An abbreviated version of the following paper was published in the Transportation Research Record, National Academy Sciences (1450), 1995, pp 3-7)
Airport Congestion and Noise: Interplay of Allocation and
Distribution
Wayne Brough, Edward Clarke and Nicolaus Tideman
I. Introduction
One of the central ideas of economics is that, in principle, questions of allocation are separable from questions of distribution. In practice, however, the two kinds of questions are often strongly intertwined. In this paper we show how, for the issue of airport take-off and landing rights, certain linkages between allocation and distribution have led to poor treatment of both issues. We also discuss what needs to be done to re-establish the separation of allocation and distribution, so that both issues can be handled well.
The paper begins with a discussion of the history of allocation of airport take-off and landing slots. It then discusses the way in which distributive concerns have dominated the resolution of allocative issues. It argues that a mechanism is needed to separate allocation and distribution, permitting the simultaneous optimization of the use of scarce resources and minimization of the social cost of unintended redistribution. From this perspective, the paper compares two allocative mechanisms: freely transferable, privately owned slots and an administered market using "compensated incentive compatibility," or CIC, a version of the demand-revealing process. The paper argues that if the only concern is efficient allocation of a specified number of slots, then either mechanism will work. However, if the number of slots is to be optimized endogenously, then CIC is needed. An additional important advantages of CIC arise from its capacity to provide optimal control of noise pollution. The paper extends understanding of CIC by demonstrating that this mechanism is useful not only for controlling public nuisances like pollution, for which the mechanism was originally developed, but also for allocating a purely private good such as airport slots, and especially for controlling the externalities surrounding such a good. The conclusions extend to other settings where allocation and distribution are intertwined, and where externalities need to be managed.
II. The History of Slot Management 1/
The decade following airline deregulation sparked great controversy over the role of the market, particularly in dealing with supply-side constraints (Hahn and Kroszner, 1989). The growth in air travel put pressure on existing capacity and created increased congestion. The creation of a market in airport take-off and landing rights (slots) in 1986 was viewed as an important, albeit limited, first step toward relieving one of the more important of the constraints on the supply of airline services and improving the efficiency with which scarce airport capacity is used.
The slot trading (or "buy-sell") program replaced a system of quotas for rationing capacity at some airports. The quota approach was initially established by the Federal Aviation Administration (FAA) in 1969 for allocating capacity at the four busiest airportsO'Hare, Kennedy, LaGuardia and Washington/National. At these airports, the "high-density" rule established slots, which represented the right to take-off or land within an interval of typically thirty to sixty minutes. The high-density rule replaced a first-come, first-served rule still in effect at the majority of U.S. airports.
At the four airports where slots were allocated, 4,700 slots were originally allocated by scheduling committees. Of these, 3,500 were assigned to trunk air carriers, while the rest were assigned to commuter airlines and general aviation. The scheduling committees met twice a year and were required by the FAA to reach unanimous agreement, or the FAA would step in and impose its own allocation. Both incumbent carriers and new entrants served on these committees, where the result was often extremely protracted negotiations. Trades in slots were occasionally possible, but generally only on a one-for-one basis for another slot at the same airport. In addition to inhibiting the most efficient use of existing slots, the quota system appears to have imposed considerable restraints on new entrants.
The FAA introduced the buy-sell program when scheduling committees were no longer able to reach unanimous agreement. This program, which became effective in April, 1986, "grandfathered" existing users by allocating to them the slots they were using at the four airports as of December, 1985. The rule also permitted the holders of slots to sell them.
The buy-sell program was intended to promote both efficiency and competition. Supporters of the program believed that:
If any carrier could gain greater net revenue from the use of a slot than the existing holder, it would purchase the slot. Efficiency would increase because slots would be operated at lower costs and be used for flights that gave air travelers greater benefits.
Competition for city-pair markets would no longer be restricted the existing carriers in that city-pair market. Any carrier that perceived a potential profit from the route would be able to enter. Potential entry was perceived to be very difficult under the airline scheduling committee arrangement, because of the previously mentioned tendency toward scheduling deadlocks.
Riker and Sened (1992) present empirical evidence on the relative efficiency of the slot market. Comparing two hubs that utilize the slot market (O'Hare and Washington-National) with two that do not (Atlanta and Los Angeles), Riker and Sened observed greater efficiency (in terms of load factors), smaller fare increases, a lesser degree of service diminution to smaller communities and considerable market activity in the sale and leasing of slots.
The main concerns with the buy-sell program have been that:
The grandfathering of slots generated windfall profits for the carriers who received such allocations. While competitive bidding, or "slot auctions," had been proposed to deal with this concern, such action was strongly opposed by existing carriers, because that would have require them to pay substantial sums for rights they had been using without payment.
Smaller regions would lose service because carriers would perceive greater value from using their slots for more heavily travelled routes. While this concern has proven to be unfounded, some critics maintain that the buy-sell program will not generally lead to the most efficient allocations.(2)
A pattern of dominance in slot ownership creates a barrier to entry that is particularly difficult for new entrants to overcome. In the absence of efforts to increase capacity or better use existing capacity at the high-density airports, competitive barriers and congestion problems are likely to grow in severity.
The General Accounting Office (1990) maintains that the buy-sell program prevents new entrants from competing with incumbent carriers, pointing to a combination of entry barriers at airports (i.e., slot and gate restrictions) and in the marketing of airline services. In preliminary work, as part of a two year study of entry barriers at airports, the GAO concluded that airports with majority-in-interest clauses, long-term exclusive use leases and slot constraints have, on average, higher air fares. Larger incumbents are alleged to hoard or "warehouse" slots. This perception has led to proposals for reallocating slots through outright withdrawal and reallocation to smaller carriers, or, more moderately, through "use or lose" rules. Critics have also observed that, while there was much early activity in the exchange of slots, the market has recently behaved rather sluggishly, except where large numbers of slots have been obtained in the course of bankruptcy proceedings or through in-kind trades.
Some critics have called for the outright abolition of both the high-density rules and the buy-sell program. Short of abolition, Congress has mandated remedial action, particularly to resolve the new entrant problem. In 1990 legislation (P.L. 101-508), Congress directed the FAA to initiate rule-making that would improve access and competitive entry for smaller carriers (Federal Aviation Administration, 1992). New rules will raises the minimum "use-or-lose" requirement for slots from 65% of the week to 80%. The periodic transfer of a small number of slots from larger incumbent carriers to smaller incumbent and new-entrant air carriers is also contemplated.
