Studying the empirical implications of the liberalization of airport markets

Assessing quantifications of regulatory effects

I define economic regulation as involving significant interventions in the play of market forces, which may include the support of these forces. These mainly involve hard regulations embracing such as command-and-control instruments, taxes and subsidies, price controls, market access requirements, standards, and ownership. (Kahn, 1970, 1971; Train, 1991). But it has also become increasingly recognized that soft regulation, such as the use of nudging, shoving, and boosting as developed by Thaler (2016) can be important in many industries, including airports. ² The emphasis is primarily, however, on the more traditional forms of regulation ³ It is also accepted that institutional structures are part of any well-functioning market; their absence would lead to chaos.

Airports are not easy to define and quantitatively analysis. They are multi-faceted providers. They obviously cater for the movement of aircraft, and the aviation needs of their passengers and freight transporters. But their functions vary according to their size, their roles in aviation networks, and the types of traffic that they handle (Adikariwattage et al., 2012; Huber & Rinner, 2020). There are considerable differences, for example, between the 77,600 acres of King Fahd International Airport and a local community airport in the US or Australia. Airports can serve different functions in a network. For example, Hartsfield–Jackson Atlanta International Airport in the US handles 110 million passengers a year in normal circumstances but a large percentage of these are domestic travelers, whereas virtually all of Dubai International Airport’s 90 million users are international. Some large airports like Memphis are dominated by the cargo traffic; it is the main hub for FedEx and handled 4,470,196 tons of cargo in 2018.

Airports commonly offer a range of non-aviation services and, directly or indirectly, control others. Many provide extensive car parking and retail outlets used by non-airport travelers or employees, warehousing for cargo, and associated business activities. Airports act as purely ground transportation interchange points. Some provide or control surface access infrastructure. Defining the exact boundaries for study is, therefore, challenging and inevitably raises economic issues related to such things as the agglomeration economies generated and the two-sided markets in which they operate (Czerny & Zhang, 2015; Starkie, 2001; Tirole, 2015). Added to this, surface access to airports can vary considerably and determine their catchment areas as well as the degree to which they are competitive with each other (Akar, 2013). Having said all this, it is perhaps surprising that given airports form part of an extensive and complex supply chain, that most analyses treats them as individual businesses and give limited consideration to network interactions.

Even when definitions are agreed, it is not easy to assess the effects of any specific regulation. This is not just because regulations inevitably come in bundles that make the isolation of any one intervention problematic. At the core of the challenge is the impossibility of specifying the counterfactual if no regulations had been imposed. Assessments of the implications of regulation inevitably involve some implicit or explicit ideas of what the outcome would be if the status quo were to be maintained; the “do-nothing scenario.” Ex ante assessment of an intervention entails predicting how this will cause deviations from the do-nothing scenario. It essentially involves some form of difference-in-differences analysis. But, of course, forecasting the do-nothing scenario is itself the subject of considerable uncertainly. The models used for forecasting are largely based on, albeit sophisticated, historic extrapolations, that themselves may well morph in the future and be impacted on by unpredictable external shocks. Ex post analysis, the subject here, entails similar challenges in backcasting what would have happened without changes in regulations.

There are also the difficulties of evaluating the quality of empirical analysis. Milton Friedman (1953) provides one criterion: A model should predict well, or at least meet the needs of those who seek to use it for predictive or forecasting purposes. But here we are dealing with ex post studies. In the natural sciences, replication is an important criterion for evaluating these. Studies should be consistent in what they find. There have been arguments dating back at least to the 1980s that more replication is needed in the social sciences (Kane, 1984). A number of factors, however, mitigate against conducting the necessary work. Independent, direct replications of others’ findings can be time-consuming and take resources away from other projects that are considered to involve more original thinking. Replications are harder to publish, often because they are not seen as very original.

Concern about this lack of replication has grown in economics. ⁴ Chang and Li (2017), for instance, could only replicate 22 of 67 papers using the authors’ data and code files, and an additional seven papers with assistance from the authors. Of articles using public data and had code written for software owned by the authors, 29 of 59 could be replicated. In particular fields, the issue has been raised by Camerer et al. (2016) in experimental economics, Hamermesh (2017) in labor economics and Sukhtankar (2017) in development economics. Little, however, has been done either on network industries in general or, indeed, on regulation more widely. The number of studies involving airports is rather scant.

