A New Deal for Airports

This is a draft of a lecture given on Tuesday 16th November 1999 as part of the IEA's Regulation Lecture series organised by the late Michael Beesley of the London Business School. This is reproduced with the kind permission of David Starkie

Regulation Lectures

Introduction

Airports are different from the other regulated utilities, or so it would seem. The debate on the fundamentals of the UK approach to regulating its privatised utilities has focused on electricity, gas, telecommunications and water, but airports tend to receive a passing mention, if that. This low profile has particularly suited BAA and, when it has run the risk of being pigeon-holed with the other utilities, it has itself argued that it is different. It did so, for example, when the windfall tax was on the cards. It was not successful, however, in ducking the tax and it also received a shock at the time of this year's budget when the Chancellor announced a special review of the competition issues surrounding BAA's ownership of most of London's airports.

But there is some justification to the claim that airports are different and it was these inherent differences that led to Michael Beesley, in last year's series, suggesting a very different approach to regulating airports. This evening I want to continue with this theme. I, too, want to point to some important differences that set airports aside from the other utilities and I will argue that, to a degree, this reflects some curious or less common economic characteristics of airports. This will lead me to propose significant changes to the present regulatory regime and, perhaps, in extremis, to proposals not too different from those put forward by Michael Beesley. But, before doing so, let me sketch out the regulatory approach to airports. This too is different from the approach adopted for the other utilities and associated with it are a number of issues which I will outline.

An Outline of Economic Regulation

The economic regulation of UK airports is governed by the 1986 Airports Act, but this legislation also includes more general powers that have the potential to impact significantly upon the business of the airport companies. For example, the Government is able to direct different types of air traffic to different airports within a "single system" and, using these powers, until the early 1990s, aircraft carrying only charter passengers were not permitted to use Heathrow Airport. Other provisions in the Act enable limits to be placed on the number of air transport movements at congested airports. This might seem uncontroversial, but airports long since considered congested have shown a remarkable capacity to handle, over time, increasing numbers of aircraft (assisted by improvements in techniques of capacity management). The Government, therefore, has powers under the 1986 Act (should it so wish to use them) to restrict both the volume and type of business handled by an airport company. Arguably, these powers allow for greater intervention in the airport industry than in any of the other regulated utilities.2

Price Controls

It is Section 40 of the 1986 Act which provides, crucially, for the designation of an airport. Designation (which is done by Statutory Order) imposes on the airport an RPI-X price-cap regime. Four airports have been designated, BAA's three London airports, as well as Manchester plc. These four airports are normally subject to a five yearly review of specified charges with the review of Manchester taking place one year later than the London review. The last round of reviews was conducted between 1995 and 1997. It is only selected 'airport charges' that are subject to explicit regulation (limiting the maximum amount that may be levied) and these are defined as those charges connected with the landing, take-off, and parking of aircraft, and with the handling of passengers through terminals. Individual charges as such are not subject to a price-cap, but the latter is applied to the overall revenue yield per passenger; the airport operator, therefore, has a degree of discretion with respect to the level of each individual charge and the relationship between them (although, in practice, this discretion is limited by the pressures that airlines can, and do, exert on the airport operator through their powerful trade associations and sometimes through governments making representations on their behalf).

By statute, the regulator is obliged to perform his economic regulation function in a manner which furthers the reasonable interests of users (as well as promotes efficiency, economic and profitable operation of airports, and encourages investment). Partly for this reason, but also because of past treaty obligations and custom in the international air transport industry, the determination of X during the quinquennial review is much influenced by what is commonly referred to as the single-till approach. Importantly, the regulator`s judgement regarding an appropriate level for the price cap takes into account not only the revenue generated by airport charges but, in addition, the revenues generated by activities such retailing within terminals and the provision of rental property and other services to tenants and licencees (airline companies, car hire operators, etc.) the fees and charges for which are not subject to a price cap formula.3 The two sources of revenue are combined into a single-till as a pre-requisite for judging whether the forecast of total revenue net of operating expenditure, provides a rate of return consistent with the firm's cost of capital and whether the projected cash flow is sufficient to sustain the prospective investment programme. The expectation of an inadequate return, or an insufficient cash flow, will lead to a less stringent price cap and vice versa.

Much of this price-cap review process is common to those reviews undertaken for other utilities. However, the single-till in the airports case, is an important difference. As Michael Beesley pointed out last year, whilst the other utility regulators have sought to isolate and bear down upon the natural monopoly elements, because of the single-till such focusing has not occurred in the case of airports.