Proponents of the buy-sell program observe that such redistributions of slots, while encouraging entry, impose efficiency costs. To the extent that property rights in slots are attenuated, a degree of uncertainty is injected into the system, reducing the value of slots and encouraging the holders of slots to use them in ways that will reduce the probability of reallocation, which is not generally efficient. New entrants who receive slots may not value them as highly as the established carriers from whom they are taken. Accordingly, some observers, such as Kleit (1991), believe that the market would be improved by giving carriers clear title to slots, eliminating the incentive to make in-kind trades, and putting smaller carriers on an equal footing in bidding for slots. Kleit hypothesizes that a major reason for inactivity in the market for slots is that the visibility of the value of these rights could breed more political criticism about windfalls and consequently more redistributionist intervention.
A further difficulty with all existing slot management rules is that the use of slots does not reflect congestion costs. In determining its use of slots, a carrier can be expected to take account of the consequent delay of its own other flights, but not the delay costs it imposes upon others. As Hahn and Kroszner (1989, 106) note, "Slot ownership gives carriers a right to advertise the scheduled arrival or departure time, but not to purchase a guaranteed departure or arrival time or to be given priority in a queue." Efficiency requires further mechanisms for determining the number of slots in each period and for setting priorities among them.
The solutions that have been proposed to problems of efficient slot management have had distributive implications. Struggles induced by these distributive implications have made it difficult to get agreement on questions of efficiency. Therefore we proceed by first discussing the distributive aspect of slot management and then returning to questions of efficiency.
III. Principles of Distribution and their Application to Slot Management
There are three competing principles for settling distributive questions: The principle of efficiency, the principle of equality, and the principle of stability.
The principle of efficiency says that entitlements should be assigned in the ways that yield the greatest aggregate income. If this principle is used as an on-going distributive principle, it leads to continual redistribution, as changes occur in the identities of those who can use things best. Often it is possible to achieve efficiency without redistribution, by trade or by administered assignment and compensation. But when these non-redistributive ways of achieving efficiency are infeasible or not under consideration, appeals are sometimes made to the principle of efficiency as a distributive principle.
The principle of equality says that entitlements that are distributed by governments should be distributed in such a way that all persons benefit equally. The principle of equality is honored sometimes by assigning equal entitlements to all, and sometimes by selling or leasing entitlements to the highest bidders, with the revenue used for the benefit all.
The principle of stability says that the assignment of entitlements should not deprive people of the opportunities that they have become accustomed to having, unless they are compensated. As one application, the principle of stability says that when an opportunity becomes scarce and cannot be accorded to all who want it, entitlements should be assigned to those who have a history of using the opportunity.
All three distributive principles are attractive. But they sometimes conflict, and then it is necessary to choose among them. However, when appeals are made to these principles, there is rarely any mention of the principles that are sacrificed in the process of honoring the principle that is advanced. We now provide a review of the history of slot allocation with a commentary on the distributive principles that are honored and sacrificed by different approaches.
When slots were not scarce, all three principles could be honored by allowing everyone who wished to do so to take-off and land at airports, with charges only for the cost of managing the take-off and landing process. However, when congestion developed, the principle of efficiency was sacrificed. Equality was preserved. Stability was compromised. It continued to be possible for everyone who wanted to take-off or land to do so, but only at a cost of waiting in line.
The high-density rule of unanimously agreed quotas represented a compromise with all three principles. Unanimously agreed quotas were probably more efficient than the prior congestion, but because only limited subsequent trading was permitted, slots often remained in the hands of carriers other than those that could use them most productively. A kind of equality was achieved by allowing all carriers to participate in the scheduling committees. However, to the extent that the participants concurred in the view that a deadlock would lead the FAA to impose an unequal allocation, the bargains they reached reflected that inequality. Stability was sacrificed to the extent that existing carriers felt obliged to agree to the assignment to new entrants of some slots that they had previously held.
The buy-sell rules permitted slots to be transferred to carriers to whom they were more valuable, thereby promoting efficiency. The principle of stability was honored by setting initial allocation of slots according to usage at a time four months prior to the start of the buy-sell program. However, these rules sacrificed the principle of equality. A variation on the buy-sell rules, in which slots were initially auctioned, would have honored the principle of equality and sacrificed the principle of stability.
One alternative to the existing buy-sell program is a program of rationing slots by rental auctions or some other administered market. This has the potential to be at least as efficient as the buy-sell program, and possibly more efficient, because it may be easier under an administered market to vary the number of slots, to take account of changes in the optimal number of take-offs and landings. An administered market can honor the principle of equality by collecting the value of slots socially rather than assigning it to the carriers who had been using slots. However, by requiring carriers to pay for what they had previously been receiving without payment, an administered market sacrifices the principle of stability. Alternatively, an administered market can honor the principle of stability and sacrifice the principle of equality, if market-clearing fees are collected and allocated among carriers in proportion to their prior use.
The central distributive question with respect to slot management is: When restrictions on take-offs and landings are introduced at an airport for the sake of efficiency, should the value of exclusive take-off and landing rights go the carriers who had been using the airport, thereby honoring the principle of stability, or should they go to the public treasury, thereby honoring the principle of equality, or should there be some compromise between the two principles?
IV. Arguments Regarding Initial Slot Ownership
The argument for assigning initial slot ownership to the public is based on the general principle that the right to exclusive use of common property should be granted only in exchange for a payment corresponding to the value of that exclusive right. This principle implies that a history of use of common property does not create a right to privileged access when the opportunity to use such property becomes scarce. Consider the example of parking. If I have become accustomed to parking in front of the place where I work, and then growth of the area makes it efficient to install parking meters, I do not have a privileged right to continue parking without paying. The use of an airport slot is a similar exclusive appropriation of common property (the runway and flight path) for a span of time. When the quantity demanded at a price of zero is low enough that all who wish can be served, there is no unfairness in free access to runways. However, when the quantity demanded at a price of zero exceeds capacity, the principle of equality is violated if rights are assigned to previous users.
There are several opposing arguments that can be made for assigning initial slot ownership to those who have been using slots. One argument is that resources are wasted in the conflict over who shall receive the initial assignment of ownership, and this waste can be avoided by agreement that carriers will receive the rights. While this argument is logically valid, an argument of the same form can be used to defend any initial assignment. To make this argument telling, one must establish that the initial assignment of slots to existing carriers has a unique inherent plausibility that other initial assignments lack. Since assignment of initial ownership to the public has at least as much inherent plausibility, the argument fails.