One possible way of comparing empirical work is to use meta-data analysis. This entails statistical synthesizing of the results of prior quantitative studies adjusting for their context. While potentially useful, care is necessary when applying it in general, with specific problems when transportation networks are concerned (Button, 2019). Meta-analysis is also not “a scientific technique.” Subjective decisions are required regarding which studies to include, what weights to attach to the studies, and how to interpret the results. Including qualitative considerations is difficult. Simply, meta-data analysis may add pseudo-scientific exactitude to assessments. ⁵

In all forms of survey there is the problem that there is no systematic way of objectively assessing the quality of studies. There are also omissions. In particular, gray literature (e.g. reports and working papers) and those not written in English tend to be omitted here. There is a tendency for editors of peer reviewed journals to favor results that support existing theories or contain findings broadly conforming to “expected” magnitudes and to reject as outliers papers where findings radically differ from the norm. In the social sciences, the lack of any template for reporting results offers an added challenge, as does the focus on quantitative findings, with little attention on the qualitative effects which can often be large.

Why do airports get regulated and “deregulated”?

So why is it there often seen a need for economic regulation of airports? First, airports are important and often seen as public utilities. For example, in normal times aviation is responsible for the movement of 35% or so of world trade by value, and some industries, such as tourism, rely heavily on them. An investment in a new airport or changes to an existing one, involves expenditures of vast sums. Design and operations of airports are complex. In addition to the attention paid to commercial airports by poliocy makers, they attract considerable attention in the academic literature, albeit mainly in relation to their engineering peculiarities. That is not surprising. They are important hubs in local, national, and international multimodal transportation networks. They facilitate interactions between surface- and air-transport movements for people and cargo.

In theory, perfect markets can handle all of these things. However, market contracts are seldom complete and, given the existing allocation of property rights, airports can generate extensive negative and positive external effects. These can be both technical and pecuniary. Airports in particular are focal points for noisy and polluting activities, as well as enjoying some degree of monopoly and rent seeking power which is the primary interest here (Basso, 2008; Forsyth, 2007a). There are transactions costs in markets, and a planned and managed approach is often seen as the most efficient. These, and other factors, have led to an array of economic regulations being imposed on airports. The traditional view is that appropriate regulations are often needed to serve the public interest (Kahn, 1970, 1971). There is little dwelling on this well-trodden path in neo-classical economics.

A different perspective on regulation stems from the fact that the presence of an airport is closely entwined with the development of the surrounding region (Morrell, 2010). ⁶ Airports can be key elements for generating significant local spatial agglomeration effects (Venables, 2006). Thus, regulation to meet the public interest extends beyond trying to optimize the operation of an airport per se, to embrace its role within the larger geographical economy. Airport regulation, therefore, extends beyond the narrow confines of nodal optimization within an aviation network, to larger matters of economic development. In particular, it may aim at reducing the economic risks to private investors of putting resources into a location. Airport support in these circumstances helps guarantee air transportation services will be maintained. This is the transactions costs theory of economic regulation. It is aimed at reassuring up- and down-stream investors that their sunk investments are not burdened by excessive risk (Biggar, 2009, 2012; Moszoro & Spiller, 2016; Spiller, 2013).

This notion of protecting an airport’s viability, and the activities associated with it, can also be seen as sustaining positive pecuniary effects. The attractiveness of an airports has trade diversion and trade creation effects. The former results from a new or expanded airport transferring economic activity from other areas suffering from congestion, to new site. While this may add to the efficiency of the system, basically Venables’ spatial agglomeration effects, it will also redistribute existing economic welfare. If the latter results in a socially preferred income distribution this is a positive pecuniary externality. Economic welfare increases because of greater technical efficiency, but also from a preferred distribution of that welfare.

In all these cases it is assumed that it is the general public that benefits from regulation, albeit without an even spread of gains and costs. The concern that has grown since the 1970s is that regulations can be introduced or subsequently manipulated to benefit those responsible for regulating an industry or by those who are meant to be the regulated (Peltzman, 1976; Stigler, 1971). While often discussed in the context of airports, the concept of regulatory capture has seldom been explicitly empirically tested. This is difficult to do, again because of the counterfactual problem and the diversity of stakeholders involved.