Capital Restrictions
In addition to controls over selected prices charged by the designated airports, there are also constraints that bear upon the flows of capital in the airports industry. The airport industry in the UK is a mixture of private and public ownership. A number of regional airports are owned by publicly quoted transport and property companies, but others remain wholly within the public sector including the designated airport, Manchester. These latter cases are not subject to the usual capital market sanctions for poor performance, particularly the possibility of being taken over. But, interestingly, these capital market incentives are arguably weak in the case of BAA too. When the utilities were privatised the government chose to impose in most cases a so-called 'golden share' for strategic reasons. This has not prevented mergers and acquisitions taking place, particularly in the electricity supply industry, but in the case of the BAA the restriction was that no single share-holder or group of share-holders could control more than 15 per cent of the company. This restriction was subject earlier this year to a reasoned opinion by the European Commission which considered that it was in breach of EU rules on the free movement of capital. The Commission has now referred the matter to the European Court of Justice.

Framework reviews
Since the 1986 Act a couple of reviews of the legislative framework have taken place. The first, in 1984, essentially considered the process applied to economic regulation and resulted in proposals for "streamlining" the approach, the most significant being an intention to adopt the standard utility model with a single economic regulator. Thus far the review process applied to airports by the 1986 Act has been unusual; it has been the MMC that has conducted the initial review and made recommendations to the regulator. This differs from the usual approach where the industry regulator conducts the review and a reference is made to the Commission only in the event of a utility challenging the regulators conclusion.

The second review, in 1998, formed part of the Government's general review of utility regulation. The thrust of this review was to continue aligning the process applying to airports with that applying to the other utilities. The proposals include placing a primary duty on the CAA to further the interests of airport users; enabling the CAA to intervene when standards of performance have not been met or have led to disagreement with airport users; and providing the CAA with concurrent powers with the Director General of Fair Trading, under the Fair Trading Act of 1973, to refer possible monopoly situations to the Competition Commission for investigation (currently the CAA does not enjoy concurrent powers in this area). These changes will bring the CAA more in line with the other utility regulators.

These proposals aim to tidy the edges of the existing regulatory framework. As a consequence, they are essentially conservative and they either choose to ignore, or fail to get to grips with, deeper problems inherent within the current framework. It is to these that I now turn.

Issues Arising from the Regulatory Framework

The Nature of the Airport User
One element in the package of proposed measures aligning airport regulation with the standard utility model, is the placing of a primary duty on the regulator to further the interests of the airport user. But this proposal contains a subtle, but potentially significant, difference from the Government's proposal for the other utilities, which is to require regulators to protect the interests of consumers. Furthering interests is not the same as protecting interests and, if the object is consistency, it is not clear why the duty to be placed on the CAA is not also one of protecting the user.

This touches upon important developments in some of the regulated industries which have a bearing on how we might define the user in the case of airports. In the initial post-privatisation periods, most regulators were faced with a priority of setting price caps for services provided to millions of small customers. Subsequent deregulation in telecoms and energy has meant price cap regulation shrinking to focus on intermediate products (interconnection services in telecoms, transmission and distribution services in electricity in gas). In these intermediate 'access' markets the demand side is characterised by much smaller numbers of relatively large buyers, some of which are, in global terms, much larger than the access provider. These are not 'consumers' in the sense usually meant by politicians. (Airports regulation is ahead of the pack in this area, in that the large user issue has been prominent from the outset, although it should be noted that the activities of the airport companies also impact direct on large numbers of individual travellers.)

Given these circumstances, the utilities bill by placing a duty on the utility regulators to protect the interests of consumers is, arguably, somewhat backward looking in its conceptual framework, and might also be criticised for its lack of clear thinking on the duties of regulators in respect of the relevant intermediate markets. It is, for example, by no means obvious that the primary aim of regulation should be to 'protect' or, even less obviously, to 'further' the interests of users, not least since, as competitors in downstream markets, there should be no presumption that such users have common interests. For example, in many ,but not all, aviation markets there are the complications of bilateral air service agreements which often incorporate various barriers to entry.

In respect of congested airports, there is a further consideration. At an airport like Heathrow there are considerable scarcity rents which lead to rent seeking behaviour on the part of the different economic agents involved in the supply of air services.. Such behaviour is constrained in the case of BAA by economic regulation but why should a regulator be given a duty to allocate a higher weight to the rents of users than to the rents of suppliers, such as BAA, which is what an unqualified application of the principle furthering the interest of users would imply?