A second possible argument for assigning initial ownership to carriers is that carriers deserve initial ownership because the value of slots is the result of their activity. There are difficulties with this argument as well. While it is true that slots could not have become valuable without a history of carriers providing service, there were other factors that were essential as well: the construction of airports, the development of an air traffic system, and the general growth of cities, for example. While it can be argued that pioneers in commercial aviation provided the public good of demonstrating that commercial aviation could be profitable, any such public benefits may have been fully compensated by earlier monopoly rights under regulation. Furthermore, there is little reason to believe that the value of such public goods provided by carriers would be highly correlated with the value of the scarce slots they were using. In addition, if it were known that whenever slots at an airport became scarce, rights to slots would be given free of charge to previous users, then there would be an inefficient incentive to use slots in order to acquire such rights.
A third argument for assigning initial slot ownership to carriers is that carriers use slots to provide the public good of connectedness with other cities. This involves an implicit partnership between carriers and the cities they serve, in which the carriers invest in developing scheduled service and cities provide airport facilities. It would be a breach of the implicit understanding of this partnership to introduce substantial fees for take-offs and landings. While this argument is plausible, it does not make sense to suppose that there is a perpetual obligation to allocate scarce slots to the carriers that happened to be using them when they first became scarce. If such arrangement were explicitly agreed by contract, it would be unenforceable because of the legal ban on perpetuities. Even if such a contract could be upheld, it would be unfair to future generations to require them to abide by a decision of earlier generations regarding a privileged assignment of rights to common property.
At the same time, it is reasonable to permit commitments of slots to carriers for some span of time, in exchange for the carriers' investments in developing schedules. The quantitative question that emerges is one that has no simple answer: When slots become scarce, to what extent, and for how long, should preference be given to existing users? Such an intermediate solution to the distributive issue is somewhat problematic, because it is not possible to deduce a particular outcome from distributive principles. For the polar outcomes, one can at least point to a justifying principle, even if it is necessary to ignore the fact that another valued principled is being sacrificed. Nevertheless, it is our view that any special entitlements of carriers to slots should be based on an assessment of the extent to which special access can be justified by the carriers' investments in developing schedules.
If slots are a common property resource, then their scarcity value ought to be shared equally by the members of the community whose common property they are. If prior commitments to carriers prevent the immediate implementation of such a rule, then there should be a transition, at the fastest rate that is consistent with prior commitments, from entitlements based on past usage to social collection of their scarcity value.
When a regulatory process favors present stakeholders, or when they can profitably invest in rent-seeking, any mechanism for separating allocation from distribution must address the means by which current entitlements are to be transformed into a structure in which rational individuals will not be motivated to employ resources in socially wasteful ways, seeking to maintain existing arrangements or create new ones that will be advantageous to them.
V. Separating Allocation from Distribution
The tendency of distributive concerns to influence allocation applies to a wide variety of institutional settings. Consider, for example, the regulation of air pollution and noise. When systems involving fees or auctions are compared with traditional "command and control" (CAC) regulation, the revenues collected from fees or auctions can substantially exceed the compliance expenditures associated with CAC. Tietenberg (1985, 100-113) reviews existing empirical work and concludes that comparative distributional effects often make CAC preferable to fees or auctions from the standpoint of the regulated community. In the case of air pollution, this creates pressure to limit consideration to options that incorporate grandfathering, such as zero-revenue auctions or trading of permits based on prior use.
The case of airport slots also illustrates the tendency for the choice of allocative mechanisms to be influenced by distributive considerations. If slots are to be owned, then to some people it will seem natural that initial ownership of slots be assigned to the carriers who have been using slots. On the other hand, if take-offs and landings are allocated by fees, then many people who thought it natural for ownership to be assigned to existing users for the case of ownership would find it natural for the fees to be received by airport authorities, rather than having them allocated among carriers in proportion to earlier use.
This association between allocation mechanisms and distribution can induce those who benefit from a particular distributive arrangement to advocate the associated allocative mechanism. Sound policy-making, however, requires that such associations not be influential when allocation and distribution can be separated, as they can in the case of airport slots.
To be confident that an allocative mechanism is chosen well despite the tendency of distributive considerations to influence the choice of an allocative mechanism, it is important that the distributive issue be settled separately. If the span of time for which existing carriers deserve special access is decided first, then it is possible to assess alternative allocative mechanisms without the bias that comes from potential distributive consequences. Then, after the allocative mechanism has been chosen, one can then adapt it to implement the previously chosen compensation for existing carriers.
VI. Allocating Slots by Transferable Ownership
The basic argument for allocating slots by transferable ownership is that if the owner of a slot is unable to obtain as much value from a slot as some other carrier, then self-interest will lead the owner to sell it. Thus a free market in slots leads to the most efficient use of slots.
This argument is valid for any specified number of slots, but it does not address the question of how many slots there should be. Allowing the wrong number of take-offs and landings can be very costly. Morrison and Winston (1989), for example, have estimated that welfare gains of approximately $4 billion per year could be achieved by optimizing the number of take-offs and landings. (See also, Congressional Budget Office, 1992.)
Planes seeking to land are given priority over planes seeking to take off, because of the greater cost of delaying landings. As the capacity of an airport is approached, increases in the scheduled number of landings increase both the expected length of time that arriving planes will have to wait while other planes land and the expected length of time that departing planes will have to wait for a chance to take off. Because of the random component of arrival times, any increase in the scheduled number of take-offs increases the expected delay for all take-offs. The optimal trade-off between number of slots and expected delay depends on the value of a slot, which varies by time of day. The number of slots in a given hour is optimal if the rental value of a slot in that hour for one day is equal to the cost of the additional delay that is caused by squeezing one more slot into that hour. If the marginal delay cost were less than the rental value of a slot, then to achieve efficiency under a system of transferable ownership of slots, the airport authority would need to create new slots. If the marginal delay cost were greater than the rental value of a slot, the authority would need to eliminate one or more existing slots to achieve efficiency.
A need to change the number of slots can also arise from a change in technology, a change in the number of runways, or a change in the estimate of the spacing required for safety. If carriers own slots, any increase in the number of slots reduces the market value of the assets of carriers, leading to pressure not to increase the number of slots. If efficiency is to be achieved, it is important that such pressure not be effective. Any reduction in the number of slots requires airport authorities to acquire slots from carriers. This would produce financial losses for airport authorities, unless there were some provision for taxing the owners of slots to finance required reductions in the number of slots.
