FTC Staff Comment Before the Federal Aviation Administration In the Matter of Study of the High Density Rule

V940013
COMMISSION AUTHORIZED
BEFORE THE
FEDERAL AVIA nON ADMINISTRATION
DEPARTMENT OF TRANSPORTATION
WASHINGTON, D.C.
In The Matter of
Study of the High Density Rule
Comment of the Staff of the
Bureau of Economics
of the
Federal Trade Commission·
November 23, 1994
)
)
)
Docket No. 27664
• This comment represents the views of the staff of the Bureau of Economics of the Federal Trade
Commission. They are not necessarily the views of the Commission or any individual Commissioner.
Inquiries regarding this comment should be directed to Laurence Schumann (202-326-3359) of the FrC's
Bureau of Economics.

BEFORE THE
FEDERAL AVIA nON ADMINISTRA nON
DEPARTMENT OF TRANSPORTATION
WASHINGTON, D. C.
In The Matter of
Study of the High Density Rule
)
)
)
Docket No. 27664
I. 	Introduction and Summary
Comment of the Staff of the
Bureau of Economics
of the
Federal Trade Commissionl
The staff of the Bureau of Economics of the Federal Trade Commission (FTC) appreciates this
opportunity to respond to the Federal Aviation Administration's Notice of Study and Request for
Comments (NSRC). The FAA is conducting a comprehensive study of the High Density Traffic Airports
Rule (HDR) and requests comments on the effectiveness and viability of the HDR and any potential
alternatives to the rule. 2
The FAA adopted the HDR to help alleviate delays caused by congestion at certain high density
traffic airports (HDTAs). The HDR set the total number of operations-takeoffs and landings-allowed
during certain restricted time periods at the HDTAs and distributed this quota of operating privileges to
incumbent carriers. The HDR currently affects operations at four airports: Kennedy, LaGuardia, O'Hare,
I This comment represents the views of the staff of the Bureau of Economics of the Federal Trade
Commission. They are not necessarily the views of the Commission or any individual Commissioner.
Inquires regarding this comment should be directed to Laurence Schumann (202-326-3359) of the FTC's
Bureau of Economics.
2 See "Study of High Density Rule," FAA Docket No. 27664, 59 Fed. Reg. 15332 (April 1, 1994),
and "Notice of Public Meeting," FAA Docket No. 27664, 59 Fed. Reg. 48165 (September 30, 1994).
1

and Washington National. The number of operations allowed under the HOR is based on the estimated
capacity of each airport. The HOR requires carriers wishing to land or take off during restricted periods
to obtain the privilege to do so. These landing and takeoff privileges, commonly referred to as "slots,"
can be traded, sold, or leased.
The HOR serves an important economic role. By establishing a market for rights to land or take
off at capacity-constrained airports, the HOR helps ensure that a scarce resource, airport capacity, is
allocated to its most highly valued use. The FAA may wish to consider extending the HOR to other
airports that might be prone to congestion and delays due to excess demand for limited capacity during
peak time periods.
The HOR has been criticized for setting artificial constraints on airport operations; however, as
long as the number of slots reflects the airports' actual capacity, this criticism is inaccurate. Another
criticism has been that the HOR, by limiting entry, fosters the exercise of market power by incumbent
carriers.
Several observations are relevant to the above criticisms. The binding constraint on net increases
in operations has been the HOTAs' physical capacity, not the use of the "slot" system. If the slot system
were eliminated, the constraint on net operations would remain. And though increases in operations are
constrained by the HOTAs' physical capacity, recent (1992) amendments to the HOR created preferences
in slot lotteries that made it easier for new carriers to obtain slots at HOT As and for incumbent small
carriers to expand the number of slots they control. These slot reallocations do not increase the total
number of operations, but they may act to reduce concentration somewhat. Fundamentally, the HOR
promotes, rather than limits, new entry, because it creates a market in which potential new entrants can
obtain operating privileges. Before such a market was created in 1986, large incumbent carriers could
unilaterally prevent entry at the HOTAs by vetoing grants of operating privileges to carriers wishing to
provide new service. Moreover, the findings of empirical assessments conducted by Bureau of Economics
2

staff of slot usage in May-June 1990 (prior to the 1992 amendments) and in September-October 1993
(after the 1992 amendments) do not suppon the hypothesized exercise of market power.
The use of peak and off-peak landing and takeoff fees may be an alternative to the present form
of slot-based regulation. The FAA may wish to study this alternative. The information requirements
needed to implement peak-load pricing to allocate capacity efficiently could exceed the information
requirements needed to implement slot-based regulation. The benefits of peak-load pricing over slot-based
allocation are not apparent.
II. 	Expertise of the Staff of the Federal Trade Commission
The Federal Trade Conunission is responsible for enforcing the Federal Trade Conunission Act,)
which, among other things, prohibits "unfair methods of competition.» The staff of the Federal Trade
Conunission, upon request from federal, state, or local governmental bodies, comments on regulatory
proposals that may affect competition, consumers, or economic efficiency. In the course of this work,
as well as in antitrust and consumer protection research, nonpublic investigations, and litigation, the staff
applies established principles and recent developments in economic theory to competition and consumer
protection issues.
The staff of the Bureau of Economics has a longstanding interest in slot allocation and transfer
methods, as well as more general interest in issues involving competition and regulation in the airline
industry. This interest has been reflected in comments submitted by the Bureau of Economics staff in
previous FAA administrative proceedings on slot allocation and transfer methods, and in previous FAA
) 15 U.S.C. 41 et seq.
3

proceedings involving other aspects of airline competition and regulation. ~ In addition. the staff has
issued research reports on slot allocation and airline deregulation. 5
III. 	Background
Effective April 27, 1969, the FAA designated five airports-Kennedy, LaGuardia, Newark,
O'Hare, and Washington National-as HDTAs and established the High Density Rule (HDR) to govern
operations there. 6 The HDR, originally a temporary measure to reduce delays caused by congestion,
established limits on the total number of hourly "slots"-takeoffs and landings-during certain hours of
the day. The hourly slot quotas were based on the operating capacity of each airport as determined by
the FAA's Engineering Performance Standards ("EPS"). 	Slot quotas at Newark were suspended
indefinitely in 1970, although the airport is still classified as a HDTA. Slot quotas at the four other
HDT As were made permanent in 1973.
Before 1986, the incumbent airlines at each HDTA decided how slots would be allocated, with
any reallocation requiring their unanimous consent. The incumbent airlines could trade slots among
4 See the comments of the staff of the Bureau of Economics of the Federal Trade Conunission, High
Density Traffic Airports: Slot Allocation and Transfer Methods, FAA Docket No. 25758 (November 15,
1991); Slot Allocation Alternative Methods, FAA Docket No. 24110 (July 27, 1984). Comments on other
aspects of airline competition and regulation include, Elimination of Airport Delays, FAA Docket No.
24206 (August 30, 1994); Discussion Authority for Agreement to Shift Schedules, Department of
Transportation, Docket No. 44634 (February 17, 1987); and Charges for the Use of Metropolitan
Washington Airports, FAA Docket No. 25204 (April 13, 1987). See also Comments of the Bureaus of
Economics, Competition, and Consumer Protection of the Federal Trade Commission, M{1ssport Program
for Airport Capacity Efficiency (February 29, 1988).
S See D. Koran and J. Ogur, Airport Access Problems: Lessons Learned from Slot Regulation by the
FAA, Bureau of Economics Staff Report to the Federal Trade Conunission (May 1983), and J. Ogur, M.
Vita, and C. Wagner, The Deregulated Airline Industry: A Review of the Evidence, Bureau of Economics
Staff Report to the Federal Trade Commission (January 1988).
6 Federal Aviation Regulations Amendment No. 93-13, 33 Fed. Reg. 17896 (December 3, 1968).
4

themselves, one-for-one. 7 Increasing competition in the post-1977 deregulated era strained this system
severely;8 in many cases it broke down completely, freezing the previous allocation in place and denying
slots to new entrants. 9 The FAA's summary of the situation in 1985 stated:
As a result of the unanimity requirement and the lack of deadlock-breaking provisions,
the air carrier scheduling committees have found it difficult in recent years to reach
agreement on a schedule in full compliance with the High Density Rule .... Overall, the
scheduling committees are not currently functioning in a manner which provides for the
efficient allocation of slots, for rapid adjustment to market conditions and shifting carrier
needs and preferences, for adequate opportunity for expansion of operations, or for new
carriers to serve high density airports. 10
Starting in 1986, revised FAA regulations have encouraged development of a market-based slot
transfer system. II The initial allocation of slots was determined by "grandfathering" -that is, giving the
right to a slot to the carrier that had been operating it. 12 Subject to certain exceptions and procedures,
slot holders are permitted to sell, trade, or lease their slots. The FAA retains the right to repossess slots,
which the regulations describe as operating privileges, not property rights. 13 Slots may be withdrawn
for such reasons as making necessary allocations for international flights and implementing competition
7 See Department of Transportation, Federal Aviation Administration, High Density Traffic Airports:
Slot Allocation and Transfer Methods, Final Rule, 14 C.F.R. Parts 11 and 93, 50 Fed. Reg. 52180,
52185 (1985) ("FAA 1985 Final Rule").
8 See D. M. Grether, R.M. Issac, and C.R. Plott, "The Allocation of Landing Rights by Unanimity
Among Competitors," 71 American Economic Review 166 (1981).
9 See FAA 1985 Final Rule, supra note 7. See also Comments of the Department of Justice, Slot
Allocation: Alternative Methods and Slot Transfer Methods. Notice of Proposed Rulemaking, FAA Docket
No. 24110 and 24105 (August 6, 1984), pp. 11-12.
10 Supra note 7 at 52181.
II See FAA 1985 Final Rule, supra note 7, p. 52180, and 14 C.F.R. Parts 93.211 - 93.229.
12 According to the 1985 Final Rule (supra note 7, p. 52190), air carriers and commuter operators
holding permanent slots on December 16, 1985 were allocated those slots subject to certain withdrawal
provisions. See 14 C.F.R. Part 93.215.
13 See 14 C.F.R. Part 93.223 (a).
5

goals. 14 Funher, slots will be recalled for reallocation if used less than a minimum specified percent of
the time. IS
Amendments to the slot allocation and transfer regulations became effective November 1,
1992. 16 These amendments were adopted, in part, "to promote the availaoility of slots to new entrant
and limited incumbent carriers at the high density airports, and thereby enhance competition. "17 Among
other things, these amendments created the classification of "limited incumbent carriers," defined as
carriers and commuter operators holding fewer than 12 slots. New entrants and limited incumbents benefit
from preferences in the lotteries of unallocated slots. The 1992 amendments also restrict the transfer of
slots obtained in a lottery. During the 24 month period following a lottery, lottery-assigned slots can only
be traded for others at the same HDTA and cannot be sold to other slot-holders (or obtained through sale
or merger) if the carrier obtaining the slots would. as a result, hold more than 12 slots. The 1992
amendments also raised the minimum slot usage requirement to 80 percent.
IV. 	The Benefits of Market-based Slot Reallocation
The HDR was originally adopted as a solution to the problem of congestion and backup delays
at five (now four) of the busiest airports in the country. 18 Airport congestion can result from a common-
14 See 14 C.F.R. Part 93.223. For example, in March of 1986, the FAA withdrew 5 percent of the
slots from incumbent carriers for distribution via lottery to new entrant carriers.
IS Originally, the minimum slot usage requirement was 65 percent.
16 See Department of Transportation, Federal Aviation Administration, 	"High Density Traffic
Airports: Slot Allocation and Transfer Methods, Final Rule," 14 C.F.R. 93, 57 Fed. Reg. 37308 (1992)
("FAA 1992 Final Rule") .
17 See FAA 1992 Final Rule, supra note 16.
18 At the time that the HDR was first proposed, the FAA noted that congestion related delays of
varying magnitude were also found at airports in Boston, Miami, Los Angeles, San Francisco, and
Atlanta. The situations at New York, Chicago, and Washington, however, were described as "the most
(continued... )
6