There are, therefore, important distribution issues to address which, in this case, take on an added significance, partly because of the size of the scarcity rents and partly because non-UK/EU users of the airport utility are a sizeable part of the market. This calls into question whether strategic trade theory has a relevance. In turn, this requires consideration of whether capital markets are open or partially closed, and we can note here attempts by US negotiators to shift the scarcity rents of Heathrow to US airlines, foreign ownership and control of which remains foreclosed.

Airport designation criteria
An airport to be subject to a price-cap has to be designated under Section 40 of the 1986 Act and, as pointed out above, four airports are designated, the same four that were designated at the outset of regulation. Why these four and why others have not been added remains unclear, which is perhaps surprising given that designation is the crucial trigger for economic regulation. But economic regulation does not come without (sometimes considerable) disadvantages. Therefore, one might have expected airports to be treated on the basis of consistent and perhaps self evident criteria when deciding upon which ones to designate.

In my 1994 lecture in this series, I questioned why, at the outset, the Scottish airports of BAA were not designated. As a group, they dominated the Scottish airport market and in scale terms were similar to Manchester which was designated. Subsequently, the Government reviewed the case for designating them and, at the same time and for the first time, set out criteria which it considered relevant to designation generally. These criteria include the extent of competition from other airports /transport modes and prima facie evidence of excessive profitability or abuse of monopoly position. To include both abuse of a monopoly position and evidence of excessive profitability as alternative criteria, is curious and appears contrary to the thrust of competition law with its focus on abuse. At what point profits become excessive is arguable and the danger is that any profits in excess of normal will be judged excessive. There is, after all, nothing inherently wrong with above normal levels of profit; such levels could indicate, for example, superior organisational efficiency and cost control and signal opportunities for profitable entry.

In the event, the Government concluded that there was no case for designating the two principal Scottish airports, Glasgow and Edinburgh, because there was no evidence of abuse of monopoly position or inefficiency. However, both airports had achieved high levels of profit and rates of return. This suggests that, in spite of having included profitability in the list of designation criteria, in practice the Government was inclined (correctly) to disregard it as a singular reason for designation. When announcing its decision, the Government added that it believed the threat of designation provided a strong incentive for BAA to control its charges. Possibly reflecting this, BAA did cap its charges on a voluntary basis (initially with the formula of RPI-3) at both Glasgow and Edinburgh.

This outcome begs the obvious question of whether this approach in Scotland is not a more appropriate way of approaching the economic regulation of airports generally. Rather than link the implementation of economic regulation to evidence that market power exists, (which appears to be the basis upon which the four airports were originally designated ) would it not be preferable to hold reserve powers which are put into effect only when there is evidence that market power is being exploited and no voluntary agreement can be reached?

Service quality
Whatever the merits of the current price-cap applied to the major UK airports, it does mean that the focus of the regulatory system is upon charges per se. But, should service standards be brought within the ambit regulation as well because of the incentives that price controls give for degrading service quality? If demand with respect to service quality is inelastic (i.e. a unit decrease in quality leads to a proportionately smaller decrease in demand) then it is possible that the cost savings from degrading the quality of service will exceed revenues foregone; profits will be enhanced as a consequence. For example, the quality of airside activities such as baggage handling, trolley services, holding lounges, can probably be degraded (perhaps severely) without any consequence for short term passenger demand (in the long term adverse reputation may have an impact). On the other hand, demand for some commercial activities at an airport are probably (highly) sensitive to quality of service. Queuing by passengers will have a marked effect on demand for duty free goods and food and drink purchases, for example. Does this suggest that quality could be degraded in some activities to increase net profits, but not in others? Not necessarily. The willingness of passengers to undertake discretionary spending could very well depend on his/her treatment airside, so that this interdependence of demand may reduce incentives to degrade quality in any element of the airport service.

Although not subject to formal regulation of its quality of service by statute, BAA has for some time produced a quality of service index and provided specific service guarantees to airline passengers and to tenants; it also has consultation procedures with airlines and shares the results of market research with airlines, concessionaires and others involved in service delivery at its airports. The Monopoly and Mergers Commission has also taken into account complaints made by airlines and consumers when reviewing the price-cap for the regulator.

Emphasis is now switching to the establishment of formal agreements between the airport companies and the airlines known as 'service level agreements'. After a period of experimentation, these are being introduced although not without controversy. The main area of contention is whether the airport company should be penalised when performance falls below agreed standards. BAA considers that the processes covered by 'service level agreements' are shared between the airport and the airlines and that penalties should apply potentially to both parties. The airlines, on the other hand, believe that penalties should be payable by the airport alone; they argue that it is the airport that is the monopoly supplier whilst airlines operate in a competitive market and poor performance by an individual airline leads to passengers transferring to other airlines.