One of the consequences of take-offs and landings is noise pollution. The cost of this pollution varies by time of day, type of aircraft, and how the aircraft are flown. To achieve proper incentives for reducing the costs of such pollution, those who take-off and land must be charged with these costs of their actions. Thus efficiency requires that possession of a slot not insulate carriers from noise pollution charges.
To summarize, it is possible in principle to achieve efficiency through transferable slot ownership, but there are potential problems in adjusting the number of slots in each hour, in adapting to the redistributive consequences of changes in the number of slots in each hour, and in implementing efficient noise pollution charges. These problems are not insurmountable, but they demonstrate that transferable slot ownership has costs, which could be higher than the costs of an alternative allocative mechanism.
Congestion fees collected by airport authorities provide an alternative to transferable slot ownership for achieving efficient use of airport capacity. However, carriers strongly resist congestion fees, because of the significant redistribution that they entail from airlines to airport authorities. Morrison and Winston (1989) estimate that, in achieving approximately $4 billion per year in welfare gains, optimal congestion fees would, in the short-run, result in transfers from airport users to airport authorities in excess of $10 billion per year. Others have estimated that a workable system of congestion fees would achieve significant economies at a much lower level of revenues. CBO (1992, 40) has estimated that reasonably workable congestion fees would yield between $1 and $2 billion per year. Noise charges would generate additional revenue.(3)
Interest in incentive-based systems as alternatives to CAC regulation was heightened by the enactment of the Aviation Noise and Capacity Act (ANCA) of 1990. The act required the phase-out of noisier (Stage 2) aircraft by late in the decade, at a cost of some $900 to $4,600 million (discounted), depending on how the phase-out was to be accomplished. Airports were to accept a quid pro quo that limited their ability to impose restrictions on aircraft that met Stage 3 noise standards. The Act also authorized up to $1 billion in passenger facility charges (PFCs) to finance airport expansion.
In general, local noise charges (such as those elaborated in Section VIII) as well as slot charges (Section VII) would be more efficient than uniform cutbacks or the phase-out of a significant portion of the aircraft fleet and flat-rate PFCs. Nevertheless, for distributive reasons, carriers may prefer the existing system to a more efficient one involving government collection of congestion and noise fees. In addition, the imposition of new taxes and charges, including local PFCs, has reduced interest in congestion fees as a potential source of revenue for airport expansion and heightened the resistance of carriers to their imposition. Nonetheless, this creates an opportunity for more active exploration of approaches in which efficient pricing is separated from any revenue enhancement objective. In this context, we develop in the following sections an incentive-compatible approach that can enhance allocative and procedural efficiency while making improved trade-offs between the principles of distributional stability and equality, in part by achieving a more effective separation of distributive and allocative issues.
VII. Allocation of Slots by a Compensated Incentive Compatible Process
The origins of our proposal are Vickrey's (1961) concept of a "second-price auction" and Clarke's (1971) idea of levying charges on individuals for public goods according to the marginal social costs of accommodating their stated departures from standardized demand schedules, an idea to which Tideman and Tullock (1976) gave the name "the demand-revealing process." The principal virtue of these mechanisms is that they motivate individuals to report their preferences honestly, thereby permitting efficient allocations to be identified and implemented. Dolan (1979) described the way in which the demand-revealing process could be employed to optimize the use of a congested facility. Economics, Inc. (1980) recommended that the FAA use a second-price auction to motivate accurate revelation of willingness-to-pay for slots.
In the 1980s there was considerable interest in applying incentive-compatible techniques to the allocation of airport slots. The published research centered largely on the appropriate organization of slot auctions, which were being experimentally tested by the FAA as well as by economic researchers (Rassenti, Smith and Buffin, 1982; Grether, Isaac and Plott, 1989). The second-price auction proposed by Economics, Inc. was experimentally evaluated, in part, by the FAA, but it is not clear that the theoretical advantages of the demand-revealing approach were dealt with adequately, if at all, in the actual experimental setting.
Some form of periodic auction tied to a buy-sell aftermarket received a good deal of support in government policy circles, until it was abandoned in favor of simple buy-sell with grandfathering. The auction approach was abandoned largely because, as with congestion fees, it entailed large and uncertain revenues and transfers. Executive Branch policymakers believed such an approach could never gain Congressional acceptance, in view of the objections of carriers. (For an interesting case study of this, see Kennedy School of Government, 1987.)
As the periodic auction was originally conceived, a primary auction would be followed by an aftermarket, which would achieve efficient adjustments to the auction outcomes. However, the auction could be expected to give rise to significant misstatements of demands and would not achieve efficient marginal-cost pricing of slots. It would also generate the politically unacceptable distributional consequences described previously. In the remainder of this section we offer a refinement of the second-price auction approach that enhances its efficiency, equity and acceptability, permitting significant welfare gains in a politically viable manner.(4)
Our proposal involves a variation on second-price auctions and the demand-revealing process that was developed by Tideman (1979). This variation, called "compensated incentive compatibility" (CIC), optimizes the use of a public resource while separating allocative efficiency from the distribution of the revenues generated by efficient pricing (in this case, of slots). Our proposal can also be considered a variant of the zero-revenue auction (ZRA) approach to emissions rights allocation, which has appeared in the environmental literature (Hahn and Noll, 1983; Hahn, 1984).
In the remainder of this section, we introduce the CIC approach to slot allocation as a refinement of traditional auctions and zero revenue auctions. We then introduce further refinements that take more precise account of congestion costs, by better reflecting differences in priority status in a queue, building on Dolan's (1979) work. Finally we develop a different application of CIC for dealing with the noise from take-offs and landings.
Traditional Auctions
Under a traditional approach to the auction of slots, each carrier submits a demand schedule for slots, i.e., a statement of the quantity of slots that the carrier would wish to rent at each possible price. These individual demand schedules are added horizontally to construct an aggregate demand schedule for slots. There is an exogenous limit on the number of slots, set by legislation or regulation. If aggregate demand is less than this quantity at a price equal to the sum of noise and other external costs, then this price is the price of a slot. If aggregate demand is more than the limit at this price, then the price is such that aggregate demand is equal to the limit.
The theory of just distribution that is incorporated in a traditional auction is that slots are public property, so that if they are scarce, any carrier that wants to use slots ought to pay for them according to their market price.
