Innovative Approaches to Addressing Aviation Capacity Issues in Coastal Mega-regions (2010)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

90 5.1 Introduction From the research undertaken to date on this project, it is clear that the scarce resource of capacity is not allocated effi- ciently. Chapter 5 investigates methods in which such capacity could be allocated in a way that balances passenger service from two perspectives: flight frequency and service reliability. The balance of stakeholder roles is explored in this chapter, with the goal of developing approaches that are agreeable to all stake- holders and fit the individual needs of a congested airport. The chapter examines alternatives to the current congestion and demand management structure in which the roles at the fed- eral and local levels are unclear. It reviews a wide variety of can- didate strategies and actions. Chapter 5 further develops several strategies to increase airport throughput capacity, examining the barriers and constraints that impact their implementation. The research explores the idea that more attention should be paid to studies at individual congested airports to prioritize the value of individual flights, based on their contribution to delay and their customer service values (see Exhibit 5.0 for highlights and key themes in Chapter 5). As used in this report, the term “demand management program” is one that limits delays that occur if too many air- craft are scheduled to arrive at an airport during a particular time. Under this use of the term, demand management is not meant to refer to any program specifically designed to decrease the number of air trips made.33 5.2 The Promise of Demand Management: A Case Study The same quantity of air transport payload capacity can be provided with larger numbers of small aircraft flights or smaller numbers of large aircraft flights. It has long been C H A P T E R 5 Airport Demand Management • The research has concluded that the current system suffers from unclear responsibility: no one has the authority and accountability for the management of congestion at mega-region airports. • The management of existing resources could be improved: this chapter builds the case that capacity in the mega-regions can be increased only when the all the major players are empowered to solve the problem. • Opportunities to reduce mega-region airport congestion and improve the overall cost and quality of passenger service exist; what would be beneficial are policies and programs that encourage key decisionmakers to grasp such opportunities. • When the system fails, a trigger mechanism could be set off; with the responsibilities of each party clearly specified, the goals of accountability and transparency could be met. • There are roles for both the national and local levels in defining these roles and procedures. • The responsibility of those in charge is to make air travel reliable for passengers; this is a form of accountability beyond making the airport available for all classes of aeronautical activities. • A way to do this is to focus on the passenger experience. A congested airport does not necessarily make the airport rea- sonably available nor are delays arguably nondiscriminatory from the passenger perspective. Exhibit 5.0. Highlights and key themes included in Chapter 5. 33 See Chapter 2 for a discussion of strategies that are designed to decrease the number of passengers flying.

91 recognized that the decisions of air carriers about what recipe to use have important ramifications for the quality of service and level of accessibility provided by the air transportation system on the one hand and for the amounts of flight traffic, levels of congestion and delay, and infrastructure require- ments on the other. To explore these trade-offs, the research team analyzed June 2008 schedules for several days at one coastal mega-region airport, SFO. The aim was to document and analyze the wide diversity of aircraft sizes contained in the SFO fleet mix in order to identify situations where a different choice of aircraft size could substantially reduce delay with a minimal loss or (taking the reduced delay into account) even an improvement in the level of service provided. 5.2.1 SFO Fleet Mix The research team examined SFO arrivals on four days in June 2008: the 5th, 13th, 18th, and 25th. These days were cho- sen because they feature varying levels of congestion and delay, as measured by on-time performance. Flight information was downloaded from the FAA Aviation System Performance Met- rics (ASPM) database. In theory, the database includes all flights, including air carrier, general aviation, and cargo, that were actually flown. The database does not include cancelled flights. Altogether, there were 2,165 arriving flights on these days, of which 10 were cargo flights. In the fleet-mix analysis, the team focused on the 2,155 passenger flights. The ASPM passenger flight data was supplemented with two other variables: the estimated seats available for passen- gers based on the aircraft type and the great circle distance of the flight route. The seat information is based on U.S. averages obtained from the DOT’s T100 database, when available, and company websites in other cases. Hereafter, the term “aircraft size” is used to mean the number of seats. The average size of an SFO arrival flight over the four selected days was 135 seats, whereas the standard deviation is 80 seats, reflecting the diverse size of the fleet serving SFO. To examine the size distribution in more detail, a cumulative dis- tribution function was constructed (Figure 5.1), which indi- cates, for any given size, the proportion of flights with aircraft at or below that size. On the small end of the distribution, about 3% of the flights are 13 seats or below. These include a smattering of corporate jets. Next there are a sizable number of regional jets, of sizes ranging from 30 to 80 seats. Altogether, aircraft 80 seats and smaller account for 26% of the fleet mix. The biggest portion of the fleet—about 60%—is in the 100–180 seat range. These include the large jet mainstays of the domestic airline fleet, such as the Boeing 737, MD 80 series, and A320 series. Widebodies of 200 seats or more—including Boeing 767s, 777s, and 747s along with Airbus 340s—account for the remaining 14% of the SFO fleet mix. The diverse fleet mix at SFO means that the vast majority of total seats are provided by a relatively small proportion of flights, as shown in Figure 5.2. This figure was constructed by sorting the 2,155 flights from largest to smallest, and then computing the fraction of total seats provided by the cumula- tive sum of the seats with the largest aircraft. To aid with inter- pretation, the aircraft size for each of these flights is also plotted Figure 5.1. Aircraft size distribution, SFO (June 2008).