What has been done is to conduct studies contrasting various situations where there are regulations and where there are not. This is not quite the same thing. There have also been studies of airline regulation which, because a dominant carrier may have monopsony power at an airport, has a bearing on the behavior of the airport; the bilateral monopoly situation. Institutionally, airlines can gain such power in a number of ways. For example, until recently international air services agreements stipulated which airlines could serve individual routes, and the origins and destinations of services. Further, the lack of permitted cabotage prevents foreign carriers from providing domestic services, limiting national networks and distorting the use of airports. The perimeter rule applied to Washington Dulles Airport in the US, for example, limits the flight distances of air services.

The power of the market

It is often forgotten that one form of regulation is to “Leave it to the market.” But this has been influential in the way airport regulation has been viewed in recent years. In practical terms, it is a wider adoption of the Anglo-Saxon approach to regulation. The idea is to allow even imperfect markets to operate unless intervention offers demonstrable benefits. The problem, of course, is defining exactly what the unimpeded market situation would look like. At the very least some form of property right protection is needed, but this can take various forms. Intervention then becomes a matter of degree.

Whatever the historic justification for the economic regulation of airports, the 1980s saw a move away from extensive regulation. This partial disengagement by the authorities from airport ownership and control was part of broader trends towards liberalization seen in many countries. The positive demonstration effects of liberalizing other markets, and notably airlines, was a factor in bring about reform, as were concerns over the efficiency of airport management (Foster, 1994). But certainly, there was a feeling that circumstances had changed and market forces had a greater role to play. Although inevitably the term was highlighted by politicians and those affected, it was not strictly deregulation but a shift in the ways production is coordinated. As Coase (1937) observed firms and markets “…are alternative methods of coordination production.” In the case of “firms,” he meant planned and managed undertakings, as opposed to individuals trading with each other Regulatory reforms have shifted the emphasis of the paradigm towards less central planning and management

Much of the early interest in regulating infrastructure related to the monopoly power it can confer on owners. In the transportation context, for example, railroads were heavily regulated virtually everywhere for fear that users would be exploited. This fear extended to airports. They are, after all, large, fixed, durable, indivisible, and specialized assets. Reforming the resultant regulatory structures imposed on airports, however, has taken time to take effect and, as with most infrastructure, there are considerable transition costs (Meyer & Tye, 1981) Airports are certainly not like the airlines where there can be free market entry leading to contestable conditions that temper any existing market power that they may enjoy (Baumol et al., 1982; Button, 2010). While airlines can acquire aircraft quite quckly, it takes years to plan and develop a new airport. The finances are complex and planning concent is difficult to acquire. As a result, there has been a tradition of public ownership, either at the national or local level. The argument being that this serves the public interest.

Airport competition, however, can be more complex than is often assumed (Starkie, 2002). In many cases airlines have options of which airports to operate from, and for regions of Europe and the US the options may be quite large. This can enable airline to exercise considerable countervailing power (Button, 2010). Low-cost carriers, in particular, are mobile and move services between airports (Francis et al., 2003). Indeed, Starkie (2012) argues that low cost carriers by operating from multiple bases have the ability to threaten to shift their services between airports and this keeps their terminal fees down. The widespread adoption of hub-and-spoke networks by the larger carriers often offers long-distance passengers a range of hubs to fly through. There are also alternative network modes available. High-speed railways in China, Spain, Japan, and France, for example, offer options over medium distance routes, although the complex systems of subsidies involved violate any text-book definition of competition (Albalate et al., 2014; Barbot, 2006; Sun et al., 2017).

It has already noted that problems arise when the regularly structure is manipulated to serve other than the broad public interest. But there can be another problem which is not to be confused with government intervention failure. These problems occur when interventions in markets unintentionally reduce public welfare. This may happen, for example, when the original regulations did not meet their aims but were retained or reinforced hoping this would change the situation, or when the original interventions are based upon inaccurate information or inappropriate, albeit well intended, concepts (Winston, 2006). In a way they arise because the public sector has less information than the private. The tendency to continue to support loss making and little used airports can be a result of this.