Clearly, from a practical point of view there are a number of difficulties if service quality is to be formally regulated. Judging the appropriate quality of service is difficult and it would be uneconomic to over-provide quality and inefficient to under-provide. The quality of service to aim for is that which would exist in a competitive market for the services in question. Bearing in mind that different airlines have different (sometimes very different) requirements, it is probable that in a competitive market there would exist a variety of service qualities attached to which would be different prices. The user would then choose the quality/price package which most suited its requirements. Unfortunately, the emphasis on service level agreements tends to focus the debate on a uniform (and possibly too high) standard and, thus, on a lower level of welfare than is potentially achievable.

The structure of charges
As Michael Beesley remarked last year, little serious attention seems to be given nowadays to peak load pricing. This is perhaps regrettable given the pioneering role in this field of BAA's predecessor, the British Airports Authority. The Authority introduced a policy of pricing peak demand at Heathrow and Gatwick in the early 1970s and, although the implementation left much to be desired, it was the first attempt to introduce, on a large scale, efficient pricing signals for the use of airports. The rudiments of the policy still exist, albeit substantially modified, and now focus upon runways and the utilisation of space for parking aircraft. There has also been an overlay of environmental charges whereby noisier aircraft pay more, sometimes much more, than quieter aircraft. The policy has come under fierce attack in the past from US aviation interests that felt that the peak passenger charge for the use of terminals unfairly penalised the early morning trans-Atlantic arrivals traffic at Heathrow.

What is particularly regrettable is that the structure of charges, and its implications for economic efficiency, have not received more attention in the regulatory reviews. In fact, on the evidence of the three MMC reports reviewing Manchester, the Commission seemed to alter its position from positively encouraging the adoption of peak-load pricing (1987) to mild encouragement (1992) and then to indifference (1997), with no mention being made in the latter review of the earlier recommendations on the subject, in spite of their apparent neglect. It could be argued that by 1997 the cause was in fact lost; proposals to construct a second runway and provide substantial additional capacity were, by that time, well advanced. The time to have pressed the issue would have been in 1992 when a well-considered peak load pricing scheme would have tested the case for a second runway and informed a decision on its optimal timing.

That apart, we have a situation at London airports where demand greatly exceeds supply at the current level of charges. There is a commonly held view that to clear the current London market requires a substantial hike in charges. This, of course, would run contrary to the current paradigm which results in a continual reduction in the real level of charge in spite of the congestion, so that, as a consequence, a balance between supply and demand is achieved through quantity rationing. But, there is an argument that some increase in charges would probably improve the situation even if it were not large enough to actually clear the market. This arises from the importance of the structure of charges regardless of whether the average level of charge clears the market. The underlying point is that total demand can be broken down into a number of sub-demands (by routes, scheduled/charter traffic, transfer/originating traffic, etc.) each of which will have its own elasticity of demand. With excess demand, quantity rationing, will tend to lead to inefficiencies in the allocation of capacity between these sub-markets. An increase in charges will, however, induce different substitutions among the sub-markets (there will be a greater reduction of demand where price sensitivities are higher). This will affect the rationing of the (now smaller) demand in ways that might be expected to improve efficiency a little. For example, if low value users are priced out, there will be more capacity available for higher value users. Thus, although the overall allocation will tend to be inefficient, there will at least have been some movement in the desired direction (i.e. some reallocation from lower to higher value users).

It may also be possible to devise an approach to charges which clears, or more nearly clears, the market without greatly affecting the average charge. Suppose, for example, the passenger facilities charge (for the use of terminals) was abolished and the emphasis was placed entirely upon a fixed charge for landing aircraft (which might apply throughout the day if demand is at a constantly high level). This would bear more heavily upon smaller jets with fewer passengers, so that demand is reduced significantly in this part of the market, thus leaving demand and supply more balanced, but with an increase in the average size of aircraft.