propeny resource problem, which has the essential element that no one controls access to a resource that
is available in a fixed supply. In the absence of propeny rights for the fixed resource, competing users
overuse it. Here, the resource that is available in fixed supply is the capacity of the airpon. Physical and
technological constraints, such as the number of runways and terminals and the capability of the air traffic
control system, limit the density and frequency of operations. Overuse of the scarce resource manifests
itself in congestion and delay as the number of airplanes attempting to land or take off during the same
period exceed the airpon's physical capacity.
An economically efficient solution to a common-propeny resource problem would limit the use
of the resource and allocate rights of use to those who value them highest. The HDR, as amended in the
FAA 1985 Final Rule,19 largely accomplished this by creating a slot market. Slot quotas in place since
1969 had already set limits on landing and takeoff privileges at the HDTAs. By creating a largely
unfettered market for slots, the FAA allowed slots to be transferred to carriers with the most highly
valued flights. Other things equal, the value of a given flight rises as consumer demand for the flight
rises. Thus, the slot market-that is, the ability of carriers to buy, sell and lease slots freely-helps ensure
that the flights offered are those that consumers value the most.
The FAA's 1992 Final Rule 20 places some restrictions on the transfer of slots, which may reduce
the economic efficiency of the slot market. The 1992 Final Rule limits for a two-year period the transfer
of the slots obtained through a lottery. The rationale for adopting this two-year mandatory operating
period was "to assure that slots obtained in a lottery are likely to be used for operations by new entrants
18( ••• continued)
critical." The FAA did leave open the possibility that if congestion and delay became more of a problem
at these other cities, the HDR could be extended as appropriate. See High Density Traffic Airports:
Notice of Proposed Rule Making and Notice of Public Hearing, 33 Fed. Reg. 12580 (1968).
19 Supra note 7.
20 Supra note 16.
7

"-.
and limited incumbents and not sold back to other slot-holders. "21 But because slot transfers likely
represent efficient transfers of resources, absent any antitrust concerns,21 the restrictions on transfers
following a lottery would perpetuate for two years allocations that might be inefficient. The staff of the
Bureau of Economics suggests that the FAA consider amending the HDR in order to rescind the two-year
restriction on the sale of newly-acquired slots following a lottery.
Critics of the HDR have described it as artificially constraining capacity at the HDTAs. For
example, the National Commission To Ensure a Strong Competitive Airline Industry23 recommended
in its report, Change, Challenge and Competition: A Repon to the President and Congress, that the "FAA
review the rule that limits operations at 'high density' airports with the aim of either removing these
artificial 	constraints 	or 	raising 	them 	to 	the 	highest 	practicable 	level 	consistent 	with 	safety
requirements."24 As noted above, the hourly quotas at each of the HDTAs have been set by the FAA
based on the hourly capacity of each airport as determined by the EPS. According to the FAA,
The hourly quotas are set at the predominant IFR [Instrument Flight Rule] capacity, as
determined by the FAA. The predominant IFR capacity for each airport is the airport's
capacity under circumstances and configurations most frequently encountered when
21 Supra note 16 at 37309.
22 Slot transfers that confer market power or otherwise lessen competition would, of course, not
represent efficient transfers of resources. Nonetheless, antitrust concerns would not be limited to the 2-
year period following a slot lottery. Slot transfers are already scrutinized by antitrust authorities (see, for
example the discussion in footnote 36 below), and there is no reason to expect that transfers conducted
in the 2-year period following a lottery are more likely to be anticompetitive than slot transfers conducted
at other times.
23 The National Commission To Ensure a Strong Competitive Airline Industry, created by Congress
(Pub. L. No. 103-13, April 7, 1993), consisted of 15 voting and 11 non-voting members appointed by
the President, the Senate and the House of Representatives. The bipartisan Commission issued its report
in August, 1993.
24 Supra note 23, p. 9. The National Commission studied and made recommendations on a large
number of issues affecting airline competition and efficiency.
8

weather conditions preclude Visual Flight Rule (VFR) operation. The limitations in the
rule are predominately determined by groundside constraints. [emphasis added]25
If there is evidence that the FAA underestimated the capacity of the HOT As or if capacity at the HOT As
has increased since the HOR was adopted, then the FAA may wish to consider raising the slot quotas. 26
As long as the slot quotas reflect each airport's capacity, the slots are not "artificial" constraints on
airport operations. As long as the HOR slot quotas reflect the capacities of the HOTAs, the HOR
promotes an efficient allocation of this limited capacity by creating a market for slots.
In general, market determined prices provide important signals about resource reallocation. In
the context of the HOR, the prices at which slots trade inform decision makers of the social benefits of
additional investments in capacity. Slots for peak demand time periods, when capacity will be fully
utilized, ordinarily will be more expensive than slots for off-peak periods. By encouraging the shifting
of lower-valued flights to off-peak times, this cost differential delays the need to make expensive additions
to capacity.
These benefits of market-determined slot prices suggest two implications for the FAA's present
proceeding. First, rules restricting the transfer of slots, by constraining this market, may diminish its
efficiency. Second, if airports other than the four now covered may regularly experience delays due to
congestion, then extending the HOR to cover these additional airports may be a more efficient method
of rationing their capacity than the current method, of rationing by the queue (that is, by requiring
airplanes to wait in line for long periods, whenever the demand for landing or takeoff privileges exceeds
an airport's capacity).
2S FAA 1985 Final Rule, supra note 7 at 52181.
26 We express no opinion about how accurately the slot quotas correspond to the HOTAs' capacities.
9

V. 	The Competitive Implications of Slot Usage
The 1992 amendments to the HDR were adopted to promote competition at the HDTAs by
facilitating the entry of new carriers and by fostering the growth of smaller "limited incumbent"
carriers. 27 The rulemaking proceeding that led to the 1992 amendments was initiated under the Aviation
Safety and Capacity Expansion Act of 1990, which directed the Secretary of Transportation to initiate
rulemaking to consider more efficient methods of allocating capacity at HDT As in order to provide
improved opportunities for operations by "new entrant carriers. "28
Critics of the HDR had suggested that it restrained competition at the HDTAs by impeding the
entry of new carriers and by limiting the ability of smaller carriers to expand and become more
formidable competitors to the larger carriers. ~9 "Since there are only a few carriers offering slots for
TT FAA 1992 Final Rule, supra note 16 at 37308.
211 The "Aviation Safety and Capacity Expansion Act of 1990" was enacted as Title IX, § 9126(a) of
the Omnibus Reconciliation Act of 1990, Public Law 101-508 (104 Stat. 1388-371).
29 See, for example, Airline Competition: Effects of Airline Market Concentration and Barriers to
Entry on Airfares (General Accounting Office Report to Congressional Requesters, RCED-91-101).
According to this report, "Since an airline must obtain a slot to serve [the HDTAs], slot restrictions can
be a barrier to entry at these airports." (p. 22) Economists do not typically consider the need to purchase
an input, such as the ability to land or take off, to be a barrier to entry. Nonetheless, government
imposed constraints can be barriers to entry to the extent that they harm, rather than promote, consumer
welfare. See H. Demsetz, "Barriers To Entry," 72 American Economic Review 47 (March 1982). Taxi
medallions, for example, are often cited as barriers to entry since they place artificially low constraints
on the number of taxis operating in a city, causing harm to consumers (due to the inefficiently low
number of taxis) that far exceeds any benefits that they may confer. On the other hand, government-
imposed constraints can also promote entry by establishing property rights. For example, patents can
prevent new entry in existing markets, but they can also encourage new entry into newly developing
markets. Since existing markets that are protected by patents might not have been developed without this
protection, patents can accurately be described as facilitating rather than barring entry.
As long as the number of slots set at each of the HDTAs reflects each airport's capacity, it is the
limited capacity that acts as a constraint, not the need to purchase a slot. When viewed in this light, the
slot market does not constrain entry, but instead it facilitates entry. If the slot market did not exist, then,
given the limited capacity at the HDTAs, a potential new entrant might be unable to obtain rights to land
or take off. Indeed, prior to the adoption of the FAA's 1985 Final Rule (supra note 5), incumbent
carriers were often able to bar the entry of new carriers at the HDT As by denying them operating
(continued... )
10

sale at any given time, securing the necessary operating rights might be extremely difficult. "30 As
evidence of the anticompetitive effects of the HDR, critics have pointed to the higher concentration31
and higher ticket prices at HOT As relative to other airports. 32
Although markets that are noncompetitive may show high concentration and high prices, these
characteristics do not, by themselves, necessarily imply the existence of anticompetitive behavior. Even
if markets were perfectly competitive, fares at capacity-constrained airports would be higher, all other
things equal, than at airports without capacity constraints. Equilibrium competitive prices for air travel
at capacity-constrained airports would reflect the "scarcity rents" associated with the constraint on
29( ..• continued)
privileges. See, for example, the discussion in Koran and Ogur, supra note 7, p. 16.
Some critics have suggested that, since slots were originally distributed for free, potential new
entrants are at a competitive disadvantage because they have to purchase slots in order to begin service.
The relevant concept of cost in determining economic decisions, however, is that of opponunity cost. The
cost the current owner incurs to use a slot is the revenue that the owner would receive if the slot were
sold or leased to another carrier, because that amount of money is foregone every time the owner uses
it. This cost is identical to the cost that a new entrant must pay in order to buy or lease it. Although the
original distribution of slots increased the value of the firms that received slots for free (and the wealth
of their shareholders), now the relevant economic decisions concerning output, prices, entry, or exit are
based on the opportunity costs, which are the same whether a firm is currently operating a slot or
considering purchasing or leasing it.
30 Policies for the Deregulated Airline Industry, Congressional Budget Office (1988), p. 55.
3\ Based on slot holdings, each of the HDTAs were either moderately or higWy concentrated in 1990,
according to the standards of the 1984 Department of Justice Merger Guidelines (although newly revised
Merger Guidelines were issued by the DOJ and FTC in 1992, the relevant classification of market
concentration is the same as that in place in 1990). See Comment of the Staff of the Bureau of Economics
of the Federal Trade Commission, High Density Traffic Airpons: Slot Allocation and Transfer Methods
("1991 Comments"), FAA Docket No. 25758 (November 15, 1991), Table 1.
32 The report Airline Competition: Effects of Airline Market Concentration and Barriers to Entry on
Airfares (General Accounting Office Report to Congressional Requesters, RCED-91-101), concludes that
slot restrictions increase average or median fares eleven percent on short haul routes and four percent
from all routes to and from HOTAs.
11