Regulating the ownership of airports

Until the late 1980s virtually all of the world’s largest airports were owned by government, be it at a national or more local level. In places like the US this is still substantially the case. Historically, there had been a big expansion in the number of military airfields during the World War II and these subsequently formed the basis for the post-war development of commercial aviation. Thus, the public sector that originally constructed them directly controlled their development and operations. They were basically treated as a public utilities with their managers implicitly focusing on what were often ill-defined definitions of public interest, and in ensuring the stability of the airline networks they served.

The situation changed in the late 1980s when supply-side macroeconomics emerged with its emphasis on economic efficiency. The new circumstances favored lighter regulation and different ownership models. And with this came a movement back from government ownership. The changes have taken a variety of forms. Hong and Yoo (2001), for example, delineate them as (i) change of ownership from public to private sector; (ii) change of ownership from central to local government; (iii) change of legal status from autonomous governmental authority to public corporate; (iv) expansion of the private sector entities’ participation in airport operation without ownership; and (v) private sector financing.

However, the detailed motivations for change and the nature of changes in ownership not been uniform. While general notions of greater efficiency are inevitably always in the background, reforms have not been designed to meet any clearly defined objective or carried out on a common timetable (Gillen, 2011a). This is not really surprising. The objectives have also seldom been clear and apparently, as Graham points out, have not generally been closely linked to actions. But actions there have been.

The first major privatization occurred in the UK in 1987 and by 2007 the International Civil Aviation Organization (ICAO) found in a study of 459 airports, that 24% were fully or partially in private ownership. Within Europe by 2008, 13% of the airports, mainly the larger ones, were owned by public-private shareholders and 9% were fully privatized. Between 1990 and 2005, 38 low and middle-income countries entered into privatization transactions that attracted investment of more than $18 billion (current prices).

In the UK, the major London terminals were sold to off as, albeit lightly regulated, commercial concerns. A more common action on continental Europe has been corporatization in the sense of being turned over to non-profit enterprises. In some parts of the world public-private partnerships were developed. But elsewhere, and often despite substantial rhetoric, the changes have been relatively minor. For example, under the US Airport Privatization Pilot Program only two airports have been privatized (Luís Muñoz Marín International Airport and Hendry County Airglades Airport) with preliminary acceptance for St. Louis Lambert International Airport.

Other forms of regulation have been imposed when states have withdrawn from ownership. To combat the quasi-monopoly positions many airports occupy, the authorities have often introduced price-caps or limited the profits of airports (rate-of-return regulation). Corporatization, with its associated non-profit element, is the extreme example of the latter. ⁷ In some cases, countries have introduced competition for the market by putting the operation and control of state-owned airport assets up for franchise. ⁸ This, for example has been adopted in Argentina and a number of African counties (Button, 2008a). The idea, stemming from Demsetz’s work (1968), is that airport operations, or parts of them, are auctioned off periodically but the airport itself remains a national asset. The notion is that franchising, providing efficient and complete contracts can be drawn-up, extracts the economic rents a monopoly airport operator may enjoy as well as stimulating efficient operations (Hart, 2003).

There is both an explicit set of empirical studies on implications of public sector in airport performance, and a more general analytical literature where ownership is used as a control variable when the primary interest in on other matters. While the bulk of airline regulatory studies has focused on the US's experiences, much of the airport work is European based (Gillen, 2011b). ⁹ One reason for this is that it is here that much of the variation has developed in recent years. Most European countries followed the UK by partially privatizing their airports in the mid-1990s; see Gillen and Niemeier (2008). Studies have been broadly of two types. The first consider an airport’s efficiency before and after privatization. The second compare public and private airports.

