In the early part of 1994, US transportation officials began to question the precepts of the 'glue' that has come to seal alliances between airlines of different countries: codesharing. In typical Washington style, a study was commissioned. And then another.
Do codesharing alliances produce benefits? Could they negatively impact competition between airlines? Do consumers gain? The questions went on and on but plainly said something simple: the US transportation department had little hard evidence to show what effect codesharing had on US carriers, despite its willingness since 1987 to comply with the wishes of airlines seeking to partner via an electronic on-line booking connection and having made codesharing a central tenet of its new international aviation policy.
Crystal clear
But if government officials were uncertain about the practice's consequences, airlines were crystal clear. While the US was studying the concept of codesharing last year, Northwest was busy making $150 million from its alliance with KLM. United Airlines and Lufthansa were similarly preoccupied, developing a new alliance that will soon add 1,000 passengers a day to their combined system. And American Airlines, the lonely but die-hard opponent of broad codesharing agreements, was losing an estimated tens of millions of dollars from traffic diverted from its system.
These findings come from the US General Accounting Office, the investigating arm of Congress, which in May released the results of a year-long inquest into codesharing alliances. The watershed study, along with a Gellman Research Associates (GRA) analysis commissioned by the DOT and released last December, found things that airlines seem to be well aware of:
1 Codesharing alliances are indeed beneficial in terms of increasing traffic and, hence, revenues;
2 They tend to make their money at the expense of airlines that do not have similar types of alliances;
3 Last but not least, consumers generally benefit.
The two studies reached many other conclusions as well. For US transportation officials, however, the most important is that before revering codesharing as a way to break down diplomatic and nationalistic barriers to open aviation trade, they should know just how powerful this marketing concept actually is.
It was this lack of knowledge about the effect on the consumer that GRA wanted to remedy. Using a 'discrete choice' econometric model, which attempted to discern how a consumer chooses between competing airlines or flights, the methodology was designed 'to examine the effects of codesharing prospectively, so that the DOT could use it to assess the effects of future proposals.' The model was intended to resolve such issues as which airline - US or non-US - benefits more from codesharing, and if 'consumer welfare' was favoured.
In applying the model, GRA studied the effectiveness of Northwest-KLM and BA-USAir as on-line, codesharing alliances. By extrapolating on first quarter 1994 information for the two partnerships, GRA's results 'indicate that codesharing has a significant impact on market share.' In the markets that were measured, they show that BA-USAir has the potential to achieve a 19.2 per cent market share, but that its share in early 1994 was really 11.2 per cent. Without codesharing, a BA-USAir interline agreement would win only a 2.9 per cent share of the market. The effectiveness of the Northwest-KLM codeshare partnership was nearly complete back then and the two fell short of their estimated potential by 1.4 per cent, giving them a 45 per cent market share in the measured areas. Without codesharing, GRA believes Northwest-KLM would control a maximum 34.4 per cent of their markets.
In terms of revenue, GRA's estimate is that BA gets roughly $90 million annually from its alliance with USAir, while USAir receives close to $16 million. Northwest, gets $49.2 million per year, while KLM makes $37.2 million. In breaking down the results in dollar terms, GRA concludes that 'other US carriers' provide the brunt of the revenues for BA-USAir - $80 million - via lost traffic. Partly because of this figure, which disproportionately rewards a non-US carrier, GRA deems BA-USAir not to be in the best interests of the US airline industry. In contrast Northwest-KLM negatively impacts US carriers by only $1 million, which is cancelled out by Northwest's larger revenue gain over KLM.
All the figures go into a stew that theoretically should devine the 'net social surplus' of any given codesharing agreement. This would be fine if GRA's stated goal of creating a model for future use by DOT had been largely met. But, many agree, it has not.
The model has been compromised by a host of 'real world' dilemmas: it is only a snapshot of 46 markets that belonged to two alliances during one quarter early on in the development of the alliances. Additionally, it makes no account for non-US airline competition due to lack of data, market growth has not been quantified, and US airline competitive responses have not been factored in. All of these issues are addressed in the GRA report, to the point of disqualifying the findings before they've even been presented.
Paucity of data
Though DOT officials like James Craun, director of the newly created Office of Aviation and International Economics (the official 'sponsor' of the study) say discrepancies are necessary to act as controls for measuring other variables, he admits the study is only 'a starting point' that reflects the paucity of data airlines gave GRA. For relevance, he adds, the study will have to be repeated. Patrick Murphy, acting assistant director for international policy at DOT concurs: 'It is not a perfect product, and certainly not the definitive work. [However] the study had the benefit of quantifying BA-USAir and Northwest-KLM.'
It also had the benefit of corroborating DOT's international aviation policy statement, which viewed positively the codesharing trend. On the other hand, the General Accounting Office report, released in early May, has not met with the DOT's favour. This is partly because the study, done for representatives sitting on the Senate commerce committee, was not controlled by transport officials. Also, the findings give credence to codesharing's negative impact on those without alliances as well as a detailed review of the very real benefits of anti-trust immunity, which most alliances wish for but only Northwest-KLM now has.
Magic key
Compared to GRA, the GAO seemingly received a magic key to airline industry codesharing data. GAO's mandate has always been to gather information, so its historical research analysis is considerably different from the more academic GRA study.
Interviewing executives and researching operating data at seven US airlines and 13 non-US players, GAO's report is probably the most in-depth analysis of codesharing that has yet been performed, even though in some cases there is a reliance only on the company word. 'We were able to collect information on 85 per cent of the 61 codesharing alliances approved by DOT since 1987,' the study's primary author, Timothy Hannegan, wrote in the report's introduction.