capacity.33 Indeed, the absence of higher airfares at slot constrained airports would imply (again, all
other things equal) that landing rights at HDT As were not a valuable resource, and slots would not trade
at positive prices. 34 Thus, higher fares at HDTAs may merely reflect scarcity rents. not monopoly
rents. 35
Finally, the concentration levels at the HDT As may overstate the potential for anticompetitive
behavior because the HDTAs are not necessarily relevant antitrust markets. In antitrust analysis, a
relevant market must be defined before concentration can be measured. For each of the HDTAs, the
relevant market may include more than the capacity-constrained airport. For example, each of the four
airports governed by the HDR is near one or more regional airports that are not slot constrained: Newark
33 Scarcity rents arise when competition for a resource causes the price of that resource to rise above
the marginal cost of "inframarginal" units. Although at the margin the price of the scarce resource will
equal its marginal cost, this price will exceed the marginal cost of all other units of the resource. Hence,
the inframarginal units will earn a "profit" (or, more precisely, an economic rent) generated by the
scarcity. Here, scarcity rents reflect the limited capacity and, in tum, the limited number of slots
available. In contrast, monopoly rents result from anticompetitive output restrictions. Here, monopoly
rents would be additional rents due to slot holders' decisions to exercise market power by restricting slot
use in a way that prevents existing slots from moving to their highest valued uses.
34 In a competitive market, a slot's value would equal the net present value of its associated scarcity
rents. Thus, the GAO's conclusion, that it is slot restrictions that cause higher fares at HDTAs, may be
imprecise. Fares at HDT As will reflect the cost of purchasing slots just as they will reflect the costs of
other inputs. Fares may be higher at HDT As simply because at other airports that are not capacity
constrained, capacity is not a scarce, and consequently costly, input. If there are airports other than the
HDTAs that are capacity constrained, then scarcity rents will be reflected in higher costs for air travel
from these airports. Not all of the rents, however, will necessarily be reflected explicitly in airfares.
Some of the rents may be "paid". through diminishing the welfare of travelers who face added congestion
and delays at these airports.
35 Several studies have concluded that high fares at concentrated hubs represent the exercise of market
power, in addition to whatever scarcity rents may be generated by capacity constraints. See, for example,
S. Borenstein, "Airline Mergers, Airport Dominance, and Market Power," 80 American Economic
Review 400 (1990) and W. Evans and I. Kessides, "Localized Power in the U.S. Airline Industry," 75
Review of Economics and Statistics 66 (1993). Others have argued that higher fares at hub airports are
consistent with competitive behavior. See Kleit and Maynes, "Airline Networks as Joint Goods:
Implications for Competition Policy," 4 Journal of Regulatory Economics 175 (1992). Despite these
differing interpretations, none of these authors identifies the system of marketable slots as a source of
competitive problems in airline markets.
12

in the New York metropolitan area, Midway in the Chicago area. and Dulles and Baltimore-Washington
International in the Washington, D. C. area. If flights using other local airports are in the same antitrust
market as flights from an HOT A (for example, because consumers can choose to fly from Dulles,
National, or BWI),36 the competitive presence of the other airports can constrain the prices of flights
using the HDTAY In that case, the relevant concentration levels would be those for the larger airport
market, not for the HDTA alone.
Thus, higher concentration and higher fares at HOT As relative to other airports could be
consistent with either competitive or anticompetitive behavior. When the FAA was considering the
proposals that ultimately became the 1992 HDR amendments, the Bureau of Economics submitted
comments on those proposals that tested a number of the implications of anticompetitive theories of the
effects of the HDR. 38 Using slot usage data supplied by the FAA's Office of Slot Administration, the
36 An illustration of how the HOT As and their regional competitors could be considered in the same
antitrust market is provided by the positions the Department of Justice ("DOl") took in Eastern Airlines'
bankruptcy proceedings. The DOJ opposed United's proposed purchase of Eastern's National Airport slots
and facilities on the grounds that it would "lessen competition" for flights between Washington and other
cities since United was already the dominant carrier at Dulles International Airport and the acquisition
would have given it control of 20 percent of the slots at National. See "Justice Dept. to Dispute Deal for
Eastern Slots," Washington Post, February 15, 1991, p. B2; 60 Antitrust & Trade Regulation Repon 152
(January 31, 1991), 186 (February 7, 1991), and 272 (February 21, 1991). The DOJ, however, did not
object to Northwest Airlines' purchase of Eastern's slots and facilities at National even though Northwest
had more slots than United at National p.rior to the Eastern bankruptcy sale, because Northwest's share
at Dulles was less significant than United's. The DOJ analysis is consistent with a relevant antitrust
market consisting of at least both National and Dulles. What is pertinent, therefore, is how the sale of
Eastern's National assets affected concentration in this antitrust market, not how it affected concentration
at just National. Similarly, even though United was the largest carrier at O'Hare, the DOJ did not object
to United's purchase of Eastern's O'Hare slots, a position consistent with considering flights to and from
Midway Airport to be in the same antitrust market as flights to and from O'Hare.
37 The fact that slots are scarce at one airport and thus sell at positive prices, while they are not scarce
at another airport in the same region, is not alone an indication of whether or not the two airports are
in the same antitrust market. Two regional airports can be in the same antitrust market even if prices for
slots vary between the airports. The primary factor in determining market definition is not price, but the
degree to which travelers will switch from one airport to another in response to changes in relative prices.
38 See 1991 Comments, supra note 31.
13

1991 Comments tested two principal hypotheses. The first hypothesis was based on the standard analysis
of market power in which dominant firms have a greater incentive to reduce output than do fringe firms.
Standard theory predicts that smaller fringe firms will increase output in response to a contraction of
output by the dominant firm. 39 It has been suggested that larger carriers will attempt to restrict output
by reducing the number of days on which a slot is used. 4O In the context of an airport slot market, this
implies that larger carriers would use their slots less intensively than the smaller fringe carriers would.
If larger carriers have market power and exercise it in this manner. slot usage would be negatively related
to the market share of the carrier using the slot.
The second hypothesis that was examined in the 1991 Comments was that dominant carriers
would be unwilling to sell slots to potential new entrants at competitive rates but instead would lease their
slots to selected other carriers, with the purpose of deterring entry by carriers likely to increase overall
slot use. In this scenario, dominant airlines would choose their competitors by leasing slots to them. In
this way, the dominant firm(s) could both ensure that less efficient firms will be competing with them
for the same passengers and simultaneously deter the entry of more efficient firms. 41 The testable
implications of this hypothesis are: (1) that the dominant air carriers will be net lessors of slots; and (2)
39 See A.A. Alchian and W.R. Allen, Exchange and Production: Competition, Coordination and
Control, 3d. ed. (1983), pp. 266-268.
40 See High Density Traffic Airpons: Slot Allocation and Transfer Methods: Supplemental Notice of
Proposed RulemaJdng, 56 Fed. Reg 4667 (September 13, 1991), at 46678.
41 See U.S. General Accounting Office, Airline Competition: Industry Operating and Marketing
Practices limit Market Entry, U.S. General Accounting Office Report to Congressional Requestors,
RCED 90-147 (1990). For a theoretical discussion of this effect, see K.E. Rockett, "Choosing the
Competition and Patent Licensing," 21 RAND Journal of Economics 161 (1990), which examines a single
incumbent (with a patent) facing a limited number of potential entrants. For a general discussion of the
fragility of entry deterrence models, see D. Malueg and M. Schwartz, "Preemptive Investment, Toehold
Entry, and the Mimicking Principle," 22 RAND Journal of Economics 1 (1991), and M. Waldman, "The
Role of Multiple Potential Entrants/Sequential Entry in Noncooperative Entry Deterrence," 22 RAND
Journal of Economics 446 (1991).
14

that these leased slots will be used relatively less intensively than slots held and operated by the dominant
carriers.
The empirical analyses contained in the 1991 Comments were largely inconsistent with these two
anticompetitive theories. Slot usage for all carriers at the four HDTAs tended to be very high-typically
above 90 percent-and the data indicated no relationship between market share and slot usage. That is,
the negative relationship between market share and slot usage predicted by the anticompetitive hypothesis
did not emerge from actual slot usage data. Further, rather than being net lessors of slots, the largest
firms at each of the four HDTAs were consistently net lessees of slots. This result, together with the high
degree of slot usage, is inconsistent with the implications of the anticompetitive theories.
Modelling slot usage as a function of multiple factors, the 1991 Comments also used multiple
regression analysis to test the anticompetitive hypothesis. 42 At 0 'Hare, the most concentrated of the four
HDTAs, regression analysis indicated a positive and statistically significant relationship between slot share
and rate of use. Further, slots leased by large carriers to others were used at a higher rate than large
carriers' owned and operated slots. Both of these results are inconsistent with the anticompetitive
hypotheses. 43
42 The dependent variable in these regressions was the number of days the slot was operated in the
two month study period. The explanatory variables were: (1) the slot holder's share of slots at the HDTA;
(2) the slot holder's national market share; (3) the FAA withdrawal priority number; (4) a dummy
variable set equal to one if the slot was operated by a single operator other than the slot holder, and zero
otherwise; (5) a dummy variable set to one if the slot was operated by more than one operator, and was
not traded in the two month study period, and zero otherwise; (6) a dummy variable set equal to one if
the slot holder changed during the two month period, and zero otherwise; and (7) a set of dummy
variables each set equal to one for the particular hour or half-hour period for which the slot existed, and
zero otherwise.
43 During the period covered by the 1991 Comments (May-June 1990), United Airlines controlled 658
slots at O'Hare (44 percent). Later, United obtained more slots at O'Hare when it acquired Air
Wisconsin, which had controlled 44 slots (3 percent). This acquisition prompted a private antitrust suit
by American Airlines, which had controlled 491 slots (33 percent). To settle the case, United sold 12 of
the Air Wisconsin slots to American. Since Air Wisconsin already had a close operating relationship with
United and since Air Wisconsin's market share had been relatively modest, the likelihood that these
(continued ... )
15

...
At National and LaGuardia, the 1991 Comments found no discernible relationship between market
share and usage rate in May-June 1990. The relationship between leasing and slot usage was also
inconsistent with the anticompetitive leasing hypothesis. An exception was Kennedy, where the 1991
Comments found that firms with larger market shares tended to show lower slot usage. This result would
be consistent with the anticompetitive usage hypothesis: however, it could also be explained by special
circumstances at Kennedy. The largest slot holder at Kennedy, with 31 percent of the slots, was a bank
that had entered a purchase-lease arrangement with TWA. The second largest slot holder was Pan Am.
Both TWA and Pan Am had been experiencing financial difficulties on transatlantic flights, which
ultimately led to bankruptcy.44 Thus, the negative coefficients on market share may have simply
reflected these largest carriers' reduced output due to operating difficulties, rather than a monopolistic
output reduction.
For O'Hare, National, and LaGuardia, 	the May-June 1990 slot usage data are generally
inconsistent with hypotheses that the HDR allows carriers to exert market power. The only result that
might support the anticompetitive hypotheses, the negative and significant relationship between slot share
and slot use at Kennedy, is probably explained by the operating problems experienced by the largest
carriers at that airport during the study period. In addition, the data did not suggest that slots leased by
43( ... continued)
transactions had a significant impact on competition at O'Hare appears small, though this is an empirical
matter. Section VI below analyzes more recent slot data to assess whether the changes in shares of slots
controlled at O'Hare and the other HDTAs have had any effect on competitive behavior there.
44 Both carriers faced problems at various times due to fears of terrorism and increased competition
on transatlantic flights. See "Secretary's Task Force on Competition in the U.S. Domestic Airline
Industry: International Air Service," U.S. Department of Transportation, February 1990. International
flights use restricted international slots that are not allocated in the same way as slots for domestic flights,
and hence were not in the data used in this study. Passengers from these flights were an important source
of traffic for these airlines' domestic flights, though, and slots used for those domestic flights were
included in the data.
16

large carriers to others were used at a significantly lower rate, but suggested instead the opposite, that
the largest slot holders at HDTAs tended to be net lessees.
VI. 	Slot Usage at the HOT As in 1993
A. 	Slot Usage Rates and Leasing Patterns
In this section, we reexamine the relationship between market shares and slot usage using slot
usage data for the period September-October 1993. 45 Since the May-June 1990 period analyzed in our
1991 Comments, patterns of slot usage at the HDTAs may have been affected by the bankruptcies of a
number of different carriers or by other factors. 46 As in our 1991 Comments, the statistical analysis is
based on the set of slots meeting three criteria: (1) the slots were operated by a domestic airline on a
domestic route; (2) use of the slots was not restricted by the FAA; and (3) the slots were not exempt from
the FAA's "use or lose" rules.
So.
45 Ideally we would have preferred to examine May-June 1993 slot usage data for comparison to the
results based on May-June 1990 data reported in our 1991 comments. Unfortunately, this data had already
been archived and was, therefore, not readily available when we began work on this comment. The slot
usage data that was available to us was from the September-October 1993 period and the November-
December 1993 period. We used the September-{)ctober data for this study because we felt that it would
represent more typical or average slot usage than would the November-December data, which would
include slot usage during the Thanksgiving and Christmas holiday periods.
46 Carriers that held slots at one or more HDTA and filed for bankruptcy protection since June 1990
include Continental, Pan Am, Midway, America West, and TWA. Midway, Pan Am, and Eastern, which
was still operating while in bankruptcy proceedings during. the May-June 1990 period, have since been
liquidated.
17