Early work by Parker (1999) on the privatization of Heathrow was of the second form finds no evidence the change improved the airport’s technical efficiency Added to this he finds BAA the economic incentives to operate more efficiently after privatization were distorted by government regulation. Running counter to this, Yokomi’s (2005) study of six privatized airports in the UK found efficiency improved from 1975 to 2001 with substantial growth on the non-aeronautical side was not considered by Parker

Seventy-one academic studies of various forms of airport privatization are compared in Graham (2011). ¹⁰ The paper highlights the diversity of approaches adopted in the various analyses. Many are largely descriptive qualitative assessment, but others make use of econometric techniques, such as stochastic frontier analysis, operations research (notably data envelopment analysis), or accountancy techniques (notably ratio analysis). The overall assessment by the author is that there is no consensus on the effect of privatization on airport performance. In part, this maybe because changes of ownership have been accompanied by other the introduction of other regulations, but it may also be that in cross-sectional comparisons, that there are significant differences in the natures of the privatized airports; essentially an endogeneity problem. With times series analysis, diversity in findings may well be associated with ways in which exogenous trends are incorporated. A not uncommon problem, for example, in difference-in-differences estimations which much of this work amounts to.

Going into more detail on some of the studies, competition between airports is often taken as a measure of their likely efficiency. Müller-Rostin et al. (2010), however, find there is little empirical evidence on the intensity of competition among European airports, concluding that “competition [between airports] is still minimal and not sufficient to prevent airports from abusing their market power.” Similarly, Wiltshire (2018) finds, again in the European context, limited evidence supporting the idea that secondary airports can compete effectively with larger neighbors. The recent trend for airlines to migrate services from secondary to primary airports is seen as supporting this position.

Airports, however, differ considerably. In terms of objectives, Vasigh and Haririan (2003) find that while privatized airports tend to be revenue maximizers, public airports are more concerned with traffic levels. Oum et al. (2008) after separating airports owned by one public shareholder from those with multi-level government involvement argue that different ownership and governance structures can affect the quality of managerial performance, Oum et al. (2006) concluding that public corporations are not statistically different from major private airports.

Vasigh et al's (2014) used total factor productivity and financial efficiency indices to look at the implications of privatization of airports between 2000 and 2010 in the UK, the US, and Latin. In terms of technical and operational efficiency, US airports outperformed those airports in the UK and LA groups, with the same results for financial efficiency albeit with the Latin American airports out performing those in the UK for five of the years. But what they did not find was any no conclusive evidence suggesting a link between private involvement and airport performance.

Studying a set of European airports, Vogel (2006) finds the privatized ones to be more cost-efficiently, and to earn higher returns on total assets and revenues, with public airports gaining from higher gearing and financial leverage. From an Australian-European semi-parametric two-stage study involving 10 years of data, Adler and Liebert (2014) concluded that under monopolistic conditions, airports of any ownership form should be subject to economic regulation. In competitive conditions public and fully private airports are equally efficiently but the latter set higher aeronautical charges. Mixed ownership with a majority public holding is neither cost efficient nor low price irrespective of the level of competition. Turning to aeronautical charges, Bilotkach et al. (2012) study of European airports finds that privately owned facilities have statistically significantly lower fees.

It is not just public ownership can affect economic efficiency, the detailed form of the ownership is important. Oum et al. (2008) in their stochastic frontier analysis of the world’s largest airports, find that countries considering privatization of airports should transfer majority shares to the private sector; mixed ownership of airport with a government majority should be avoided (even a 100% government owned public firm is better); airports operated by multiple authorities should be consider changing to independent airport authorities; and privatization of one or more airports in cities with multiple airports would improve the efficiency of them all.

Likewise, Gitto and Mancuso (2012) conclude when looking at Italian airports that, government/private ownerships with a private majority is more efficient than with government majority; liberalization of ground handling services increases airside and landside efficiency; granting all services to airport management companies can decrease efficiency when competition is reduced, and the introduction of a dual-till, as discussed below, increases overall technical efficiency. argue that the extent of managerial autonomy dominates the effect of ownership (Oum et al., 2003). Airport size also seems to matter in all of this.

So, a rather messy picture emerges. There is no simply formula that tells us what the implications of state ownership or, more perhaps more correctly, reductions in state ownership, are. It is not even clear in many studies whether ownership per se matters; its form and wider market often seems more germane.

Allocating airside capacity

From an airline’s perspective, the most important asset any airport offers are its landing and take-off slots. ¹¹ This makes them valuable assets for airlines that have use of them as this generates market power and can benefit most from their use. The pricing and allocation of slots are thus often heavily regulated, and especially so at congested airports.