A traditional auction is efficient if the number of carriers is so great that no carrier considers it worthwhile to take account of the effect of his reported demand schedule on the price. When the number of carriers is not so great, there is a modification of the traditional auction that is efficient. This is to require each carrier to pay for each slot it obtains according to the marginal social cost of that slot. That is, the price of a carrier's first slot would be the lowest price at which all other carriers had an aggregate demand for all but one slot, the price of the second slot would be the lowest price at which all other carriers had an aggregate demand for all but two slots, and so on. This method is an application of the principle behind Vickrey's (1961) second-price auction and is the method of slot allocation recommended to the FAA by Economics, Inc (1980). Because each carrier would pay for slots according to their marginal social cost, it would have an incentive to report its demand schedule accurately. However, if the very considerable revenue that would be generated were to be returned to carriers, then there would be incentives for misstatements of demands.
Zero Revenue Auctions
In a zero revenue auction (ZRA), a system administrator assigns each carrier a provisional allocation of slots. As with a traditional auction, all carriers report their demand schedules. The intersection of aggregate demand with an aggregate quantity constraint determines the equilibrium price of slots. Each carrier's final allocation is its quantity demanded at the equilibrium price. Each carrier pays for slots at the equilibrium price and receives back an amount of money equal to the product of the equilibrium price and its provisional allocation. Net payments for all carriers taken together are zero. The theory of just distribution that is incorporated in the ZRA procedure is that each carrier has an entitlement to its initial allocation of slots, but this entitlement is protected by a liability rule rather than a property rule: Each carrier may be required to sell some of its slots to any other carrier that values them more highly.
One problem with the ZRA procedure, noted by Hahn and Noll (1983), and by others as well, is that when a carrier recognizes that it can affect the equilibrium price of permits, it will have an incentive to report something other than its true demand schedule. This problem can be remedied by the CIC method of auctioning slots.
Compensated Incentive Compatibility (CIC)
Under compensated incentive compatibility, the system administrator assigns each carrier a demand schedule as well as an initial quantity of slots. The assigned demand schedule should be the best estimate of that carrier's actual demand that the administrator can make without using information supplied by the carrier itself. (The use of that information would generate an incentive for biased reports.) The administrator can, however, use information supplied by similarly situated carriers to estimate any given carrier's demand. As with the other auction methods, the carriers also report their actual demand schedules. Assignments of slots to those who value them most highly lead to reallocations of some slots. But for any carrier whose assigned demand schedule is perfectly accurate, these reallocations occur without harm or benefit.
The amount that a carrier pays for its final allocation of slots is illustrated in Figure 1. Here D1 is the administrator's estimate of the carrier's demand schedule, while D2 is the schedule that the carrier actually reports. The schedule labeled S is the carrier's "derived supply schedule of slots." The quantity on this schedule at any price, p, is found by first determining the total number of slots that can be supplied to all carriers at a marginal social cost of no more than p. (For the case of a regulated number of slots, one would simply use that number.) One then subtracts the quantities demanded at a price of p by all carriers other than the one whose demand is illustrated in the figure. The result is the quantity of slots that can be supplied to this carrier at a marginal social cost of no more than p, and the schedule of all such quantities is the derived supply schedule of slots for the carrier. The carrier's initially assigned quantity of slots is labeled q0. The quantity of slots that the carrier would receive if it had reported a demand schedule of D1 is labeled q1 The quantity of slots that it actually receives, given that it reports a demand schedule of D2, is labeled q2.
Figure 1: COMPENSATED INCENTIVE COMPATIBILITY
Note: The following diagram has been corrupted. It will be reproduced. (Under construction)
Marginal Value of
a Slot
2
D1
q0 q1 q2
Number of Slots
The fee that is charged to the carrier whose demand is shown is the shaded area in Figure 1. This price is the sum of the integral of the carrier's assigned demand schedule from q0 to q1 and the integral of the carrier's derived supply schedule from q1 to q2 Either integral or both can be negative, in which case it represents a payment to rather than a payment from the carrier.
The reason that two different schedules are used to price different parts of one carrier's reallocation of slots is as follows. The movement from q1 to q2 is induced by the deviation of the carrier's reported demand schedule from its assigned schedule. To motivate an honest statement of the schedule, this movement must be priced according to marginal social cost, that is, according to the derived supply schedule. The movement from q0 to q1 is induced by the fact that q0 is not the quantity of slots that would be efficient to assign to the carrier if it had reported that its assigned schedule was accurate. Using the assigned schedule to price this movement reduces unintended redistribution for this movement to the minimum that is feasible.
If the administrator were to estimate all demands perfectly and assign quantities corresponding to the market clearing ones, then each q0would equal the corresponding q1 and q2 Therefore there would be no charges on any carrier. The theory of just distribution that is incorporated in this solution is that every carrier has a right to receive, without charge, the quantity of slots that it demands at a market clearing price.
An efficient variant of ZRA is created by using CIC with any initial allocations of slots that sum to the total available quantity. At the other extreme, if all carriers are assigned quantities of zero and demand schedules that have zero quantity at all prices, then CIC reduces to the efficient auction procedure described above. But every intermediate distributive outcome can also be achieved efficiently through CIC. If one's theory of just distribution says (as ours does) that carriers have a temporary but not a permanent right to their initial usage of slots, then this can be incorporated into the CIC approach by starting with initial allocations that correspond to prior usage and then reducing the initial allocations to zero quantities, either gradually or all at once after some chosen period of time.
In general, CIC does not achieve budget balance. In two-carrier examples, the sign of the budget imbalance depends on the sign of the covariance between the deviations of the two reported demand schedules from the corresponding assigned schedules. A positive covariance yields a budget surplus and a negative covariance a budget deficit. We do not know whether this relationship extends to examples with more than two carriers. To achieve the budget balance that must obtain in the end, the budget deficit or surplus could be allocated among carriers in proportion to their assigned quantities. This would induce a slight incentive for misstatements of demands, but it would be difficult to know what misstatements would be profitable. To eliminate the incentive for misstatements all but infinitesimally, the budget surplus or deficit could be assigned to the general funds of the government.
The idea of assigning demand schedules that represents the best possible estimate of actual schedules (without using information supplied by each carrier to estimates its demand) incorporates an assumption that overcompensation and undercompensation for movements from a carrier's initial allocation of slots have equal social costs. If undercompensation is regarded as more costly than overcompensation, then the appropriate adjustment is to use an assigned schedule that is biased toward inelasticity. The assignment of perfectly inelastic schedules would eliminated the possibility of unintended harm to carriers (and guarantee a budget deficit) if all slots were initially allocated, and any were then reallocated.
