The economic market power of Amsterdam Airport Schiphol: A Rejoinder

Airneth Column
September 2010
By Prof. Dr. Andreas Polk
on behalf of the GAP study team

David Starkie, in his comment published at Airneth, welcomes the study on the assessment of market power of Amsterdam airport Schiphol and points at some methodological aspects and economic judgements, which - according to his point of view - beg some important questions. We strongly welcome the public interest the report arises and appreciate the discussion and proposals of the comment. This rejoinder reflects on some important ideas of the comment and gives answers to some issues. 
 
As David notes, the report was commissioned by the Netherlands Competition Authority, NMa. It is part of the evaluation of the Dutch Aviation Act, which aims at analysing if current market conditions still justify sector specific regulation of airport charges and conditions. The evaluation process contains an analysis of the extent of market power of Amsterdam airport Schiphol, which is the study to which David refers. The study team, of which I am a member and on behalf of which I reply, has been asked to take among others two important aspects into account: First, the focus is on aviation and aviation-related services, which generally speaking contain all the services an airport provides to airlines necessary for take-off and landing of aircraft¹. Beyond the scope of the study are the non-aviation services Schiphol airport provides, i.e. the airport's real estate or consumer business². 

Second, the study needs to be in line with European competition law, and therefore also contains a chapter on the definition of the relevant markets. We agree with David that competition law is sometimes a difficult framework for economists to exercise their skills, as we love trade-offs so much (surely not a synonym for not making up ones mind). But we believe that economists have to be aware of the legal institutions they argue in, and they need to be able to communicate their expertise within this framework to competition experts, lawers and policy makers alike. If the institutional setup is ignored, a study may be a nice intellectual exercise, but it is likely to remain unheard in the political decision making process. Consequently, we appreciate that the NMa requested the study to be within the framework of European competition law, and not beyond it. 

Having said this, let's turn to David‘s most important arguments and have a closer look at some details.
 
Dominance
The comment starts with the relationship between the existence of dominance and its potential abuse. It argues that the existence of dominance per se does not matter. Rather, the central question is if the airport has incentives to abusively exercise a dominant position. We can understand this as a reference to the commonly used two step approach in European competition law, which first investigates if a firm is dominant and, if affirmative, analyses if the firm abuses its dominant position. Our study refers to this approach³. 

Unfortunately, the comment leaves open what follows from this argument⁴. To our understanding the author in fact argues that the non-avation activities might restrict the airport's incentives to increase charges for aviation-activitities, because aviation-activities exert a positive externality on the non-aviation services. The airport will internalize these effects by reducing charges. More simply stated: The more traffic the airport attracts, the more customers shop at the airport, which in turn increases revenue in the non-aviation shopping and parking business. Thus, according to this argument, a dominant airport will voluntarily not exercise market power. It has incentives to lower its airfares to a competitive level⁵ in order to internalize the externality and attract customers for the non-aviation business areas. So why discuss market power at all instead of releasing all airports from regulation? 

The argument may indeed play a role, and the author correctly indicates at this possibility. But in terms of regulation, the incentives to internalize must be very strong and effectively restrain the market power of the airport. Furthermore, it would have to restrain the airport from exercising its market power in all relevant areas of aviation activities (i.e. O&D passengers, transfer passengers and cargo). From our point of view this is not the case, and during our investigations we didn't come over any indication that this aspect is strong enough to render regulation superficial. However, we agree that this aspect is indeed an interesting question and should be subject of further research. An appropriate study would have to cover non-aviation and aviation activities alike. (The aspect may also affect the choice of the best regulatory regime, i.e. single-till vs. dual-till.) However, for good reasons the non-aviation services were explicitely not subject of this study to which the comment refers⁶. But even if we took this aspect into account, the comment confines itself to simply indicating at what could be a problem, but does refrain from giving any evidence or at least indication that the argument might indeed be important in practice and beyond mere theoretical ideas.

Moreover: Yes, the Australian approach towards airport regulation is different , as the comment correctly notes, and we think it's an interesting experiment. But what does this proove? The Australian system of income taxation and the Australian social security system also differ more or less from the European as well as from the American approach. Are they also superior, just because they’re Australian? The Australian approach of (non) regulation is indeed an interesting experiment and we surely will see future research on how it performs compared to the regulatory approaches used in so many other countries around the world. But without further research it appears too speculative from our perspective to solely rely on this argument, as long as there is no evidence that the internalization effect is strong enough to outweight all other indicators pointing towards the existence of market power of Amsterdam airport Schipol. 

Market Definition 
With reference to the proposed market definition, the comment appreciates the definition of separate sub-markets, but states that the separation of markets does not go far enough. Especially David remarks that, despite our market separation, the study would treat all passenger-serving airlines as belonging to the same market. So to our understanding, the comment argues that some airlines like Ryanair have a different business model in terms of choosing airports, and this should be taken into account. 