A similar outcome in efficiency terms can be achieved by the introduction of a secondary market in take-off and landing 'slots' (entitlements for use of the runway at particular times). If incumbent airlines are free to trade their entitlements, the consequence will be that slots will be transferred to those airlines that place a higher value on their use. However, the allocation of slots is governed by European legislation and the Commission has so far opposed the trading of slots. This has not stopped a 'grey' market developing at both Gatwick and Heathrow, but in my view it would be better to legitimise the process and make it transparent. Nevertheless, a secondary market is not without its disadvantages; in particular it involves a lump sum financial transfer from entrants to incumbents with consequences for the balance sheets and financial strength of often competing airlines. Reducing the discrepancy between market clearing prices and the prices actually charged for the use of runways, remains the preferred course of action.

The Single-till
I have already outlined the unusual but central role of the single-till in the overall approach to the price regulation of airports. It is an approach that has been criticised especially for forcing down charges at congested airports below market clearing levels and sometimes below the (resource) costs of providing airside services. Given the total revenue requirement, the amount of revenue from airport charges (and thus the average level of those charges) is determined by the anticipated level of profits from retailing and property activities. As more volume is squeezed out of a congested facility, retail revenues are increased and, if forecast to continue, the price-cap review will (all other things being equal) increase X, thus reducing charges at a time when economic efficiency requires these to increase. Indeed, if the price-capped airside activities contribute a minor share of total revenues, the gearing effect means that modest changes in revenue requirements give rise to large changes in (price-capped) charges.

This is the more evident distortion resulting from the single-till approach, but there are others. Although the retail and property activities are formally excluded from the scope of the price-cap, nevertheless the approach, by taking into account these revenues when determining the price-cap, implicitly extends the range of activities subject to regulation. It means, for example, that the retailing activities are possibly subject to inefficient investment incentives. A further complication is that the overall regulatory approach requires the regulator to assess the airport company's cost of capital and an incorrect assessment will further distort investment incentives. The incorporation of retailing and property activities into the assessment, inevitably complicates this exercise and increases the potential scope for error. Their incorporation requires, for example, the regulator to take a view on the cost of capital for retailing activities but this is not an area in which regulators generally have much or indeed any experience.

In spite of the distortions that the single-till imparts to the overall process, it is, nevertheless, an approach which focuses upon the important complementary nature of the relationship between the airside and retailing activities; an increase in the demand for flights from a particular airport will increase the demand for related goods and services and for rented property at that airport.4 But there is a twist to this relationship that is particularly important. The retailing and property activities enjoy locational rents due to the fact that superior locations have an enhanced value (in just the same way that retailing properties in, say, Bond Street or Oxford Street enjoy similar economic rents from their own unique location).5 And because the retailing and property activities gain these locational rents, increases in traffic volumes at an airport will often produce significant increases in their profitability.6 For a profit maximising airport company with market power, selling in both markets, the effect of the demand complementarity linked to the locational rents, is to attenuate the normal, downward pressure on profits which would arise when increased air traffic volumes have to be bought at the expense of lower prices.7 This means that, as long as an airport has spare capacity and combines both activities, the incentive will be to set charges lower than if runways were a stand-alone facility.

This outcome has some potentially efficient properties. For example, if an airport is operating with excess capacity, efficient pricing of runways and terminals implies a level of charge that will fall short of cost recovery (including a reasonable return on capital employed). Absent the retailing activities, these efficient prices would not be achieved. If, however, there are economic rents from retailing activities, the complementary relationship will encourage the airport to increase these rents by lowering airside charges. That is, the returns from retailing may support a more efficient level of charges (although it is also possible that this situation could actually give rise to charges for airside activities that are below marginal cost).

Once the capacity of an airport is reached, however, the pattern of incentives will change. There is now nothing to be gained in not pricing the use of runways and terminals at market clearing levels; increased turnover in and profits from retailing cannot in the short-term be secured by reducing charges for runways and terminals below the level which equates demand with the capacity available. In these capacity constrained circumstances, the prices charged for the use of airside activities by an efficient airport business which combines these with retailing, will be the same as the prices charged by an efficient airport business wherein runways and terminals are a stand-alone facility. Of course, once sufficient capacity is added,, the situation will revert; the pricing of runways and terminals will take into account the contributory revenues from retailing and property.

Reform Options

More focused regulation
In the light of these important economic characteristics there are various options for regulatory reform which suggest themselves. The more obvious option would be to remove the single-till constraint whilst retaining price-cap regulation. In other words, maximum airport charges would be determined by a requirement to allow an appropriate return on airside assets disregarding the revenues and costs of retailing and property activities. This might be referred to as more 'focused' regulation.