Table 147 lists the slot holdings and the Herfindahl-Hirschmann Index (HHI) of concentration
at the four HOT As as of October 31, 1993. 48 All remain either moderately or highly concentrated under
the standards contained in the Department of Justice and Federal Trade Commission Horizontal Merger
Guidelines;49 however, the level of concentration has risen considerably at all four airports. Both
LaGuardia and National were moderately concentrated as of October 1993, just as they were in June
1990. LaGuardia's 1993 HHI of 1747 exceeds its 1990 HHI (1241) by 506, and National's HHI of 1657
exceeds its 1990 HHI (1282) by 375. Kennedy and O'Hare were highly concentrated as before;
Kennedy's 1993 HHI of 2233 exceeds its 1990 HHI (1816) by 417, and O'Hare's 1993 HHI of 3731
exceeds its 1990 HHI (3122) by 609.
47 The tables discussed in this section are generally numbered so as to correspond to similar tables
in the 1991 Comments. That is, Table 1 is comparable to Table 1 in the 1991 Comments, Table 2 is
comparable to Table 2 in the 1991 Comments, and so on. Tables that do not correspond to tables found
in the 1991 Comments are denoted by letters following the number of the preceding table (e.g., Table
lA, Table 2A, and so on).
48 In Table 1, slots held by American Airlines include slots held by AMR Slot Holdings (FAA code
(033), a subsidiary of American's parent firm, AMR. USAir's slots include slots held by USAir Shuttle,
Inc. (FAA code USS), a subsidiary of Shuttle, Inc. USAir manages and operates the USAir Shuttle for
Shuttle Inc. Finally, United's slots include slots held by Air Wisconsin Inc. (FAA code AWl), which was
purchased by United in 1992.
49 The Department of Justice and Federal Trade Commission jointly released horizontal merger
guidelines on April 2, 1992. Under the Guidelines, markets are "moderately concentrated" if their HHls
are between 1000 and 1800 and "highly concentrated" if their HHls exceed 1800.
18

TABLE I
DOMESTIC SLOT HOLDINGS AND CONCENTRATION
O'HARE 	NATIONAL
(ORn) 	(DCA)
HOLDER 	CODE NUMBER 	SHARE 	NUMBER 	SHARE
OF SLOTS 	OF SLOTS
FAA 	0001 	0 	0 	0 	0
SHAWMUT BANK N.A. 	0004 	0 	0 	32 	5.9
STATE ST. BANK & TRUST. 	0020 	0 	0 	0 	0
WILMINGTON TRUST CO. 	0023 	14 	1.0 	0 	0
BT COMMERCIAL CORP 	0026 	5 	0.4 	0 	0
CITIBANK 	0030 	0 	0 	26 	4.8
FIRST BANK NATIONAL ASSOC. 	0034 	39 	2.8 	0 	0
CONTINENTAL SLOT TRUST 	0035 	33 	2.3 	43 	8.0
AMERICAN AIRLINES 	AAL 	533 	37.7 	67 	12.4
DELTA AIRLINES 	DAL 	57 	4.0 	88 	16.3
FEDERAL EXPRESS 	FOX 	5 	0.4 	0 	0
BUSINESS EXPRESS 	GAA 	0 	0 	0 	0
MIDWEST EXPRESS 	MEP 	0 	0 	I 	0.2
NORTHWEST AIRLINES 	NWA 	19 	1.3 	87 	16.1
TRANS WORLD AIRLINES 	TWA 	0 	0 	I 	0.2
UNITED AIRLINES 	UAL 	674 	47.7 	42 	7.8
USAIR 	USA 	35 	2.5 	152 	28.2
TOTAL 	1414 	539
HHI 	3731 	1657
19

TABLE I CONTINUED
KENNEDY 	LAGUARDIA 	ALL HDTAs
(JFK) 	(LGA)
NUMBER 	SHARE 	NUMBER 	SHARE 	NUMBER 	SHARE
HOLDER 	CODE 	OF 	OF 	OF
SLOTS 	SLOTS 	SLOTS
FAA 	0001 	0 	0 	13 	2.5 	13 	0.5
SHA WMUT BANK 	0004 	44 	39.3 	20 	3.9 	96 	3.7
STATE ST BANK 	0020 	3 	2.7 	0 	0 	3 	0.1
& TRUST
WILMINGTON 	0023 	0 	0 	0 	0 	14 	0.5
TRUST
BT COM'L CORP 	0026 	I 	0.9 	0 	0 	6 	0.2
CITIBANK 	0030 	0 	0 	53 	10.4 	79 	3.1
FIRST BANK 	0034 	0 	0 	0 	0 	39 	1.5
NATIONAL
ASSOC.
CONTINENTAL 	0035 	0 	0 	23 	4.5 	99 	3.8
SLOT TRUST
AMERICAN 	AAL 	22 	19.6 	67 	13.1 	689 	26.7
AIRLINES
DELTA AIRLINES 	DAL 	12 	10.7 	112 	21.9 	269 	lOA
FEDERAL 	FDX 	0 	0 	0 	0 	5 	0.2
EXPRESS
BUSINESS 	GAA 	1 	0.9 	0 	0 	I 	0.0
EXPRESS
MIDWEST 	MEP 	0 	0 	0 	0 	I 	0.0
EXPRESS
NORTHWEST 	NWA 	9 	8.0 	41 	8.0 	156 	6.1
AIRLINES
TRANS WORLD 	TWA 	4 	3.6 	0 	0 	5 	0.2
AIRLINES
UNITED 	UAL 	7 	6.3 	35 	6.8 	758 	29.4
AIRLINES
USAIR 	USA 	9 	8.0 	148 	28.9 	344 	13.3
TOTAL 	112 	512 	2577
HHI 	2233 	1747
20

One factor that makes the analysis of the 1993 slot usage data somewhat more complicated than
the analysis of the 1990 slot usage is the increased prevalence of noncarrier slot holders. The growth in
noncarrier holders of slots since 1990 is due, in large part. to the severe financial difficulties faced by
many carriers in recent years. In the May-June 1990 period. only one noncarrier institution, the Shawmut
Bank, held any slots. The Shawmut Bank held slots in trust that were formerly held by TWA, and
Shawmut leased these slots back to TWA on a long-term basis.
Table I indicates that six financial institutions in addition to the Shawmut Bank held slots during
the September-October 1993 period. The slots held by these institutions as part ·of bankruptcy proceedings
were generally leased back to the airline that previously held the slots. Other slots held by noncarriers
represented collateral for secured obligations and, again, these slots were generally leased back to the
airline that previously held the slots. Because slots held by noncarriers were typically leased back to
carriers on a long term basis, we have recalculated slot market shares and concentration under the
assumption that slots held by noncarriers should, in most cases, be treated as being held by the relevant
carrier. 50 Table lA contains the results of this analysis. This adjustment raises further the HHI at each
of the HOTAs, moving each into the "highly concentrated" range as defined in the 1992 Merger
Guidelines. In our analyses that follow, we examine whether, in these conditions of higher concentration,
carriers utilized their slots in an anticompetitive manner.
so Generally the carriers transferring slots to trusts have received long-tenn leases on the transferred
slots. Continental, for example, has ten-year leases on the slots held by the Continental Slot Trust; the
USAir Shuttle has ten year leases on the slots held by the Shuttle Slot Trust (Citibank); and Business
Express has ten-year leases on slots transferred to the State Street Bank. BT Commercial Corp., a
subsidiary of Banker's Trust, held slots originally held by America West and transferred as collateral for
secured obligations. First Bank National Association holds slots originally held by Northwest Airlines and
transferred as collateral for certain secured obligations. The one exception to this general rule is that the
slots of the Wilmington Trust are not allocated to any carrier because many of its slots were transferred
from Pan Am, which no longer exists. Also, we have kept separate the slots held by the FAA and Federal
Express.
21

TABLE IA
DOMESTIC SLOT HOLDINGS AND CONCENTRATION
WlTH NONCARRIER SLOTS ASSIGNED TO CARRIERS
O'HARE 	NATIONAL
(ORD) 	(DCA)
HOLDER 	CODE NUMBER 	SHARE 	NUMBER 	SHARE
OF SLOTS 	OF SLOTS
FAA 	0001 	0 	0 	0 	0
WILMINGTON TRUST CO. 	0023 	14 	1.0 	0 	0
AMERICAN AIRLINES 	AAL 	533 	37.7 	67 	12.4
AMERICA WEST AIRLINES 	AMW 	5 	0.4 	0 	0
CONTINENTAL AIRLINES 	COA 	33 	2.3 	43 	8.0
DELTA AIRLINES 	DAL 	57 	4.0 	88 	16.3
FEDERAL EXPRESS 	FDX 	5 	0.4 	0 	0
BUSINESS EXPRESS 	GAA 	0 	0 	0 	0
MIDWEST EXPRESS 	MEP 	0 	0 	1 	0.2
NORTHWEST AIRLINES 	NWA 	58 	4.1 	87 	16.1
TRANS WORLD AIRLINES 	TWA 	0 	0 	33 	6.1
UNITED AIRLINES 	VAL 	674 	47.7 	42 	7.8
USAIR 	USA 	35 	2.5 	178 	33.0
TOTAL 	1414 	539
HHI 	3741 	1930
22

TABLE IA CONTINUED
KENNEDY 	LAGUARDIA 	ALL HDTAs
(JFK) 	(LGA)
NUMBER 	SHARE 	NUMBER 	SHARE 	NUMBER 	SHARE
HOLDER 	CODE 	OF 	OF 	OF
SLOTS 	SLOTS 	SLOTS
FAA 	0001 	0 	0 	13 	2.5 	13 	0.5
WILMINGTON 	0023 	0 	0 	0 	0 	14 	0.5
TRUST
AMERICAN 	AAL 	12 	19.6 	67 	13.1 	689 	26.7
AIRLINES
AMERICA WEST 	AMW 	I 	0.9 	0 	0 	6 	0.2
AIRLINES
CONTINENTAL 	COA 	0 	0 	23 	4.5 	99 	3.8
AIRLINES
DELTA AIRLINES 	DAL 	12 	10.7 	112 	21.9 	269 	10.4
FEDERAL 	FDX 	0 	0 	0 	0 	5 	0.2
EXPRESS
BUSINESS 	GAA 	4 	3.6 	0 	0 	4 	0.2
EXPRESS
MIDWEST 	MEP 	0 	0 	0 	0 	1 	0.0
EXPRESS
NORTHWEST 	NWA 	9 	8.0 	41 	8.0 	195 	7.6
AIRLINES
'.
TRANS WORLD 	TWA 	48 	42.9 	20 	3.9 	101 	3.9
AIRLINES
UNITED 	UAL 	7 	6.3 	35 	6.8 	758 	29.4
AIRLINES
USAIR 	USA 	9 	8.0 	201 	39.3 	423 	16.4
TOTAL 	112 	512 	2577
HHI 	2520 	2347
23