Regarding charges within the European Union, the 2009 Airport Charges Directive applies to all airports handling over five million passengers per annum. It sets out minimum transparency requirements around how charges are calculated and mandating airports to consult airlines. At London’s Heathrow, charges are regulated by the Civil Aviation Authority. There are caps on increases in landing fees per passenger at the rate of inflation minus a stipulated percentage—the “(RPI minus X) formula.”

As in most major airports outside of the US, allocation of slots is also the subject to command-and-control regulation and at Heathrow they are allocated by the Airport Co-ordination Limited. In practice, most slots are grand-parented to incumbents usually within schedule coordination processes developed by the International Air Transport Association (IATA). Globally, about 175 airports are “schedule coordinated” in this way. In Europe, the process is mandatory for coordinated airports. Trading is often subsequently allowed with the aim of making optimal use of slots as demands patterns across the network change. For example, in 2012, the European Parliament adopted common rules for the allocation of slots that expressly allows carriers to buy, sell, and lease slots at EU airports. ¹² A review of the Directive by the European Commission (2014), however, found that while there have been some positive results, implementation of even such a light-touch framework had been inconsistent.

There are debates about the various traditional ways of pricing slots using landing weight as a proxy for cost so as to achieve a degree of equity across users (Levine, 1969; Morrison, 1987; Morrison & Winston, 2007). A large number of economists have, for example, focused on employing fees more closely reflecting the opportunity costs of a landing/take-off. Modifying the existing regime is one approach (Schummer & Abizada, 2017). The idea follows that of Pigou (1920) who argued that when there is excess demand for infrastructure one should incorporate a congestion element in the charge. ¹³ An alternative, following Coasian lines, is for slots to be periodically auctioned (Button, 2008e, 2008f). A third possibility favored by economists and, as noted earlier, used at some airports, is to allow secondary slot markets, whereby slots are initially allocated by administrative means but then their “owners” may trade or, in some cases, sell them (de Wit & Burghouwt, 2008). ¹⁴

Evidence on the outcomes of different regimes is limited. Most of the empirical work on the economic approaches to slot allocations involves case studies and, perhaps not surprisingly, these provide diverse results. Because they are more common, the studies usually concern slot trading. Early applied work in the US by Kleit and Kobayashi (1996), Sened (1997), and others generally concluded that slot markets had worked well and slots were not being underutilized. Later however, Fukui (2010, 2012) looking at four US airports between 1994 and 1999, and Pai (2010) found that secondary slot trading had mixed results when it came to the efficient use of capacity. This was supported in a follow-up regression analysis study of UK airports (Fukui, 2014) that found slot trades among partner carriers contributed to slightly more competition in terms of competitors per route, whereas the slot trades between rival carriers had a negative effect on the number of competitors.

While looking at the degree of competition that emerges after slots are trade is one metric for analysis, another used involves comparing fares across airports (Borenstein, 1989). Higher fares being correlated with greater monopoly control of slots. Larger airports, allowing for a variety of factors, for example, competition from other transport modes such as high-speed rail, and nearby airports can impose some discipline, seem to be associated with higher fares. Bel and Fageda (2010), for example, using data for 100 large airports in Europe found this. Low-cost carriers and airlines with a high market share seem to exercise countervailing power. But the extent of this varies across empirical studies. Again, replication is lacking.

Regulating supplementary revenue

The landside revenues earned by airports from parking fees, taxi rights, food concessions, etc. vary considerably, but in many cases form a major part of an airport’s revenue flow (Czerny, 2013; Graham, 2009). The extend this occurring depends on the nature of the customers being served and the role of an airport in the aviation network (Chen et al., 2020). These facts are, in turn, influenced by the wider regulatory framework governing airports.

Two airports are needed to offer connecting services and at least three for interconnecting services. In the first case, fees connected with the land accessing and egressing at airports are important, as well as money generated from inside terminal concessions. Many European airports, for example, largely involve short-haul flights by low cost carriers catering mainly for the leisure market. The airlines like to ensure rapid loading and thus an extended time between checking-in and boarding, basically holding passengers is the terminal, is common. Airports make money by offering refreshment and retail services. In the second case, the relative earnings for connecting hub airports are usually greater from on-site sources where captive travelers await connecting flights. Counteracting this, the banking of flights by airlines can limit these waits.