More focused regulation should bring about improvements in economic efficiency, since it would have the effect of reducing the distorted investment incentives previously outlined. In terms of its impact on charges, the effects here are contingent upon circumstances. At congested airports it is to be expected that average charges will increase, perhaps significantly, but, to the extent that charges have been held by regulation below marginal cost, the effects of the increase in economic terms will be beneficial. At uncongested airports it is more difficult to judge the impact of more focused regulation on charges. This will depend upon whether the existing price-cap is actually binding on the airport company's charging decisions. The powerful economic incentives I have noted, whereby an airport company with significant earnings from retailing and with spare capacity has an incentive to keep charges lower than would otherwise be the case, means that the maximum allowable average charge may exceed what the company wishes to charge. If this is the case, charges will decline once the single-till constraint is removed. Alternatively, the price-cap may be binding, in which case charges will rise, but not necessarily to inefficient levels.

There will also be longer term impacts. Removal of retailing and property from the regulatory assessments can be expected to affect both the pricing of, and investment in, these activities. As these sectors adjust, factors such as their price-cost margin will change, and these changes will have feed back effects on pricing and investment in these sectors. For example, if price-cost margins in retailing increase over time, the marginal profitability of extra demand will increase, and this will tend to increase incentives to price lower and invest more in (complementary) runways and terminal facilities. These incentives apply equally to congested and uncongested airports.

Overall, it would seem that from an economic viewpoint there is probably much to be gained by doing away with the single-till approach. The risks from doing so are reduced substantially by the complementary nature of airport activities, and the incentives that these complementarities impart for efficient behaviour. However, if the regulatory framework was to be altered in this way it would be expedient to encourage airport companies to develop commercial activities that depend for their profitability upon the number of passengers using the airport. Less desirable would be moves to demerge, or hive off, these activities into companies separate from those owning and operating the runway and terminal infrastructure. Although a separate retailing or property company would continue to benefit from airport growth, the company retaining the airside assets may have less incentive to increase output. This is because the owners of such assets will no longer have a shared interest in the performance of the retailing and property assets (an interest which does exist if the company has an integrated structure). If the airport is operating at capacity this is perhaps of less significance but in the longer term it is likely to affect the incentives to invest. There might then be a requirement for greater regulatory scrutiny of the investment plans of the airport company to ensure that investment was at an appropriate level.

An alternative view is that, as a consequence of the demand complementarities, there would be incentives for separate companies to enter into contractual relationships that compensated the airport company for the beneficial effects that any expansion in its activities would have on the retailing company . To the extent that this was successful, we can note that all that would happen is that the effects of common ownership (integration) would be replicated by contractual arrangements, and the economic effects of demerger would be insubstantial. That is, demerger will only have significant effects if the contractual alternative is infeasible or costly (e.g. because of transactions costs) in which case these effects would not be beneficial.

Abolition of the price cap
A further option to consider is full de-regulation of both airport services and retailing activities. This involves the classic trade-off between the economic effects of market power and the distortions introduced by regulatory intervention. The option is, therefore, more/less attractive the greater/lesser the degree of competition among airports, and it raises associated questions concerning, for example, the common ownership of the three London airports.

But putting aside these questions for the time being, the economic factors that have led me to argue for the abolition of the single-till approach to regulation also lend weight to an argument for doing away with a permanent price-cap. The existence of economic rents from complementary retailing and property activities has the effect of reducing some of the adverse effects of market power. It provides incentives for lower pricing and higher investment in airside activities because the extra air traffic will generate higher rents in retailing and property. The bundling of both types of activity is, therefore, a factor that, at least in terms of its effects on the market for runway and terminal services, tilts the balance of advantage towards deregulation more than would be the case if runways were operated on a stand alone basis.

There is yet another factor associated with the economic characteristics of the industry which tilts the argument towards deregulation. Compared with the more traditional 'natural' monopoly examples, supply in the airport industry is probably characterised by increasing, rather than decreasing, long-run costs at quite moderate levels of output. That is to say, if we double the potential output of a sizable airport by doubling the capacity available for use, total costs will more than double. This observation was made, somewhat in passing, in MMC2(1991) and although there is not the hard, statistical evidence to support the proposition, it is, as I argued with David Thompson in 1985,8 a likely outcome of the complex way in which airports grow in size. The source of the airport monopoly, therefore, is not the usual economies of scale in the long-run production function, but the fixity of 'locational' inputs (i.e. good sites) and economies of scope associated with established air service networks The significance of this increasing cost argument is two-fold. First, even in the absence of congestion, prices in excess of average costs are not necessarily inappropriate. And second, in increasing-cost industries, regulation of prices based on allowances for normal or reasonable rates of return on capital may lead to inefficiently low prices. Thus, even though deregulation may, in spite of the moderating factors previously mentioned, lead to inefficiently high prices, the outcome is not necessarily worse than the regulatory outcome..