Noise pollution charges can be incorporated into the CIC simply by announcing the charges and asking carriers to report their demands given those charges. To incorporate the distributional premise that carriers are entitled to continue all or part of their past noise for some span of time, one would supplement the noise pollution charges with credits corresponding to those entitlements. With further refinements, priority status in queues can be incorporated as well.
CIC with Priority Pricing
Dolan (1979, 432) developed a slot allocation method that incorporated congestion costs defined in terms of the demands of carriers (or passengers) for a guaranteed arrival or departure time or a higher priority in a queue. He showed in a static queuing model that the priority prices that would minimize congestion were equivalent to the taxes generated by a demand-revealing process. A corresponding modification of CIC is to ask each carrier to state, for each slot in which it expresses an interest, not just the amount of money that it is willing to pay for the slot, but rather the schedule of amounts they would be willing to pay for different priority status. The system administrator would assign not just initial quantities and demand schedules, but a priority status for each slot in each carrier's initial allocation and a priority demand schedule for each slot in which each carrier might have an interest. The final allocation would be the set of slot allocations, each with assigned priority status, that maximized aggregate reported value, subject to the constraints of available capacity. As with the earlier version of CIC, each carrier would pay, according to its assigned schedules, for the movement from its initial allocation to the allocation that it would have had, had it reported that its initial allocation was accurate, and it would also pay the net cost to all other carriers that resulted from the deviation of is reported schedules from its assigned schedules.
Incentives for the System Administrator
As Dolan (1979, 434) has suggested, "The system administrator must be presented with the proper incentive to make reliable estimates . . ." which "should not be based on the amount of priority fees collected." The appropriate incentive could be provided by giving the administrator a commission (say some fraction of 1% of the measured increase in social surplus from improved allocation, from which a smaller percentage of aggregate redistributive harm from inaccurate schedules would be subtracted). In the short-run, with a fixed amount of runway time, the administrator would maximize the aggregate value of landings and take-offs, net of transacting and administrative costs and the assigned cost of redistributive harm. In the long run, the administrator would have an incentive to also equate the marginal value of additional capacity and the long-run marginal cost of capacity. With the proper motivation provided, the system administrator should be permitted to choose an arrangement more, or less, complex than the CIC/priority pricing approach described above, as long as all carriers were treated fairly. The administrator can, for example:
Use complex algorithms and a series of iterations to find optimum slot combinations (between city pairs, at particular times of day). Such a combinatorial approach recognizes that carriers value flights not just slots, and that "an airline's demand for a slot from a flight originating airport is not independent of its demand for a landing slot at a flight destination airport" (Rassenti, Smith and Buffin, 1982, 402). Using CIC for combinations of slots can improve on the RSB mechanism by incorporating incentive compatibility. For a suggestion along these lines, see also Banks, Ledyard and Porter (1989).
Use the CIC as a means of primary allocation of individual slots according to carriers' reported demands, while letting the market perform the more complex task of coordinating take-offs and landings. Following the primary allocation by CIC, the administrator can allocate slots evenly or randomly to particular sub-hour segments and let carriers buy, sell or trade them to refine the allocation.
The administrator's choice of technique will be determined by the usual marginal tradeoffs between externality costs and administrative transactions costs of reducing externalities, including the costs of obtaining and providing the relevant information.
Consider a typical airport that currently experiences about $200 million per year in delay costs (out of $8 billion for the system as a whole by the FAA's estimate). If, as the FAA estimates, 57% of delay costs are weather-related, then there is about $95 million in potentially controllable delay costs. An administrator who was given a commission of 1% of savings would have a potential income of just under a $1 million. Of course, there would be administrative costs as well as delays that were too costly to avoid.
The administrator's incentives can be structured to ameliorate other external effects in addition to congestion. Earlier, we mentioned that the entry of an additional carrier into a market may have social benefits not reflect in the return to the carrier. The provision of service to small communities may also have external benefits. Appropriately designed incentives for the administrator can drive the system toward maximum social efficiency, reflecting any externalities that can be quantified monetarily.
One of the externalities that is of special concern is noise. The next section describes the way that CIC can be used to integrate information from the generators and recipients of noise to optimize its control. This section also serves to illustrate the more general application of CIC to the control of externalities, such as pollution.
VIII. Managing Noise Pollution by CIC
Properly set noise pollution charges will motivate carriers to invest efficiently in quieter airplanes, to operate them in such a way as to economize appropriately on noise, and to operate at times of the day when noise is less costly. The allocation of noise pollution fees to those who are adversely affected by the noise can neutralize unjustified redistribution that would otherwise result from variations in the level of noise. But what is the right level of charges for noise?
Noise is a "local public bad." That is, the harm that one person suffers from noise does not add or detract from the harm that his neighbors suffer from the same noise. But as with local public goods, the effect abates with distance from the source. The total cost of airplane noise is the sum over all persons who are harmed by the noise of the harm that they experience. A CIC process can be used to motivate those who live in the vicinity of an airport to report this harm accurately and motivate carriers to economize on that harm. At the same time, the CIC process also manages distributive effects of airport noise, within a prescribed theory of just distribution.
To manage noise, one must begin with a technology for measuring noise. A good measurement technology should achieve cardinally meaningful measurements. That is, if one plane is measured to produce twice as much noise as another at a given location, then all persons in that location should experience twice as much noise from the second plane as from the first. We will assume that such a measurement technology exists. We will also assume that it is easy to know how much noise any plane produces in any place.
To create a noise management system, one begins by dividing the total area that is affected by noise into sub-areas that can be treated as each experiencing a uniform level of noise, for any pattern of airplane operation. One must also divide the week into time intervals in which noise has uniform costs. Let the number of sub-areas be S, and let them be indexed by s. Let Is denote the set of individuals who are affected by noise in sub-area s. Let the number of time intervals be T, and let them be indexed by t. Let there be R carriers, indexed by r.
In each sub-area, s, in each time interval, t, there is a marginal social cost function, F, for noise:
Note: The expressions have become corrupted. They will be reproduced at the conclusion of this paper.
Cst = Fst(Qst) = S fit(Qst) (1)
iIs
where Cst is the cost of noise to sub-area s at time t, fit is a function that expresses the ith person's marginal cost of noise for time interval t as a function of Qst, the amount of noise in sub-area s in time interval t. If noise has neither economies nor diseconomies of scale, then the functions fit will be constant functions. But these functions can also have regions over which they are increasing or decreasing.
The noise level in any sub-area at any interval of time, Qst, is the sum of the amounts of noise produced by individual carriers, qrst. That is,
R
Qst = S qrst
(2)
r=1
Each qrst is a function of the matrix of prices of noise in all times and places:
qrst = grst(P) (3)
where P is a matrix whose stth component is the marginal price of noise is sub-area s in time interval t.