Table 2 examines the hypothesis that carriers with large slot holdings will attempt to restrict output
by reducing the number of days on which those slots are used. The table shows for each slot holder and
each HOTA the proportion of total days that the holder's slots were used (TPCT) and the proportion of
weekdays (WPCT) that the slots were used during the September-October 1993 period. 51 As was the case
in the 1991 Comments, slots are typically utilized at very high rates. 52 Virtually all of these slots were
used at rates exceeding 90%, and the highest rate of slot utilization actually occurred at O'Hare, the most
concentrated of the four HDTAs. Further, the pattern of slot use does not support the anticompetitive
hypothesis. As a general matter, slots held by firms with larger slot market shares appear to be used no
less intensively than slots held by firms with smaller shares.
Looking at each airport individually, the lowest rates of total use (TPCT) at O'Hare were for slots
held by the carriers with the smallest shares, and weekday use rates (WPCT) were all very high and
appear unrelated to share. A similar pattern occurred at National: usage rates are typically very high,
especially on weekdays, and slot usage appears unrelated to share. At LaGuardia, neither the values of
TPCT nor the values of WPCT suggest a relationship between slot use and the number of slots held. At
Kennedy, the holders with the fewest slots have the highest values of both TPCT and WPCT; however,
these firms hold only one slot each, so it is difficult to infer much from this result. Among holders with
multiple slots at Kennedy, no particular pattern relating slot usage to the number of slots held emerges.
SI The distinction between total and weekday use of slots was an issue pertinent to the proposed
regulations that gave rise to our 1991 Comments. The results reported here continue to distinguish
between total and weekday slot usage in order to facilitate comparisons with the slot usage patterns
previously reported.
S2 A comparison of Table 2 in this comment with Table 2 in our 1991 Comments indicates that slot
usage rates were slightly higher in the September-October 1993 period than they were in the May-June
1990 period.
24

TABLE 2
AVERAGE SLOT USE
O'HARE
NUMBER 	TPCT 	WPCT
SLOT HOLDER 	CODE 	OF SLOTS
HELD
UNITED AIRLINES 	UAL 	674 	.95 	.98
AMERlCAN AIRLINES 	AAL 	533 	.96 	.98
DELTA AIRLINES 	DAL 	57 	.97 	.99
FIRST BANK NATIONAL ASSOC. 	0034 	39 	.93 	.96
USAIR 	USA 	35 	.97 	.98
CONTINENTAL SLOT TRUST 	0035 	33 	.95 	.98
NORTHWEST AIRLINES 	NWA 	19 	.95 	.97
WILMINGTON TRUST CO. 	0023 	14 	.97 	.98
BT COMMERCIAL CORP 	0026 	5 	.91 	.94
FEDERAL EXPRESS 	FDX 	5 	.91 	.98
NATIONAL
SLOT HOLDER 	CODE 	NUMBER 	TPCT 	WPCT
OF SLOTS
HELD
USAIR 	USA 	152 	.91 	.97
DELTA AIRLINES 	DAL 	88 	.95 	.97
-.
NORTHWEST AIRLINES 	NWA 	87 	.92 	.96
AMERlCAN AIRLINES 	AAL 	67 	.94 	.97
CONTINENTAL SLOT TRUST 	0035 	43 	.90 	.96
UNITED AIRLINES 	UAL 	42 	.92 	.96
SHAWMUT BANK 	0004 	32 	.95 	.98
CITIBANK 	0030 	26 	1.00 	1.00
MIDWEST EXPRESS 	MEP 	1 	.80 	.98
TRANS WORLD AIRLINES 	TWA 	1 	.89 	.98
25

TABLE 2 CONTINUED
LAGUARDIA
SLOT HOLDER 	CODE 	NUMBER 	TPCT 	WPCT
OF SLOTS
HELD
USAIR 	USA 	148 	.93 	.98
DELTA AIRLINES 	DAL 	\12 	.95 	.98
AMERICAN AIRLINES 	AAL 	67 	.93 	.98
CITIBANK 	0030 	53 	1.00 	1.00
NORTHWEST AIRLINES 	NWA 	41 	.91 	.95
UNITED AIRLINES 	UAL 	35 	.93 	.98
CONTINENTAL SLOT TRUST 	0035 	23 	.91 	.97
SHA WMUT BANK 	0004 	20 	.95 	.97
FAA 	0001 	\3 	.94 	.98
KENNEDY
SLOT HOLDER 	CODE 	NUMBER 	TPCT 	WPCf
OF SLOTS
HELD
SHAWMUT BANK 	0004 	44 	.93 	.95
AMERICAN AIRLINES 	AAL 	22 	.97 	.98
DELTA AIRLINES 	DAL 	12 	.93 	.92
NORTHWEST AIRLINES 	NWA 	9 	.95 	.96
USAIR 	USA 	9 	.96 	.95
UNITED AIRLINES 	UAL 	7 	.94 	.93
TRANS WORLD AIRLINES 	TWA 	4 	.98 	.98
STATE ST BANK & TRUST 	0020 	3 	.92 	.94
BT COMMERCIAL CORP 	0026 	1 	.98 	1.00
BUSINESS EXPRESS 	GAA 	1 	1.00 	1.00
26

As in our 1991 Comments. our primary interest is to test the implications of hypotheses predicting
anticompetitive behavior of carriers controlling large shares of slots at the HOTAs. Table 2A contains average
slot use for the slots operated by each carrier. Particularly notable is the generally high level of slot usage. Only
one carrier (Business Express at National) operated its slots less than 90% of the time, and weekday usage is
commonly as high as 98%. As in Table 2, there does not appear to be an obvious pattern supporting the
anticompetitive hypothesis. Slot usage rates among the carriers operating at each of the HOTAs does not appear
to be inversely related to the number of slots operated by each carrier.
27

TABLE 2A
AVERAGE SLOT USE BY OPERA TOR
O'HARE
NUMBER 	TPCT 	WPCT
CARRIER 	CODE 	OF SLOTS
OPERATED
UNITED AIRLINES 	UAL 	712 	.95 	.98
AMERICAN AIRLINES 	AAL 	540 	.96 	.98
DELTA AIRLINES 	DAL 	48 	.97 	.99
NORTHWEST AIRLINES 	NWA 	45 	.93 	.96
USAIR 	USA 	30 	.98 	.98
CONTINENTAL AIRLINES 	COA 	30 	.94 	.98
TRANS WORLD AIRLINES 	TWA 	6 	.99 	1.00
SIMMONS AVIA nON 	SYM 	3 	.98 	.98
NATIONAL
CARRIER 	CODE 	NUMBER 	TPCT 	WPCT
OF SLOTS
OPERATED
USAIR 	USA 	211 	.93 	.98
DELTA AIRLINES 	DAL 	86 	.95 	.98
AMERICAN AIRLINES 	AAL 	68 	.94 	.97
NORTHWEST AIRLINES 	NWA 	56 	.92 	.96
CONTINENTAL AIRLINES 	COA 	49 	.90 	.96
UNITED AIRLINES 	UAL 	31 	.94 	.97
TRANS WORLD AIRLINES 	TWA 	31 	.93 	.97
BUSINESS EXPRESS 	GAA 	7 	.88 	.92
28

TABLE 2A CONTINUED
AVERAGE SLOT USE BY OPERATOR
LAGUARDIA
CARRIER 	CODE 	NUMBER 	TPCT 	WPCT
OF SLOTS
OPERATED
USAIR 	USA 	197 	.94 	.98
DELTA AIRLINES 	DAL 	112 	.96 	.98
AMERICAN AIRLINES 	AAL 	67 	.93 	.98
UNITED AIRLINES 	UAL 	37 	.93 	.97
NORTHWEST AIRLINES 	NWA 	34 	.92 	.95
CONTINENTAL AIRLINES 	COA 	26 	.91 	.96
TRANS WORLD AIRLINES 	TWA 	19 	.97 	.99
PENN COMM 	PCA 	11 	.94 	.96
BUSINESS EXPRESS 	GAA 	7 	.92 	.96
TRANS WORLD EXPRESS 	RBD 	1 	.95 	.95
NORTHEAST EXPRESS 	NEE 	1 	.92 	.95
KENNEDY
CARRIER 	CODE 	NUMBER 	TPCT 	WPCT
OF SLOTS
OPERATED
TRANS WORLD AIRLINES 	TWA 	43 	.93 	.95
AMERICAN AIRLINES 	AAL 	22 	.98 	.99
DELTA AIRLINES 	DAL 	12 	.93 	.92
TRANS WORLD EXPRESS 	RBD 	10 	.94 	.94
BUSINESS EXPRESS 	GAA 	9 	.96 	.97
USAIR 	USA 	7 	.96 	.95
UNITED AIRLINES 	UAL 	6 	.93 	.92
CARNIVAL AIR LINES 	CAA 	I 	.98 	1.00
NORTHEAST EXPRESS 	NE 	1 	.92 	.93
PENN COMM AIRLINE 	PCA 	1 	.93 	.93
29

Table 3 compares the usage rates of slots held and operated by each carrier with slots leased by the carrier
to other carriers. 53 The anticompetitive leasing hypothesis suggests that slot holders' owned and operated (0&0)
slots will be used at a higher rate than the slots leased to other firms. Accordingly a positive difference between
the usage rates of 0&0 slots and leased slots would be consistent with this h.\'pothesis. Table 3 lists for each
HDTA (I) the number of 0&0 slots for each carrier that owns and operates at least one slot (2) the number of
slots that each carrier leased to other carriers; (3) the total and weekday usage rates for the 0&0 slots and the
leased slots; (4) the difference (M between the 0&0 usage rate and the leased usage rate for total days and
weekdays; and (5) two measures of the statistical significance of this difference, tu and teo tu is at-statistic
calculated under the assumption that the variances of 0&0 and leased usage rates are different. te is at-statistic
calculated under the assumption that the two distributions have the same variance.
S3 For purposes of Table 3, a carrier's "held" slots include those leased to it from noncarrier slot
holders (Le., financial institutions) on a long term basis and held in trust by noncarrier slot holders as
collateral for secured obligations. See footnote 50.
30

TABLE 3
RATES OF USE IN OWNED AND OPERATED VERSUS LEASED SLOTS
O'HARE
OWNED AND OPERATED 	LEASED TO OTHERS
SLOT HOLDER 	CODE 	N 	TPCT 	WPCT 	N 	TPCT 	WPCT 	.1 	tu 	te 	.1 	tu 	te
TPCT0&:0 - TPCTL 	WPCT0&:0 - WPCTI.
AMERICAN AIRLINES 	AAL 	459 	.956 	.980 	74 	.957 	.977 	-.0004 	-.09 	-.1 I 	.003 	1.00 	1.07
CONTINENTAL AIRLINES 	COA 	13 	.951 	.977 	20 	.956 	.978 	-.0049 	-.28 	-.32 	-.0012 	-.13 	-.15
DELTA AIRLINES 	DAL 	33 	.969 	.992 	24 	.961 	.984 	.0076 	.70 	.69 	.009 	1.84 	1.88
NORTHWEST AIRLINES 	NWA 	32 	.928 	.962 	26 	.948 	.972 	-.021 	-1.39 	-1.34 	-.0010 	-.99 	-1.01
UNITED AIRLINES 	UAL 	613 	.952 	.977 	61 	.956 	.979 	-.004 	-.54 	-.70 	-.001 	-.32 	-.36
USAIR 	USA 	19 	.981 	.982 	16 	.949 	.980 	.0322 	2.93' 	3.04' 	.002 	.26 	.25
• Statistically significant at the .05 level.
31