It is a broader perspective and puts some badly needed hard numbers on the codesharing controversy. Does codesharing work? Look at British Midland, the relatively small UK carrier that has in the past few years implemented eight codesharing alliances with the likes of United and American. In 1994, those eight partners produced 100,000 extra passengers for British Midland, mostly taken from competitor British Airways. Since beginning to codeshare with United in April 1992, the UK carrier averaged 2,072 extra passengers per month. In contrast, the interline agreement the two companies had before developing closer ties averaged just 151 passengers per month.
Other smaller agreements are also analysed: American-South African Airways produced $2 million last year for each carrier; United-Ansett gave UAL $14 million in extra revenue. But like GRA, the GAO report concentrates on the large, strategic alliances that produce dramatic results for the carriers involved and equally dramatic traffic shifts from competing carriers. Specifically, the study looks in detail at the alliances of Northwest-KLM, BA-USAir and, to a lesser extents, United-Lufthansa.
Officials involved in the Northwest-KLM alliance, which brings together 88 US cities with 30 European and Middle East destinations, believe that their combined transatlantic market share has increased from 7 per cent before the alliance began in 1990 to 11.5 by 1994 (see chart). In 1994, revenues from the alliance were estimated at 5 per cent of Northwest's total international revenue base, or $150 million, and approximately $100 million for KLM, or three per cent of its total international base - these figures differ markedly from GRA estimates.
Market share
Between 1991 and June 1994, the market share of the alliance, between 34 of the US codeshare destinations and 30 European and Middle Eastern cities, almost trebled from 1.2 per cent to 3.3 per cent. The increases for Northwest-KLM are attributed largely to the level of integration of the alliance and the anti-trust immunity the two airlines possess, which protects the two carriers from legal challenges and allows integrated pricing and marketing initiatives.
The equivalent opportunity in this partnership is less visible between BA and USAir - the most controversial in Washington. Like the GRA study, GAO has found that BA is profiting significantly over USAir in the relationship, roughly by five-to-one. Compared to USAir's estimated $20 million in additional revenue from cooperating with BA last year, BA told GAO it brought in $100 million between April 1994 and March 1995, with codeshare services directly contributing $45 million of that.
But unlike the GRA study, which talks of BA's unfair dominance in the relationship, GAO's report points towards BA's total $400 million investment as an equalising force between the two carriers. Nonetheless, the structure of the alliance is weighted towards benefiting BA's bottomline over USAir's: the codesharing only goes one way in this relationship, with USAir feeding BA transatlantic flights.
Non-codeshare revenue
This of course has a direct impact on codesharing traffic, but BA also benefits from a halo effect on interline traffic from US cities other than the 52 US codeshare points: 'A comparison of BA's traffic data for April through December 1994 with the same time period a year earlier shows that the number of USAir-BA interline passengers has increased by 60 per cent, from 36,396 to 58,164.' The interline traffic figure is part of the non-codeshare revenue - which also includes cost savings and frequent flyer programme hookups - that account for $55 million of the annual alliance benefit.
Though DOT officials hoped that the GAO would find that the revenue and traffic gains from these alliances come solely from market stimulation, the report's researchers say they could not find much evidence to support this idea. Money and traffic, it appears, comes primarily from other airlines coffers and distribution systems.
The GAO heard plenty of complaints: Continental Airlines says it lost $1 million in revenue last year in the transatlantic market to Northwest-KLM.
Delta told researchers it lost 'about $25 million' in 1994 because of BA-USAir. Another unnamed carrier says it lost $40 million, or 11 per cent, of its US-UK business to the same alliance. Additionally, a 625 per cent increase in codeshare traffic - from 6,589 to 47,749 - between the last six months of 1993 and the same period last year for BA-USAir is suggested to have come primarily from interline traffic between BA and other US carriers. In this case, among other examples United's US domestic traffic interlining with international BA flights dropped 15 per cent, Northwest's came down by 9 per cent and TWA lost 6 per cent.
All of which provides troubling evidence for the DOT. In its international policy statement last November, it hailed codesharing as the perfect market system that would help assure US airline rights around the world. Equally important, it has pushed aside American's long-standing complaints that there are liabilities for the US airline industry in codesharing. How the department rectifies these positions may soon be on the docket, after congressional hearings over the international policy. Perhaps most important, however, is the recommendation from both reports that the DOT strengthens its analytical departments. Simply having to contract out a report on such a significant industry trend as codesharing speaks volumes about the weakened state of DOT's research end, and its consequent dependence upon US airlines to provide analytical information to back up negotiating positions.
Knee jerk reaction
In Washington the knee-jerk reaction is to blame the UK for all the US's aviation ills, but it is difficult to sympathise completely with this notion when, for lack of appropriate data collection and analysis, the US has stumbled in most Bermuda 2 negotiations. Sources say that the UK was studying the importance of codesharing as early as March 1989, and that the consequent 1991 US-UK negotiation, which gave United and American access to Heathrow and BA the right to codeshare to the US, is a testament to that early research. 'Our concern is that the DOT is not well positioned to measure and monitor the impact of competition and fares, and the well being of the US airline industry,' says Hannegan.
In Washington the knee-jerk reaction is to blame the UK for all the US's aviation ills, but it is difficult to sympathise completely with this notion when, for lack of appropriate data collection and analysis, the US has stumbled in most Bermuda 2 negotiations. Sources say that the UK was studying the importance of codesharing as early as March 1989, and that the consequent 1991 US-UK negotiation, which gave United and American access to Heathrow and BA the right to codeshare to the US, is a testament to that early research. 'Our concern is that the DOT is not well positioned to measure and monitor the impact of competition and fares, and the well being of the US airline industry,' says Hannegan.
Source: Airline Business