The efficient pricing rule is that Trst, the payment by carrier r for noise produced in sub-area s in time interval t, should be the incremental cost of the noise generated by its planes. That is,
Qst
Trst =
Fst(Q)dQ
(4)
Qst\r
where Qst\r is the amount of noise produced in sub-area s in time interval t by all carriers other than r. Noise payments must be expressed as these integrals rather than the simpler and more customary products of price and quantity, to achieve a marginal price that is equal to marginal social cost even when carriers produce enough noise to affect the marginal social cost.
The practice of charging carriers the marginal costs of their noise will motivate them to economize efficiently on the production of noise. But it is also important to motivate individuals who are harmed by noise to report that harm accurately and to economize on the harm they experience. Economizing on harm can be achieved, for example, by insulating houses or, if people are particularly sensitive, by moving away from the airport area. Both the goal of motivating accurate reports and the goal of inducing efficient cost-reducing activity are accomplished by charging individuals for the net marginal costs to all others of the harms that they report. Such a net cost is the difference between the costs to carriers of the reduction in noise that is induced by the individual's reported preferences and the benefits to other individuals of the reduced amounts of noise.
What has been described so far is an example of an intricate, but nevertheless standard application of the demand-revealing process. The introduction of CIC modifies the process by giving individuals entitlements to non-zero valuations of noise reduction. One reasonable way of doing this is to define those with entitlements as the possessors of land, with entitlements proportional to the assessed value of their land. (Landlords would have an incentive to take account of the concerns of tenants.) The entitlement to noise concern can then be defined as an entitlement, in each interval of time, to the median, among all persons in the vicinity of the airport, of the reported cost of noise per dollar of assessed value of land. The charge or credit for each individual is then calculated by first determining the amount by which noise is reduced as a result of the departure of his preferences from median preferences per dollar of assessed value of land. The cost of this reduction in noise to carriers (or benefit in the case of less-than-median concern) is computed, and from this the value of the noise reduction to other individuals is subtracted (or the cost of the increase in noise is subtracted from the benefit to carriers).
In the same way that individuals are given entitlements to noise concern, carriers could be given entitlements to produce noise. One plausible specification of such initial entitlements would be that each carrier would have an entitlement to the level of noise that would be generated by its existing flights if it converted to Stage 3 aircraft at a rate of 20% per year. CIC would permit departures from this initial entitlement to equate each carrier's costs of noise abatement with the local benefits of noise control. Of course, except as a transitional accommodation, this would compromise the principle of equality and give carriers an unjustified privilege. The egalitarian application of the CIC would be to give carriers entitlements to the amount of noise that the median person wishes to produce, which would be none.
Due to restrictions on length, the following section was not included in the published paper.
IX. Advantages and Disadvantages of CIC
The advantages of CIC lie largely in those extensions which allow the administrator to internalize social costs. However, implementation of the system could initially set aside these more complex refinements, dealing simply with the allocation of slots, possibly with surcharges to reflect priorities.
The advantages of even simple CIC can be illuminated by identifying the ways in which it improves upon a simple zero revenue auction. Like ZRA, CIC has the ability to maintain the wealth distribution prescribed by existing permits. But it also can readily accommodate a smooth transition from the existing distribution to one that eliminates the privilege inherent in permits. Like ZRA, CIC attacks the problem of barriers to entry from a reluctance of carriers to sell slots. Unlike ZRA, CIC gives carriers an incentive to reveal their true demands. But CIC also has disadvantages of ZRA noted by Hahn and Noll (1983, 75): "The drawbacks of the Zero Revenue Auction are that it is somewhat more difficult to understand than simple grandfathering and it makes participation in the market mandatory, rather than voluntary."
Much of the benefit of CIC depends on the assumed ability of an administrator to determine the value of a good (slots) on the basis of reported values of others in similar circumstances. In our view, airport slots would appear to lend themselves particularly well to accurate valuations, as long as the administrator has an adequate incentive to make accurate valuations.
Slot holders in turn retain the right to retain slots, as long as they are willing to pay the opportunity cost to others of using the slots. While this helps provide a more competitive market, entry may remain inhibited where slots are one of several scarce resources (including gates) that must be obtained in order to make entry effective.
Like other market arrangements, CIC is vulnerable to collusive behavior. For example, two carriers operating in a duopoly situation could report artificially low demands and divide slots between themselves instead of paying their actual value. If they were given initial allocations, they could report coordinated artificially high demands and make the price of slots astronomically high for new entrants, just as a duopoly could refuse to deal with a new entrant under current arrangements. In the absence of collusion, there is no motivation to misstate the value of slots; if a misstatement is consequential, it reduces the carrier's profits.
Capacity constraints are a particularly significant factor inhibiting entry. However, the entry-inhibiting effects of capacity constraints can be lessened by adding capacity when this is worthwhile. CIC and other incentive-compatible techniques reveal the true value of additional capacity so that it can be added efficiently.
Any system that merely allocates slots according to private willingness to pay will not necessarily maximize social surplus, because it will not take full account of the benefits of greater competition. Borenstein (1988) has suggested that the results of a market for slots should be modified by adjusting private willingness to pay for the benefits of additional competition.
X. Conclusion
The CIC method advanced here offers a way of allocating airport/airway resources efficiently, while providing for a separate determination of who should properly receive the financial return from the scarcity value of these resources, as well as a description of how the distribution of that scarcity value should evolve over time. If we do not implement such a procedure, we can look forward to growing congestion at the largest 20 or so airports, leading to a repetition of the costly scenario that was played out over the years since the High Density rule was put into effect in 1969. That is, the gradual imposition of quotas and administered scheduling regimes, followed by pressure for market trading, which will be accepted reluctantly because, while it improves efficiency, it also reveals the substantial wealth that is conferred by current arrangements.
Implementation of CIC could generate considerable welfare gains, both at the airports where High Density and buy-sell rules have been implemented and at the 20 to 35 other congested airports. An initial evaluation of CIC could be developed from the FAA's analyses of allocation by auction, completed in 1980. With legislative authorization, such a system might then be implemented through a pilot/experimental effort and phasingperhaps first with the 15% of traffic at the High Density airports. An experimental effort with these four airports could be followed by some adaptation of the mandatory trading procedures contained in CIC, supplemented by incentive-compatible priority pricing. The experiment could then be expanded to other airports experiencing heavy congestion.