Many comments are appropriate on this aspect, and we will only refer to which appear to be the most important ones here. First, market definition is not a means in itself, it rather serves a certain purpose. So what are the gains from a further subdivision? The comment does not indicate this, nor does it give any proposal about which further separation could bring new insights. We doubt that a further subdivision brings more to gain. 

Second, market definition must rely on observable evidence, it is not sufficient to simply assume a market. The analysis in the market definition in Chapter 4 gives many indications why the chosen market definition appears appropriate from our point of view. Considering the comment's arguments, there is for instance no indication that the airport adresses different types of carriers differently. Rather, the price structure of Schiphol airport is the same for all airlines and independent of the particular business models. The airport does not price discriminate between different airlines, if KLM, Ryanair or alike. The comment indicates at what might play a role at an airport somewhere else, but in the context of Amsterdam airport Schiphol bilateral bargaining is not the point. 

Third, to our understanding it is not clear which definition the author prefers, not even if he prefers a broader or narrower definition. First, he argues in favor of a further separation of the relevant markets (which is a narrower market definition, even though he refuses to indicate which definition then appears appropriate), just to state a few sentences later that he assumes that there is now "a single European market" in aviation, which in fact favors just the opposite, namely a broader market definition⁷.

And finally the comment mixes up aspects of demand-side substitution with those of supply-side substitution. The reference to para. 21 could be misleading in this context. Market definition relies heavily on the analysis of demand-side substitution, which becomes clear even if a reader confines himself to the executive summary. The reference which the comment gives as proof in this context, namely that it plays a role if an alternative airport is able to "take over almost all traffic", refers to supply-side substitution, which is clearly rejected to play a role at Amsterdam airport Schiphol⁸. 

Counter arguments 
Finally, let's turn to of what David calls "Counter arguments". For the sake of length we restrict ourselves to one of them, even though the comment as being published on the Internet, well claimed for more attention on certain aspects. 

The comment correctly indicates that the aspect of sunk costs may well work in both directions, but immediatly finds relief in what David sees as one of the healing inventions of modern economic systems (it's bilateral bargaining, not to put you on the rack). The proposed solution is simple: If the airport negotiates contracts with the airlines, the problem of market power vanishes into the air, just because there are negotiations! Those guys of you with kids may know the experience that bilateral bargaining may well be one-sided (The others, guess how!). So is the pure existence of bargaining really sufficient to render questions of market power superfluous? We think not. 

This aside, there is another problem with the argument: The analysis of market power cannot be detached from what we observe at the specific airport under investigation. Even though Schiphol is allowed to bargain bilaterally on regulated airport charges and conditions, it has to offer the terms of the agreement to all airlines in case the negotiations were successfull. Thus bilateral bargaining over the regulated domain is less attractive for the airport and it does not play a major role for the determination of aviation charges.

Conclusions 
We highly appreciate David's comments, and it was a pleasure to discuss some of the issues in greater detail at a recent conference. We appreciate the open discussion, and this rejoinder also intends to stimulate well-considered comments. The comment hints at some interesting questions for future research: Among these are the issue of costs and benefits of regulation, the appropriate approach towards it, or the lessons we can learn from the Australian experience (did I say experiment?) of freeing up airports.

-------------------------------------------------------------

1. This includes the provision of infrastructure for take-off and landing of aircraft, aircraft parking, processing of passengers and freight, security checks, as well as granting concessions for refuelling or catering.

2. The general competition law framework is considered to be sufficient to ensure competition in these business fields by the Netherlands Competition Authority.

3. Take the discussion about the access to the airport for the provision of ground handling services as an example, which the comment cites as an example in Footnote 2.

4. The cited references in the comment do not help us to shed light on it. The relevant section of the study discusses the following aspect: Even though the airport does not exercise (sic!) market power in the market for the access to the airport for the provision of ground handling services, the airport could possibly abusively employ its position in the market for the provision of rental space at the airport to increase profits (cp. executive summary, para. 35-45, and Chapter 6.5 of the main report). The study leaves this question open to a certain extent, as this aspect is not regarded as being crucial for the overall assessment of market power. But it is misleading to take this reference as an indication for the general approach of the study and the outcome for all identified markets, as this does not do justice to the analysis conducted in the whole report.

Then, the author constructs a contradiction to this statement by turning to an early part of the main text (which aims at identifiying all stakeholders and services the airport provided as a prerequesite for the definition of the relevant markets and the analysis of market power), which - at that stage of the inquiry - simply refers to the main question of the study rather than answering it. This becomes clear from a more detailed inspection of the main text of the study, but – at least from our point of view - the structure of the executive summary also makes this point clear.

5. Whatever "competitive" might mean in this context, as almost any company has market power to some extent. The relevant question is how much market power is necessary to justify regulation.

6. One aspect which needs to be taken into account is the extent of market power in the non-aviation sector, which is likely to exist to a certain extent, and how this relates to the incentives to lower airfares.

7. There is not even a single common market in the downstream transportation markets.

8. Compare para. 211-223 and Appendix D of the study.