from attempting to take over Cost and management synergies could be the failing Swissair or Sabena positions. Slot con- significant but require common ownership and trol mechanisms make it impossible to establish a control (as between Swissair and Sabena) and hub network position without acquiring all of the cannot be seriously exploited under an arms- liabilities of the failing company. There is consid- length alliance. The consolidation of systems and erable risk that the local market will simply functions that would drive meaningful savings receive much less service than it would with an requires loss of direct control and other risks that efficient market that did not have these artificial independent management groups rarely find constraints.
September 2002 Aviation Strategy
Analysis
• Letting badly run airlines go bust can serve the effective US style Chapter 11-type process. public interest in the US because the bankruptcy • One can achieve meaningful savings by consol- laws there ensures that airlines with viable core idating management and key systems, but only networks are not prematurely destroyed, and pro- with common ownership and control. These deci- vides an imperfect but largely workable mecha- sions could be taken at Swissair and Sabena nism for reallocating assets to more productive once there was truly one common bottom line. uses while protecting creditor rights and encour- • Given local marketing, industrial and airpolitical aging new investment. European bankruptcy constraints, fully merging major cross-border air- laws are highly similar to US law on paper, but line brands would probably be counterproductive, appear totally ineffective in the case of large air- however, big savings may still be possible even lines. The restructuring needed at many Second when two airlines continue to have separate crew Division carriers may be impossible without an rosters and brands.
CUSTOMISED COMPANY AND MARKET BRIEFINGS If you are interested in a briefing on a particular airline, manufacturer, lessor or industry sector/market, Aviation Economics is able to produce in-depth reports customised to your requirements.
Contact: Keith McMullan or Tim Coombs +44 (0)20 7490 5215 [email protected]