The case of airport noise also provides an excellent potential experimental setting for determining how to better reconcile distributive and allocative issues. The application of CIC to airport noise should show how it is possible to achieve significant efficiency gains, while preserving substantial distributional stability and leaving the issue of final distribution of benefits to be determined by distributive principles.
Finally, by demonstrating the possibility of achieving a smooth, efficient transition from an existing structure of entitlements to a more attractive one, the experiment would produce results that could be extended to other settings where more effective separation of allocation and distribution would provide more effective control of externalities.
References
Banks, J., J. Ledyard and D. Porter. "Allocating Uncertain and Unresponsive Resources: An Experimental Approach," Rand Journal of Economics, Vol. 20, No. 1, Spring, 1989.
Borenstein, S. "On the Efficiency of Competitive Markets for Operating Licenses," QJE, May, 1988.
Clarke, E. "Multipart Pricing of Public Goods," Public Choice, Fall, 1971.
Congressional Budget Office. "Paying for Highways, Airways, and Waterways: How Can Users Be Charged," Chapter 3, "Airways," May, 1992.
__________. "Reducing the Deficit: Spending and Revenue Options," February, 1993.
Dolan, R. "Incentive Mechanisms for Priority Queuing Problems," Bell Journal of Economics, Fall, 1979.
Economics, Inc. "The Allocation of Runway Slots by Auction," FAA Report No. FAA-AVP-3, April, 1980, Vol. I-III, esp. Vol. III, "Theoretical and Technical Issues for Implementation," Chs. 3 and 4.
Federal Aviation Administration. "High Density-Traffic Airports Slot Allocation and Transfer Methods," 57 FR 37308, August 18, 1992.
General Accounting Office. "Airline Competition: Industry Operating and Marketing Practices Limit Market Entry" (GPO/RCED-90-147), August, 1990.
Grether, D., Isaac, M. and Plott, C. The Allocation of Scarce Resources: Experimental Economics and the Problem of Allocations of Airport Slots. San Diego; Westview, 1989.
Hahn, R. "Market Power and Transferable Property Rights," QJE, November, 1984.
Hahn, R. and R. Kroszner. "The Mismanagement of Air Transportation: A Supply-Side Analysis," The Public Interest, Spring, 1989.
Hahn, R. and R. Noll. "Tradable Permits," Yale Journal Of Regulation, Volume 1, 63, 1983.
Kennedy School of Government. "The Department of Transportation and Airport Landing Slots," Case Program (C16-87-781.0), 1987.
Kleit, Andrew N. "Competition Without Apology," Regulation, Summer, 1991, p. 72.
Morrison, S. and C. Winston. "Enhancing the Performance of the Deregulated Air Transportation System," in Brookings Papers on Economic Activity (June, 1989) pp. 110-112.
Rassenti, S., Smith, V. and Buffin, R. "A Combinatorial Auction Mechanism for Airport Time Slot Allocation," Bell Journal of Economics, 1982 (13,2, Autumn, 402-17).
Riker, William and Itai Sened. "A Political Theory of the Origin of Property Rights: Airport Slots," American Journal of Political Science, November, 1991.
__________. "Common Property and Private Property: The Case of Airport Slots," November, 1992, unpublished.
Tideman, N. and G. Tullock. "A New and Superior Process for Making Social Choices" (with Gordon Tullock), Journal of Political Economy 84 (December, 1976), pp. 1145-1159.
Tideman, T. N. "Property, Liability and Incentive Compatibility: Towards Efficient Management of Urban Externalities," COUPE Papers in Urban Economics, 1979.
Tietenberg, T. H. Emissions Trading, Resources for the Future, Washington, D.C., 1985, chapter 6.
U.S. Government Printing Office. Regulatory Program of the U.S. Government, April 1, 1992-March 31, 1993.
Vickrey, William. "Counterspeculation, Auctions, and Competitive Sealed Tenders," Journal of Finance 16 (1961), pp. 8-37.
Warford, J. Public Policies Towards General Aviation, Brookings, 1971.
1. An abbreviated version of this section appears in the Regulatory Program of the United States Government U..S. Government Printing Office (1993, 14-15). For another discussion of the evolution of the "buy-sell" rule and the reasons it was adopted in lieu of other approaches such as auctions or congestion fees, see William Riker and Itai Sened (1991).
2. Borenstein (1988), for example, shows that for licenses generally, and for slots in particular, allocation of licenses by competitive markets is likely to lead to too much entry in crowded markets and too little entry in markets that could support few (i.e. one or two) firms. Borenstein takes note of the existing policy of reserving slots for commuters serving smaller communities, acknowledging that this has mitigated the loss-of-service criticism and may have improved social efficiency as well. In any case, the reservation of about a quarter of the slots at the four airports for commuter and general aviation use has preserved access to the major airports for small communities.
3. . The following figures provide some indication of the financial magnitudes involved. In 1992, airports collected about $1.4 billion in landing fees. This revenue is now supplemented by recently authorized passenger facility charges (PFCs) of about $1 billion per year, which airports collect for expansion of airport facilities. To estimate congestion fees, CBO extrapolated from an estimated $300 million level for annual slot rentals at the four High Density airports to about $1,700 million for all congested airports (CBO, 1993, 261). Taking account of the fact that some increase in fees is warranted to reduce demand to the point where marginal congestion costs are equal to price, we estimate that a workable level of slot fees would yield not much more than $2 billion per year. Noise fees, however, would need to be added to this. Tietenberg (1985, 105) reports that one study concluded that optimal noise charges would be 2.84 times as much as existing landing fees at one airport (Logan). But the existence of noise charges would reduce the value of slots. Still, the combination of optimal slot fees and optimal noise charges could amount to several billion dollars per year. Thus these fees and charges could well exceed the sum of current landing fees and PFCs at the noisier and more congested airports.
4. This paper abstracts from the issues of equitable and efficient recovery of explicit airport costs and pricing for commercial versus general aviation. In general, commercial aviation largely pays its own way while general aviation does not. In the 1994 Budget, the Administration has proposed a significant increase in GA registration fees, which will improve cost recovery. In addition, localized congestion fees for GA could improve efficiency in the use of existing capacity (Congressional Budget Office, 1992). Warford's (1971) treatment of the integration of registration fees, fuel taxes and landing fees for GA points toward a long-sought way of achieving both a fair allocation of system costs and a more efficient use of capacity. In addition to general aviation tax/pricing issues, there are issues concerning the prices that commercial users pay for both enroute air traffic control and airport use.