In addition to these economic arguments supporting the case for deregulation, there is a further argument which derives from the earlier observations made regarding the nature of the airport user. As we observed, the airport user from a regulatory standpoint is not so much the individual airline passenger but rather a relatively small number of airlines providing services in downstream markets. These airlines, when represented through their trade associations, are relatively large and sophisticated customers more than capable of challenging the airport operators. They have often done so and frequently have used the courts to put their case. Arguably, therefore, the airlines have a measure of counterveiling power. However, the counterveiling power of airlines would clearly be greater if the London airports were all in separate ownership. For example, it is likely that Ryanair's recent threat to leave Stansted unless a proposed increase in charges was ameliorated, would have carried more weight if Stansted had not been in common ownership with Gatwick and Heathrow.

This suggests that there is a trade off worth contemplating between abolishing or reducing the level of price control and separating the ownership of proximate airports. To achieve the level of restructuring required to advance airport competition and to ease the regulatory burden might not be too difficult. As the voluntary divestment of generating capacity in the electricity supply industry and the voluntary separation of Centrica from British Gas indicate, it does not necessarily require considerable political will (in the gas case the Government had rejected the MMC's divestment proposals but the company decided that restructuring was in the best interests of the shareholders). If BAA were confronted with the possibility of a much reduced regulatory burden in the event of divestment, it might be a trade-off that the company was willing to take seriously.

Taking into account the high level of existing demand relative to available capacity at both Heathrow and Gatwick, in the event of separate ownership, competition will centre less on prices and more on the second of the two competitive dimensions set out by George Yarrow earlier in this series. Separate ownership in this airports case is more likely to spur product development and innovation especially with a view to adding more capacity to the congested infrastructure. Although, as I have argued, BAA currently has an incentive to pursue the same end, competition is likely to sharpen creativity. Consequently, it might be argued that for the capital markets there might be little to loose by BAA being broken up into potentially competing parts (the sum of the profitable parts may even exceed the sum of the existing profitable whole) and everything to gain from a less regulated environment.

The principal difficulty with deregulation is likely to be associated with distributional issues (rather than issues of economic efficiency). At congested airports, removal of the price cap can be expected to lead to user detriments in the form of higher prices to users and to significant increases in the profits of the airport companies. Consequently, even if efficiency is improved (if profit gains to the airport company exceeded loss of profits to incumbent airlines), the weighting given to the user interest may be decisive and make deregulation difficult, if not impossible, to achieve. Increased competition between airports could help to mitigate such distributional effects, but, if deregulation was to be pursued as a serious option, it might be necessary to consider it alongside other measures designed to offset the adverse effects on users.

One such measure would be to allow the airlines take a stake in the airport. In Australia and New Zealand, for example, some domestic airlines own their own terminals and have thus vertically integrated some of the service functions typically provided by the airport company. However such developments provide opportunities for foreclosing entry into the air services market and new entrants, particularly in Australia, have faced this situation. Although access rules could be formulated to try to ease this problem, an alternative (structural) approach would be to demerge airport retailing into a separate company which is then established as a joint venture between the owners of the runway and terminal assets, and the (downstream) airline companies.9

A joint venture of this type would maintain existing incentives for the airport company to increase output and, in addition, provide incumbent airlines with similar incentives which they do not necessarily have at the present time (given the strategic competition among airlines for landing slots at congested airports). Such a company structure could provide a means for compensating the airlines for losses following deregulation. Whether the required amount of compensation could be achieved would depend upon the willingness of the airport company to provide incumbent airlines with enough equity in a joint venture. This might be possible for the following reason. Increased charges at congested airports could be expected to lead to an increase in the average number of passengers per aircraft movement and, because of this, there would be an increase in total retailing turnover and profits which both parties to the joint venture could share; there could conceivably be a net overall gain to the airport company after allowing for the lump sum transfer of enough equity to compensate the airlines for any increase in charges.

Conclusions

Overall, I believe the balance of the argument is against a continuation of the single-till approach. In the longer term, the approach introduces distorted investment incentives particularly by extending the scope of regulation; it requires the regulator to make difficult judgments regarding the cost of capital not only for the air transport sector but also for retailing and property. In the shorter term, at congested airports, it also leads to distorted pricing signals. These distorted economic incentives are probably reason enough to abolish the single-till approach. But a more telling reason for its abolition is that it is an unnecessary complication. A business which combines the landing of aircraft with retailing (and has spare capacity), will have, without the intervention of a regulator, an incentive to reduce charges to airlines and to expand output. The special factors in this favourable situation are strong demand complimentarities combined with location rents.