"•
TABLE 3 CONTINUED
RATES OF USE IN OWNED AND OPERATED VERSUS LEASED SLOTS
NATIONAL
OWNED AND OPERATED 	LEASED TO OTHERS
SLOT HOLDER 	CODE 	N 	TPCT 	WPCT 	N 	TPCT 	WPCT 	6 	tu 	te 	6 	tu 	te
TPCT0&0 • TPCTL 	WPCT0&0 - WPCT"
AMERICAN AIRLINES 	AAL 	55 	.941 	.976 	12 	.929 	.961 	.0120 	.61 	.77 	.015 	1.36 	1.30
CONTINENTAL AIRLINES 	COA 	32 	.888 	.980 	\I 	.940 	.928 	-.052 	-2.87" 	-2.41" 	-.023 	-2.56" 	-2.1·"
DELTA AIRLINES 	DAL 	78 	.951 	.980 	10 	.903 	.928 	.0482 	2.21" 	2.89" 	.052 	2.14 	4.55'
NORTHWEST AIRLINES 	NWA 	35 	.912 	.956 	52 	.922 	.966 	-.010 	-.58 	-.58 	-.010 	-.82 	-.87
TRANS WORLD AIRLINES 	TWA 	19 	.951 	.979 	14 	.950 	.975 	.0012 	.07 	.06 	.004 	.27 	.30
UNITED AIRLINES 	UAL 	20 	.931 	.966 	22 	.918 	.961 	.013 	.82 	.80 	.005 	.44 	.43
USAIR 	USA 	162 	.924 	.979 	16 	.943 	.967 	-.018 	-1.\ 3 	-0.88 	.013 	1.08 	I. II
• Statistically significant at the .05 level.
32

TABLE 3 CONTINUED
RATES OF USE IN OWNED AND OPERATED VERSUS LEASED SLOTS
LAGUARDIA
OWNED AND OPERATED 	LEASED TO OTHERS
SLOT HOLDER 	CODE 	N 	TPCT 	WPCT 	N 	TPCT 	WPCT 	Ii 	tu 	te 	Ii 	tu 	te
TPCT0&:0 - TPCTl 	WPCT0&0 - WPCTL
AMERICAN AIRLINES 	AAL 	38 	.931 	.987 	29 	.925 	.974 	.0057 	.38 	.37 	.013 	2.41' 	2.36'
CONTINENTAL 	COA 	12 	.904 	.961 	11 	.912 	.970 	-.008 	-.25 	-.25 	-.009 	-.47 	-.48
DELTA AIRLrNES 	DAL 	83 	.961 	.981 	29 	.921 	.971 	.0404 	3.05' 	3.87' 	.010 	1.49 	1.61
NORTHWEST AIRLINES 	NWA 	14 	.906 	.939 	27 	.905 	.952 	.0016 	.07 	.07 	-.013 	-.57 	-.65
TRANS WORLD 	TWA 	6 	.967 	.977 	14 	.940 	.967 	.0269 	1.18 	1.06 	.0100 	.56 	.55
UNITED AIRLINES 	UAL 	16 	.940 	.974 	19 	.921 	.977 	.0189 	1.17 	1.14 	-.003 	-.29 	-.28
USAIR 	USA 	161 	.948 	.985 	40 	.950 	.970 	-.002 	-.18 	-.14 	.014 	2.56' 	2.82'
• Statistically significant at the .05 level.
33

TABLE 3 CONTINUED
RATES OF USE IN OWNED AND OPERA TED VERSUS LEASED SLOTS
KENNEDY
OWNED AND OPERA TED 	LEASED TO OTHERS
SLOT HOLDER 	CODE 	N 	TPCT 	WPCT 	N 	TPCT 	WPCT 	~ 	tu 	te 	~ 	tu 	te
TPCT0&0 - TPCT l 	WPCT0&0 - WPCTI.
AMERICAN AIRLINES 	AAL 	17 	.975 	.988 	5 	.974 	.967 	.0012 	.06 	.06 	.0202 	1.13 	1.74
DELTA AIRLINES 	DAL 	12 	.929 	922 	0
BUSINESS EXPRESS 	GAA 	4 	.943 	.946 	0
TRANS WORLD AIRLINES 	TWA 	35 	.930 	.948 	13 	.948 	.948 	-.019 	-1.32 	-1.21 	.0001 	.00 	.00
UNITED AIRLINES 	UAL 	6 	.932 	.919 	I 	.984 	.977 	-.052 	-1.27 	-.058 	-.92
USAIR 	USA 	7 	.96 	.95 	2 	.959 	.965 	-.001 	-.04 	-.04 	-.015 	-.37 	-.36
• Statistically significant at the .05 level.
34

At O'Hare, the two largest airlines (in terms of slot share), United (47.7% share) and American (37.7%
share) used their 0&0 slots just as intensively as other carriers utilized slots that these two carriers leased to them.
The differences are very small and in all cases statistically insignificant. In one case - 	USAir in the all days
analysis - 	the difference between a carrier's usage of its 0&0 slots and lessees' usage of its leased slots is
positive and significant. Although this finding is consistent with the anticompetitive hypothesis, USAir's modest
share at O'Hare (2.5%) indicates that its practices would be unlikely to raise competitive concerns.
At National, one carrier, Delta, leased slots to carriers that used the slots a smaller percentage of the time
than it used its 0&0 slots. Delta was the second largest carrier at National (16.3% share). By contrast, for the
largest carrier at National, USAir (28.2% share), there is essentially no difference between its usage of its 0&0
slots and lessees' usage of its leased slots (the differences are either very small or negative, and not statistically
different from zero at the five percent level of significance.) Thus, the carrier with the largest share of slots at
National does not lease its slots in a manner consistent with the anticompetitive hypothesis.54
The findings at LaGuardia are somewhat more problematic. USAir also is the largest carrier at LaGuardia
(28.9% share of slots). Based on total-days usage, USAir leased slots to carriers that used the slots just as
intensively as USAir used its 0&0 slots; however, based on weekdays-only usage, USAir leased slots to carriers
that used the slots significantly less often than USAir used its 0&0 slots. This result is consistent with the
anticompetitive leasing hypothesis. Delta, the second largest carrier at LaGuardia (21.9% share), and American,
the third largest carrier at laGuardia (13.1 %), also appear to lease slots to carriers that use the slots less often than
Delta or American use their 0&0 slots. Thus, at LaGuardia some leasing patterns appear consistent with the
anticompetitive leasing hypothesis. 55 At Kennedy, the differences in slot use of 0&0 slots and leased slots are
all statistically insignificant.
54 In our 1991 Comments, we found that USAir leased its slots at National in a manner consistent with the
anticompetitive leasing hypothesis.
55 In our 1991 Comments, we found that leasing patterns at LaGuardia were not consistent with the
anticompetitive leasing hypothesis.
35

Overall, the patterns of use of 0&0 slots and leased slots are somewhat mixed with regard to the
anticompetitive leasing hypothesis. At O'Hare, the most highly concentrated of the four HDTAs, and at Kennedy,
the differences in slot usage are inconsistent with anticompetitive leasing hypothesis. At National and LaGuardia,
the competitive implications of the differences in 0&0 and leased slot usage rates are less clear. At National, the
largest carrier, USAir, appears to lease slots to carriers that use them at least as intensively as USAir uses its 0&0
slots. This finding is contrary to the predictions of the anticompetitive leasing hypothesis. Nonetheless, Delta, the
second largest carrier at National, tends to use its 0&0 slots at a significantly higher rate than the carriers that
lease slots from it. At LaGuardia, where USAir is again the largest carrier, USAir, Delta, and American lease slots
to carriers that use them less intensively than these carriers use their 0&0 slots. This result. which passes standard
tests of statistical significance, is consistent with the anticompetitive leasing hypothesis.
Table 4 lists the net leasing positions of the carriers holding slots at each HDTA in September-October
1993. The anticompetitive leasing hypothesis predicts that carriers with a large share of the slots at an HDTA will
tend to be net lessors. Accordingly, this anticompetitive hypothesis would be supported if the difference between
the number of slots operated and the number of slots held is negative. S6 The results in Table 4 are somewhat
mixed, but do not generally support the anticompetitive hypothesis. At O'Hare and National, the largest carriers
are net lessees-theyoperate more slots than they hold, counter to the prediction of the anticompetitive hypothesis.
At LaGuardia, the largest carrier, USAir, operates four fewer slots than it holds, while the next two largest carriers,
Delta and American, operate the same number of slots that they hold. At Kennedy, because Shawmut Bank's 44
slots are treated as being held by TWA, TWA holds 48 slots and operates 43, making TWA the largest holder at
Kennedy and a net lessor of slots. The next two largest carriers at Kennedy, however, American and Delta, are
not net lessors of slots.
S6 As was the case with Table 3 (see footnote 53), in Table 4 a carrier's "held" slots include those leased to
it from noncarrier slot holders (i.e., financial institutions) on a long term basis and held in trust by noncarrier slot
holders as collateral for secured obligations. See footnote 50 above. The regression analysis in the next section
allocates the slots held by noncarriers in the same manner.
36

TABLE 4
SLOTS OWNED. OPERATED. AND NET LEASING POSITION
O'HARE
SLOT HOLDER 	CODE 	SLOTS 	SLOTS 	SLOTS OPERATED -
HELD 	OPERATED 	SLOTS HELD
UNITED AIRLINES 	UAL 	674 	712 	38
AMERICAN AIRLINES 	AAL 	533 	540 	7
DELTA AIRLINES 	DEL 	57 	48 	-9
USAIR 	USA 	35 	30 	-5
NORTHWEST AIRLINES 	NWA 	38 	45 	-7
FEDERAL EXPRESS 	FDX 	5 	0 	-5
NATIONAL
SLOT HOLDER 	CODE 	SLOTS 	SLOTS 	SLOTS OPERATED -
HELD 	OPERATED 	SLOTS HELD
USAIR 	USA 	178 	211 	33
DELTA AIRLINES 	DAL 	88 	86 	-2
NORTHWEST AIRLINES 	NWA 	87 	56 	-31
AMERICAN AIRLINES 	AAL 	67 	68 	1
UNITED AIRLINES 	UAL 	42 	31 	-II
MIDWEST EXPRESS 	MEP 	1 	0 	-1
TRANS WORLD AIRLINES 	TWA 	33 	31 	-2
37

TABLE 4 CONTINUED
SLOTS OWNED, OPERATED, AND NET LEASING POSITION
LAGUARDIA
SLOT HOLDER 	CODE 	SLOTS 	SLOTS 	SLOTS OPERATED -
HELD 	OPERATED 	SLOTS HELD
USAIR 	USA 	201 	197 	-4
DELTA AIRLINES 	DAL 	112 	112 	0
AMERICAN AIRLINES 	AAL 	67 	67 	0
NORTHWEST AIRLINES 	NWA 	41 	34 	-7
UNITED AIRLINES 	UAL 	35 	37 	2
KENNEDY
OWNED AND OPERATED
SLOT HOLDER 	CODE 	SLOTS 	SLOTS 	SLOTS OPERATED -
HELD 	OPERATED 	SLOTS HELD
TRANS WORLD AIRLINES 	TWA 	48 	43 	-5
AMERICAN AIRLINES 	AAL 	22 	22 	0
DELTA AIRLINES 	DAL 	12 	12 	0
USAIR 	USA 	9 	7 	-2
NORTHWEST AIRLINES 	NWA 	9 	0 	-9
UNITED AIRLINES 	UAL 	7 	6 	-1
BUSINESS EXPRESS 	GAA 	I 	9 	8
38