In either case, an airport not only has the opportunity to gain revenue from its role serving airlines, but also from the others services it provides to passengers. While in some cases the two sources of revenue may be considered independent of each other, in others there is a significantly high degree of cross-elasticity in demands. The associated interacting prices, imply two-sided markets. In the airport case, revenues from food and some others concessions do not influence the management and pricing of airside activities (Fasone et al., 2016; Ivaldi et al., 2012). ¹⁵ From a regulatory perspective, such low cross-elasticities mean the two markets can be treated as independent. They only need to be considered jointly when there are high-cross elasticities. Indeed, in this latter situation, an optimal strategy of a profit maximizing airport selling to both airlines and passengers may be to price below direct marginal costs on one side of the market to attract customers on the other; technically a static lost-leader strategy involving skewed pricing. ¹⁶

A problem that has emerged is whether each side of the market should be regulated separately. Froehlich (2010) argues, for example, that airports, while having features of two-sided platforms, are better treated within a conventional vertical-relationship framework. On the other hand, Appold and Kasarda (2011), while accepting that airports do have features amenable to other forms of analysis, feel airports are close enough to being two-sided markets to use them for regulating on that basis.

This has morphed somewhat in to the “one-till/two till” debate, involving whether airport price regulation should be applied only to airside activities (hence assuming two-tills) or to all airport activities (one-till) (Button, 2017; Czerny, 2013; Czerny and Zhang, 2015; Czerny et al., 2016; Gillen & Martin, 2013). The underlying distinction does not strictly revolve around the two-sided/one-sided dichotomy, however, because some airside sales, as noted, have little influence on airline activities, but there are significant overlaps. The practical challenge has been particularly pronounced in policy debates over price-cap regulation of the London airports (Czerny, 2006; Starkie, 2001, 2008).

Malavolti (2016) argues for a single-till approach to regulation on the grounds that concessionary revenues from shop sales vary according to the time people spend at an airport and thus relate to the transfer time between flights. This effectively takes retailers and airlines as the two-sides of an airport’s market. The interacting user externalities result because rapid transfers reduce airline costs but also reduce shops’, and thus airports’ concession, incomes. This, however, is unlikely to meet the Rochet and Tirole (2006) criteria for a two-sided market (Czerny, 2013). While food concessionaires will probably enjoy positive external gains from having more flights at the airports, additional or better eating facilities are unlikely to provide external benefits to the airlines by stimulating many more passengers with local origins or destinations to the airport.

Another problem is that not all airport concessions involve genuine reciprocal network externalities with airside activities. ¹⁷ Some are just complimentary ways of gleaning additional revenues from spatial monopoly power. This blurring between two-sidedness and dual-till extends, in varying degrees, to academic studies. Gillen (2011a), for example, looked at airports more generally as two-sided markets, arguing essentially that an airport is a platform linking airlines and passengers in the way Rochet and Tirole (2006) describes, but does not specifically focus on access. He therefore includes within concessions some single-sided markets such as restaurants, souvenir stands, and bookstores, thus making his analysis more akin to examining the dual-till issue than to strict analysis of two-sided markets. Van Dender (2007) finds that per passenger concessions revenue does fall with passenger numbers suggesting possibly some link between concession prices and passenger numbers. Whether this is causal is, however, not examined. ¹⁸

Ex post studies examining whether regulating supplementary revenues have affected the performance of airports, or of the implications of the alternative approaches are limited. They certainly help the finances of airports, and balance-sheet analysis highlights their importance to airport specializing in low-cost airline traffic (Fu & Zhang, 2010). There are also problems in isolating the implications of liberalizing airport regulations and airline deregulation. Many of the developments at airports have, at least partly, been the result of significant consolidation of legacy carriers based around strategic alliances and mega-hubs, and from the emergence of low-cost airlines (Francis et al., 2003). In this context, perhaps the most complete study is Bilotkach et al. (2012) examination of a panel of 61 European airports over 18 years. Analysis takes aeronautical changes as the dependent variable and in its analytical framework separates out airports subjected to single-till regulation and find them negatively correlated with aeronautical changes