The added significance of the complementarities combined with economic rents, is that they also make it less likely, and possibly much less likely, that any market power that does exist in the market for landing aircraft and handling passengers will be abused. In fact, because the airport company has incentives to maximise throughput, the problem is more likely to be one of excessive use and congestion (rather than the classical problem of a monopolist restricting output).10 It is important, therefore, not to jump from the proposition that the operator of, say, Heathrow has significant market power (which is certainly the case) to the proposition that the operator can be expected to abuse that market power. Consequently, the special economic circumstances in the airport case tip the balance of the argument not only towards the abolition of the single-till approach, but also towards the abolition of price-cap regulation.

In addition to the underlying pattern of economic incentives there are other aspects to take into account which lead to the conclusion that serious consideration should be given to removing the formal price-cap at airports currently designated. Foremost amongst these is the new Competition Act which will give the CAA considerable powers to police and punish abuse of a dominant position thereby establishing ex-ante incentives not to abuse market power. With this in mind, if designation was set aside, it might be appropriate if this was accompanied by undertakings in respect of pricing at Heathrow (and possibly Gatwick and Manchester) given under general competition law; such undertakings may have an economic effect similar to the imposition of a relatively loose price-cap at Heathrow.

The overall thrust of such moves would be to shift airport regulation in the direction of the approach adopted in New Zealand, where the privatised airports are subject to reserve powers of price control under the Commerce Act, and, importantly, towards the approach now adopted for BAA's main Scottish airports. At the latter, charges are capped on an informal basis following an agreement with the CAA. This establishes an important precedent broadly along the lines of the approach I am suggesting here.

Finally, there is the question of timing if radical change is to take place. One necessary step is to notify the US authorities of any proposal to abolish the single till, as is required under the 1994 amendment to Article 10 of the Bermuda II Agreement. (Article 10 was amended so that there is no longer a requirement for the UK to maintain a single- till approach). A further consideration is how change can be introduced within the timetable of the regulatory review process. The next review (of BAA) was due to take place, starting next year, with completion expected in 2001, but the timetable has been put back a year to allow time for a decision on T5 and to allow for completion of the general review of aviation policy that the present Government has set in train. This delay provides an opportunity to undertake an appraisal of the regulatory framework as a whole, the outcome of which may make a further quinquennial review unnecessary.

Endnotes

This paper is based on joint work with Professor George Yarrow to whom I am indebted. I am also grateful to David Thompson for his comments. All views expressed here, however, are entirely my own.
Because of externalities associated with airports the planning control system is also onerous and the Government is heavily involved in the decisions regarding the scale and location of new capacity.
These fees and charges have, nevertheless, been considered by the MMC during the quinquennial review process with a view to judging whether their level was contrary to the public interest.
Reductions in real air fares can be expected to lead to additional demand for the complementary good because of changes in the cross-price elasticities, i.e. besides there being more passengers, a lower real air fare will encourage each passenger to spend more in airport shops.
As Martin Kunz ("Airport Regulation: The Policy Framework" in W. Pfä?¬er et.al. Airports and Air Traffic, Peter Lang, Frankhurt am Main, 1999) points out (14) these locational rents are part of the process of allocating limited space efficiently.
These related activities are not always undertaken directly by the airport companies but are frequently franchised. The terms of the franchise normally enable the airport to extract much of the rent. Note however that Section 41 of the 1986 Act potentially constrains the ability of the airport companies to increase these rents by limiting the number of franchisees.
The effect of the rents is to shift to the right the marginal revenue curve for airside services.
David Starkie and David Thompson (1985) "Privatising London's Airports", Institute for Fiscal Studies, Paper 16, London.
The "deal" put forward last year by Michael Beesley was essentially one between the Government and BAA (with the airlines offered the comfort of charges being pegged in real terms). However, in paragraph 49 of his paper he also hinted at the possibility of a joint venture between BAA and the airlines in relation to the development of Terminal 5. It would seem, though, that he viewed this "joint venture" in terms of airlines bidding for access rights to the new Terminal.
On this issue see Peter Forsyth, "Price Regulation of Airports: Principles with Australian Applications", Transport Research - E, 33, Issue 4, 297-309, 1997