B. 	Regression Analysis
As in our 1991 Comments, we further examine the anticompetitive hypotheses regarding slots using
multiple regression analysis to analyze the more recent (September-October 1993) slot usage data. The dependent
and independent variables used in these regressions are generally the same as those used in the 1991 Comments. 57
In the regressions, the dependent variable is the number of days a slot was used, NDOPER. The dependent
variable is measured by using either all 61 days in the sample period or only the 43 weekdays in the period.
The independent variables in the regressions consist of (I) SLMS, the slot holder's share of the total
number of nonrestricted and nonexempt slots at the HDTA at the end of the two-month period; (2) NAMS, the
slot holder's national market share (based on Revenue Passenger Miles) for the two month period; (3) WITHD,
the FAA withdrawal priority number; (4) LEASE, a dummy variable equal to one if the slot is operated by a
single operator other than the slot holder and zero otherwise; (5) MULTO, a dummy variable equal to one if the
slot is operated by more than one operator, and was not traded in the two month period and zero otherwise; and
(6) PI' a dummy variable equal to one if the slot's hour or half-hour period starts at time 1. The dummy variable
PT is dropped, where T is the last hour or half-hour period of the day for which slots are allocated at each airport.
The inclusion of SLMS allows us to test the anticompetitive hypothesis that slots held by carriers with a
large share of the slots at the HDTA will be used less intensively than slots held by carriers with smaller shares. 58
57 One variable has been dropped because it is not applicable to the September-October 1993 data. MULTH,
a dummy variable equal to one if the slot holder changed during the two-month period and zero otherwise, was
dropped because no slot was transferred from one holder to another during the September-October 1993 period.
58 The regressions include the interactions of SLMS with LEASE and MULTO; consequently, the coefficient
on SLMS measures changes in NDOPER as SLMS changes only in those cases in which LEASE and MULTO
equal zero. For cases in which LEASE equals I and MULTO equals 0, the effect of a change in SLMS on
NODPER equals the sum of the coefficients on SLMS and LEASE*SLMS; when MULTO equals I and LEASE
equals 0, this effect is the sum of the coefficients on SLMS and MULTO*SLMS; and when both LEASE and
MULTO equal 1, this effect is the sum of the coefficients on SLMS, LEASE*SLMS, and MULTO*SLMS. For
the majority of slots in our sample, both LEASE and MULTO are zero; consequently, our discussion of the effects
of slot market share on the number of days a slot was used (i.e., our dependent variable NDOPER) generally
concerns the coefficient on SLMS. We have, however, calculated the effects of changes in SLMS on NDOPER
for those cases in which either LEASE, MULTO, or both equal 1. Those cases in which these effects are
significant are discussed in footnotes.
39

...
The variables LEASE and LEASE*SLMS (which interacts these two variables) allow us to test the anticompetitive
leasing hypothesis, which predicts that slots leased by large holders of slots to other carriers will be used at lower
rates. The other variables in the regressions control for other factors that might affect usage rates, but which do
not necessarily imply anticompetitive behavior. NAMS controls for differences in slot use resulting from the
degree to which different carriers have more or less extensive route structures. 59 All other things equal, the more
extensive a carrier's route network, the more flexibility the carrier will have to adjust routes and, consequently,
the more efficiently a carrier will be able to utilize slots. The slot withdrawal number, WITHD, controls for
possible differences in the strength of a holder's "property right" associated with a slot and any influence that this
may have on usage rates. The dummy variables PI are included in order to control for differences in slot use
arising from differences in the time of day over which a slot is defined. A slot covering a period of the day in
which air travel tends to be high will, all other things equal, be used at a higher rate than a slot covering a less
busy time of day.
The larger number of noncarrier slot holders in the 1993 data makes the analysis here somewhat more
complicated than in the 1991 Comments. In September-October 1993, as in May-June 1990, the Shawmut Bank
held TWA's former slots in trust and TWA, in tum, had a long-term lease with Shawmut for use of the slots.
Thus, Shawmut's slots were treated as held by TWA in the regression analysis. The slots held by the other
noncarrier holders have been treated similarly.60
The results from the 1993 regressions are reported in Tables 5.1 through 5.4. 61 As explained below, they
do not tend to support the anticompetitive hypotheses.
59 The national market shares for the two slot holders that do not offer carrier se~-yice, Wilmington and Federal
Express, are set to zero.
60 See footnote 50 supra.
6\ For expositional convenience, the coefficients on the time-period dummy variables are not reported.
Estimates of these coefficients are available upon request.
40

TABLE 5.1
REGRESSION RESULTS
O'HARE
DEPENDENT VARIABLE: NUMBER OF DAYS SLOT OPERATED, SEPT - OCT 1993
(t-statistics in parentheses)
VARIABLE 	COEFFICIENTS
ALL DAYS 	WEEKDA YS ONLY
CONSTANT 	52.58 	40.49
(51.94)* 	(70.65)*
SLMS 	-0.91 	-0.52
(- 1.65)* 	(-1.66)*
NAMS 	7.39 	4.66
(3.18)* 	0.54)*
WITHD 	.18 x 10~ 	.77 X 10-4
(0.17) 	( 1.22)
LEASE 	0.24 	0.23
(0.77) 	( 1.25)
LEASE*SLMS 	-0.40 	-0.37
(-0.45) 	(-0.74)
MULTO 	-0.20 	-0.02
(-0.46) 	(-0.07)
MULTO*SLMS 	0.42 	-0.40
(0.37) 	(-0.61)
N 	1414 	1414
MEAN DEP. VAR. 	58.20 	~.	
42.07
R-SQUARED 	.325 	.122
ADJ R-SQUARED 	.308 	.100
F 	18.96* 	5.4842*
LOG LIKELIHOOD 	-2881.62 	-2077.27
* Significant at less than the .05 level (one-tail test)
41

TABLE 5.2
REGRESSION RESULTS
NATIONAL
DEPENDENT VARIABLE: NUMBER OF DAYS SLOT OPERATED, SEPT - OCT 1993
(t-statistics in parentheses)
VARIABLE 	COEFFICIENTS
ALL DAYS 	WEEKDA YS ONLY
CONSTANT 	61.10 	43.64
(24.85)· 	(37.86)·
SLMS 	-0.84 	2.03
(-0.41) 	(2.12)·
NAMS 	6.47 	1.14
( 1.66)* 	(0.62)
WITHD 	-0.48 x lO-l 	-0.12 X lO-l
(-5.35)* 	(-2.83)*
LEASE 	-0.97 	0.37
(-1.02) 	(0.83)
LEASPSLMS 	0.58 	-4.71
(0.11) 	( -1.87)·
MULTO 	-1.30 	-0.21
(-1.31) 	(-0.46)
MULTO·SLMS 	3.91 	-1.11
(0.70) 	(-0.42)
N 	539 	539
MEAN DE? VAR. 	56.65 	41.80
R-SQUARED 	.232 	.144
ADJ R-SQUARED 	.196 	.104
F 	6.48* 	3.61·
LOG LIKELIHOOD 	-1461.69 	-1053.23
* Significant at less than the .05 level (one-tail test)
42

TABLE 5.3
REGRESSION RESULTS
KENNEDY
DEPENDENT VARIABLE: NUMBER OF DAYS SLOT OPERATED, SEPT - OCT 1993
(t-statistics in parentheses)
VARIABLE 	COEFFICIENTS
ALL DAYS 	WEEKDA YS ONLY
CONSTANT 	76.39 	48.99
(4.24)- 	(3.80)-
SLMS 	-0.52 	2.57
(-0.16) 	(1.14)
NAMS 	10.57 	4.81
(1.92)- 	( 1.22)
WITHD 	-0.38 x 10-2 	-0.13 x 10-2
(- 1.08) 	(-0.54)
LEASE 	0.98 	1.67
(0.48) 	(1.14)
LEASE-SLMS 	1.92 	-3.47
(0.32) 	(-0.80)
MULTO 	0.12 	0.13
(0.11) 	(0.16)
MULTO·SLMS 	0.37 	-3.38
(0.10) 	(-1.21 )
N 	112 	112
MEAN DEP. VAR. 	57.70 	40.93
R-SQUARED 	.181 	.174
ADJ R-SQUARED 	.043 	.035
F 	1.31 	1.25
LOG LIKELIHOOD 	-266.24 	-228.68
- Significant at less than the .05 level (one-tail test)
43

TABLE 5.4
REGRESSION RESULTS
LAGUARDIA
DEPENDENT VARIABLE: NUMBER OF DAYS SLOT OPERATED, SEPT - OCT 1993
(t-statistics in parentheses)
VARIABLE 	COEFFICIENTS
ALL DAYS 	WEEKDA YS ONLY
CONSTANT 	63.06 	43.29
(20.72)* 	(37.64)*
SLMS 	2.24 	2.47
( 1.11) 	(3.23)*
NAMS 	-1.35 	4.06
(-0.34) 	(2.71)*
WITHD 	-0.13 x 10.2 	-0.37 x 10';
(-1.92)* 	(-1.42)
LEASE 	-0.57 	0.53
(-0.67) 	(1.65)*
LEASE*SLMS 	-1.22 	-4.50
(-0.35) 	(-3.41)*
MULTO 	-1.07 	-0.23
(-1.25) 	(-0.70)
MULTO·SLMS 	2.34 	0.01
(0.71) 	(0.0l)
N 	512 	512
MEAN DEP. VAR. 	57.30 	42.00
R-SQUARED 	.163 	.179
ADJ R-SQUARED 	.089 	.105
F 	2.18· 	2.43*
LOG LIKELIHOOD 	-1375.75 	-887.44
• Significant at less than the .05 level (one-tail test)
44

1. 	O'Hare
O'Hare is the most concentrated of the four HOT As. The high level of concentration might suggest that
the potential for anticompetitive behavior would be greatest here. [n fact, the two coefficients on SLMS, the slot
holder's market share, are negative and statistically significant at the .05 level using a one-tail test. While these
findings are consistent with the anticompetitive hypotheses. these coefficients are very small. 62 The coefficients
on LEASE and LEASE*SLMS also are statistically insignificant. as are the coefficients on the control variables
except NAMS, national market share. Thus, this regression appears to indicate that slot usage at O'Hare is
primarily a function of the extensiveness of a carrier's route structure.b3
2. 	National
At National, the coefficient on SLMS is negative In the all-days regression. but it is statistically
insignificant. In the weekdays-only regressions. the coefficient on slot market share is both positive and significant,
which is inconsistent with the anticompetitive usage hypothesis. Both coefficients on LEASE are statistically
significant a the .05 level using a one-tail test, and the coefficient on LEASE*SLMS is statistically significant at
the .05 level in the weekdays only regression. For weekday use, the largest slot holder at National, USAir, leased
62 In the ALL OAYS (WEEKDAYS ONLY) regression, the coefficient of -0.91 (-0.52) indicates that an
increase in slot market share of 50 percent would decrease slot usage by 0.46 (0.26) days over a two month
period. In the WEEKDAYS ONLY regression, the sum of the coefficents on SLMS and LEASE*SLMS equals -
.89 (see footnote 58), which is statistically significant at the .05 level using a one-tail test (t-statistic equals -1.85).
This result indicates that an increase in market share of 50% would decrease slot usage by 0.45 days for those
slots that were leased. Approximately 14 percent of the 1414 O'Hare slots included in our sample were leased.
In our 1991 Comments, we found a positive and significant relationship between SLMS and slot usage at O'Hare,
a result inconsistent with the anticompetitive hypothesis.
63 As indicated in Table 2, there is little variation in usage rates across the various airlines holding slots at
O'Hare. Since the two largest carriers at O'Hare are also the two largest carriers nationally, NAMS may be
capturing some of the relationship between slot use and market share. When NAMS is dropped from the
regressions, the coefficient on SLMS falls to -.199 (t-statistic -.39) in ALL OAYS regression and -.072 (t-statistic -
.25) in the WEEKDAYS ONLY regression.
45

slots to carriers who used these slots significantly less intensively than USAir utilized its owned and operated
slots. 64 This finding is consistent with the anticompetitive leasing hypothesis, although the effect is offset by the
positive, significant coefficient on SLMS in the weekdays-only regression.
The coefficient on WlTHD is significant in both the all-days and weekdays-only regressions. In both
regressions, this coefficient is negative, indicating that slots with higher withdrawal numbers (i.e., those least likely
to be withdrawn) are used less intensively-a somewhat surprising result.
3. 	Kennedy
Virtually all coefficients estimated by both regressions are statistically insignificant at standard levels.6s
Accordingly, no relationships between slot shares and slot usage or leasing and slot usage are discemable. In fact,
the F-statistic testing the joint significance of all the explanatory variables is statistically insignificant, which
indicates that the explanatory variables, taken together, cannot explain the observed variation in the dependent
variable.
4. 	LaGuardia
For the all-days regression, the constant term and WITHD are statistically significant at standard levels. 66
The weekdays-only regression appears somewhat more informative. The coefficient on SLMS, slot market share,
is positive and significant, a result that is inconsistent with the anticompetitive hypothesis. On the other hand, the
negative, significant coefficient on LEASE*SLMS indicates that slots leased by larger carriers to other carriers
tend to be used relatively less intensively, a result consistent with the anticompetitive hypothesis (and with the
64 By combining the coefficients on the LEASE and LEASE*SLMS variables with USAir's slot market share
at National, the analysis indicates that the slots that USAir leased to other carriers were used 1.18 fewer days over
this two month period than USAir utilized its owned and operated slots. This difference is statistically significant.
65 In the case of the all-days regression, the coefficient on national market share, NAMS, would be significant
at the .05 level using a one-tail test. This indicates that a carrier with a more extensive national route structure
tend to use slots at Kennedy somewhat more intensively.
66 The negative coefficient on WlTHD suggests that slots less likely to be withdrawn by the FAA (i.e., those
with higher withdrawal numbers) are less highly used-a somewhat counterintuitive result. The F-statistic for the
regression is statistically significant.
46

· .
results reported in Table 3).67 According to this analysis. the two carriers with the largest slot shares at
LaGuardia, USAir and Delta, leased slots to other carriers who utilized them less intensively, other things equal,
than these two carriers utilized their owned and operated slots. 68
5. 	Summary of Regression Results
Generally, the regressions using the September-October 	1993 data do not indicate any consistent
relationships between slot usage and either holders' slot market shares or holders' leasing behavior. The results,
therefore, do not generally support the two anticompetitive hypotheses examined. 69 Many of the findings do not
meet standard tests of statistical significance; in these cases, the hypothesis that the variables to which they apply
had no effect on slot usage one way or the other cannot be rejected. The negative, significant coefficients on slot
market share at O'Hare do support the anticompetitive hypothesis, but would account only for very small effects.
The analyses that provide statistically significant results supporting the anticompetitive leasing hypotheses also
contain offsetting results with regard to slot market share. At National and LaGuardia, the weekdays-only
regressions produce negative, statistically significant results that are consistent with the anticompetitive leasing
hypothesis, but these same analyses contain positive, statistically significant coefficients on market share that are
inconsistent with the anticompetitive slot usage hypothesis.
67 For those cases in which slots were leased (i.e., LEASE equals I), the sum of the coefficients on SLMS
and LEASE*SLMS is -2.02, which is significant at the .05 level using a one-tail test (t-statistic equals -1.86). This
results indicates that an increase in market share of 50% would decrease the usage of leased slots by 1 day. This
result is consistent with the results reported in Table 3. Approximately 22 percent of the 512 LaGuardia slots
included in our sample were leased.
68 The analysis indicates that the slots leased by USAir (Delta) to other carriers were used 1.24 (0.45) fewer
days over this two month period than were the carrier's owned and operated slots. These differences are
statistically significant.
69 These results, of course, depend on the data and the model that was estimated.
47

vn. 	Peak and Off-peak Pricing as an Alternative to the HDR
Some critics of the HDR have suggested that the use of peak and off-peak landing fees may be a preferred
alternative to the HDR.70 In a previous analysis of the use of peak and off-peak aircraft operations fee
structures, 7 1 the staff of the FTC concluded that properly structured takeoff and landing fees can be an important
step toward more efficient pricing of operating rights. 72
The marginal cost of an aircraft operation consists of: (l) the resource costs imposed on the airport; (2)
the delay costs imposed on aircraft operations and passengers, including the delay costs imposed on other flights;
and (3) the noise costs imposed on residents living near the airport. During peak periods, the first two, and
possibly all three, of these marginal costs increase. Increased congestion during peak periods increases delay costs
imposed on passengers and airlines using the airport during these times. 73 The resource costs imposed on the
airport also increase (e.g., additional costs are imposed on the air traffic control system).
70 See, for example, S. Morrison and C. Winston, "The Dynamics of Airline Pricing and Competition," 80
American Economic Review 389 (May 1990). They write, "Because slots limit the effect of competition on fares
and the number of competitors, they should be eliminated and replaced by congestion-based takeoff and landing
fees. Congestion pricing would reduce travel delays efficiently ... and could enhance competition." (p. 392).
... 	71 See Comments of the Bureaus of Economics, Competition, and Consumer Protection of the Federal Trade
Commission, Massport Program for Airport Capacity Efficiency (February 29, 1988).
72 Both peak-load pricing and the HDR have the desirable attribute of being market-based methods of
allocating airport capacity. When resources are allocated by non-market methods, such as through administrative
action, the benefit of the good or service is awarded as an economic rent. Economic theory predicts that private
parties will make expenditures in pursuit of such gain until the gain is dissipated. See. e.g., Posner, "The Social
Costs of Monopoly and Regulation," 83 Journal of Political Economy 807 (1975), for a discussion of the
proposition that pursuit of economic rents converts those rents into social costs. Thus, in order to obtain goods
or services awarded by non-market methods, firms expend real resources-such as for attorneys, lobbyists, and
consultants-until the expected gains are dissipated. These expenditures represent a social cost that is largely
avoided by market-based allocation methods.
73 Examples of these costs include the extra fuel used by aircraft waiting to take off and land and the
additional time wasted by delayed passengers. For a more detailed discussion of this point, see the FTC Comments
on the Massport Program, supra note 71, Appendix A, and Alfred E. Kahn, The Economics of Regulation:
Principles and Institutions, Cambridge: MIT Press (1988), pp. 87-116.
48

Economically efficient prices would cover all of these costs. Because the marginal costs of a landing will
increase during peak periods, the economically efficient price of a landing also will increase during peak periods.
Thus, using a peak/off-peak pricing structure would be consistent with economically efficient pricing.
Although economically efficient pricing will have a peak/off-peak structure, whether or not a regulatory
body or airport administrator would (or could) choose the appropriate levels of peak and off-peak prices both to
avoid congestion and to utilize capacity optimally is not clear. Some critics promote the use of price rather than
quantity, that is, slot regulation, as the appropriate regulatory instrument. The underlying reason for this preference
is not always clear, and the relative merits of a quantity-based versus price-based system are not clear ex ante.
While market prices are superior economizers of information compared to administratively set output levels,
administratively set prices do not necessarily possess the same advantage. 74
The amount of information required to implement peak and off-peak pricing may be formidable. An
airport's capacity is fixed, but the demand for air travel from a particular airport can fluctuate substantially. Not
only does demand fluctuate between peak and off-peak periods during each day, but peak and off-peak demand
can fluctuate from day to day and month to month. The demand for air travel (both at peak and off-peak times)
will fluctuate with changes in the business cycle, seasonally, and as a result of purely random events such as
terrorist threats and the weather. If administratively determined peak and off-peak prices cannot respond quickly
to cyclical and random changes in peak and off-peak demand, then the welfare of air travelers may be reduced.
If prices are set too low, travelers may be faced with congestion and congestion related delays; if prices are set
too high, airport capacity may be under-utilized as the number of operations falls below levels necessary to control
congestion.
The HDR was adopted in order to allocate existing capacity. If that capacity is known and reflected in the
hourly slot quotas, the prices at which slots trade will fluctuate as demand fluctuates: higher prices during high-
74 See F. A. Hayek, "The Use of Knowledge in Society," 35 American Economic Review 519 (1945), and M.
L. Weitzman, "Prices and Quantities," 41 Review of Economic Studies 477 (1974).
49

demand peak periods, and lower prices during low-demand off-peak periods. Thus, given the level of capacity,
market forces will lead naturally to peak and off-peak prices. To implement slot-based regulation, the only
information that regulators or airport administrators need is an estimate of the capacity of the airport. Since airport
capacity does not generally change from day to day or hour to hour, the degree of regulatory oversight necessary
to implement slot-based regulation is modest.
Thus, while using peak/off-peak pricing of takeoff and landing fees would likely enhance efficiency
compared to rationing airport capacity by the queue, it appears likely that such a pricing structure is not superior
to slot-based regulation. The information required to set peak-load prices at levels that adequately discourage
congestion may be considerable given the degree of cyclical and random fluctuations in the demand for air travel.
vm. 	Suggested Areas for Study
To the extent the FAA might be concerned about the effects, if any, of slot market regulation on other
dimensions of competition in the airline industry, the FAA might consider expanding its slot usage data base to
include such information as the size and destination of the airplane using a particular slot, the prices at which
carriers sell slots to one another, and rates at which slots are leased. Such data would allow a more refined analysis
of how slot usage might affect or evidence competition among airlines. 75 If collecting this additional data is
burdensome, the FAA might consider collecting the data for a limited study period or for a limited, representative
sample of the slots. In addition, the FAA might want to consider, as part of its study of the HDR, a reexamination
of the operating capacity at each of the HDTAs. Such a study would help to ensure that slot quotas still accurately
reflect capacity and do not inappropriately constrain airport operations.
The staff of the Bureau of Economics supports the FAA's efforts to encourage the use of market-based
systems to allocate scarce airport resources, including the use of price-based and quantity-based allocation schemes.
75 The FAA may also wish to examine whether different airlines' financial circumstances affected their slot
usage rates.
50

We suggest that the FAA consider under what conditions use of quantity-based regulation systems, such as the
HDR, may be more efficient than price-based regulation systems.
IX. 	Conclusions
The high density rule was created in reaction to congestion and delays at the high density traffic airports.
Although the rule has been criticized as imposing artificial constraints on operations at the four HOTAs, it is the
limited capacity of these airports, not the HOR, that ultimately constrains their operations. The HOR is basically
a mechanism for allocating the limited capacity of the HOTAs to competing carriers. By creating a market for
slots, the HDR helps ensure that slots are allocated to the flights that are valued most highly by consumers. 76
We concluded from our empirical examination of slot usage at the four HOTAs both before and after the
1992 amendments to the HDR were adopted that slot usage was not consistent with the exercise of market power
by incumbent carriers. The 1992 amendments were designed specifically to promote entry by new carriers and
foster the growth of smaller incumbent carriers at the HDTAs. Given the important economic function of the HDR
as a mechanism for efficiently allocating limited capacity, the FAA may wish to consider rescinding one part of
those amendments, the two-year restriction on the sale of slots obtained through a lottery. Further, the FAA may
wish to study the feasibility of extending the HDR to include additional airports regularly prone to congestion and
delays attributed to excess demand at peak time periods. Alternatively, the FAA may wish to consider allowing
airports that are not slot-regulated to implement peak-load pricing of landing and takeoff fees as a means of
allocating scarce capacity more efficiently than by simple rationing by the queue. As we have emphasized in our
76 Markets are being used to allocate scarce resources efficiently in a number of regulatory settings. The 1990
Clean Air Act, for example, allows pollution emission credits to be purchased and sold. By allowing a market for
pollution credits to exist, Congress created incentives that will help insure that air quality standards are met
through the cheapest and most efficient means. As described in the New York Times (October 23, 1990, p. A18),
"The idea in the legislation is that if one company can remove a ton of sulfur dioxide at its plant more cheaply
than another company can do it at another plant, then the ton should be removed in the cheapest place, and the
utility doing the extra work should be able to sell the credit to other utilities, thus reducing emissions at the lowest
possible overall cost." More recently, the Federal Communications Commission has used an auction to allocate
spectrum to firms providing "personal communications services."
51

comments, the HDR is not the problem-rather, the problems are the limited capacity, high demand, and resulting
congestion and congestion-related delays at the HDTAs. The HDR is an attempt to solve these problems in an
economically efficient manner.
52