92 using the secondary vertical axis. Figure 5.2 reveals that 87% of the seat capacity is provided by the 66% of the flights using Airbus 319 (124 seats) or larger aircraft. Conversely, 10% of the SFO flights using the smallest aircraft account for less than 2% of the total seats. One could argue that the “value” of a flight depends on not only its size but also its distance. Longer distance flights gen- erally have higher fares and serve trips of longer duration. Moreover, the time savings from making the trip by air instead of by surface mode is roughly proportional to distance. Figure 5.2 therefore contains a second share curve that is based on seat-miles instead of seats. This curve is generally higher than the seat-share curve. For example, the 40% of the flights flown with the largest aircraft generate 60% of the seats but 66% of the seat-miles. This difference reflects the positive correlation between aircraft size and flight distance. The only exception is for the smallest aircraft in the fleet, bizjets of 15 seats or fewer, on which many of the flights are quite long, which accounts for the sharp up-tick in the seat-mile share curve at the far right of the figure. The relationship between size and distance is shown more directly in Figure 5.3, which plots aircraft size against flight length on a log-log scale. The data in this figure are differen- tiated according to whether the flight was a scheduled flight appearing in the Official Airline Guide. The correlation for the scheduled flights is evident, with the trend-line indicating that aircraft size increases proportionally with the square root of flight distance. No such relationship is evident for the non- scheduled data. The small corporate jet flights have lengths ranging from around 100 to several thousand miles, while the handful of non-scheduled large jet flights—often diversions or ferries—are on average somewhat shorter. Aside from distance, aircraft size is related to segment traf- fic density—the quantity of passenger traffic per unit time. If the density is low, smaller aircraft are needed to attain an acceptable level of flight frequency. As traffic increases, airlines can use larger aircraft, exploiting economies of scale while still maintaining a convenient number of daily flights. Figure 5.4 depicts this phenomenon. Based on the June 18, 2008, SFO arrival schedule, Figure 5.4 summarizes the service provided by individual passenger carriers on individual flight segments in terms of the number of flights (plotted on the hor- izontal axis) and the average seats per flight (plotted on the vertical axis). Different symbols are used to differentiate the segments according to length. Seats per day for a segment, which is directly related to traffic density, is the product of the two coordinates. A series of isoquants indicate combinations of aircraft size and flight frequency that yield the same quantity of seats per day. Segments on which small (<100 seats, in this discussion) aircraft are used have low densities, almost always less than 300 seats/day. Within this set of segments, the key determinant of aircraft size is distance, with smaller Embraers assigned to segments 300 miles or less, larger Canadairs serv- ing the 601- to 1,200-miles segments, and a mixture of the two types employed on the 301- to 600-miles segments. Although all the segments served by small aircraft are low density, not all low-density segments are served by small air- craft. The variability is particularly notable for segments of Figure 5.2. SFO fleet mix profile (June 2008).

300–600 miles. For example, one airline provides 280 seats per day from Portland (a 551-miles segment) with two MD80s, whereas another airline uses five Canadair flights to provide 270 seats from Boise (522 miles). Similarly, an airline flies eight Embraers a day from Medford, a distance of 329 miles, whereas another provides almost as many seats (239 vs. 260) with two 737 flights from Burbank, which is 326 miles away. 5.2.2 Economies of Aircraft Size The fleet-mix behaviors observed in the previous discussion are shaped by two main economic factors: economies of scale in the cost of operating aircraft and the service advantages of higher flight frequency. Cost economies are illustrated in Fig- ure 5.5, which plots aircraft direct operating cost per seat Figure 5.3. Aircraft size versus segment length, SFO arrivals (June 2008). Figure 5.4. Daily seat capacity production, SFO (June 18, 2008). 93

94 against segment distance for two aircraft types, the 144-seat Boeing 737-400 and the 50-seat Embraer 145. Unit costs for the regional jet are consistently higher, but the ratio increases with stage length, from 1.7 at 100 miles to 1.9 at 1,000 miles. More important, however, the absolute difference in cost per seat increases rapidly with distance. Therefore, the cost of increasing schedule convenience by offering more flights is the lowest on short-haul flights. On the other hand, the benefits of high frequency are probably greater for these flights, as they often are used for short-duration business trips and also must compete with the automobile. Finally, short-haul segments have traditionally been served by commuter airlines, which in the past were subject to less stringent safety regulation if they operated aircraft of 60 seats or fewer. Pilot contracts with large jet carriers have also limited the sizes of aircraft that can be operated by their lower-paid counterparts working for com- muter affiliates. 5.2.3 Operational Impacts of Up-gauging At SFO, as in most airports, small aircraft use the same run- ways as large ones and occupy them for about the same length of time. Thus, when the airport is congested, the operational impact of a small flight is no less than that of a large one. Indeed, the slower approach speeds and longer in-trail separa- tion requirements of small aircraft can result in longer effective service times. Thus, when airlines and other airport users pro- vide capacity with more small flights rather than fewer large ones, the result can be higher levels of congestion and delay. This does not mean that such a choice is a bad one, but it does imply that the service benefits of operating small aircraft must be weighed against the congestion costs. These trade-offs were analyzed using the four June 2008 days described, based on a deterministic queuing analysis. The approach can be visualized using a queuing diagram, as shown in Figure 5.6, which is based on June 5th operations. The hori- zontal axis is the time of day; the vertical axis is the cumulative count. There are two count curves, one for the schedule and one for actual arrivals. The schedule curve gives the number of flights that are scheduled to arrive at or before a given time. It is constructed by sorting the flights in order of scheduled arrival time. The horizontal coordinate of the point corresponding to the nth flight is the time when it is scheduled to arrive, and the vertical coordinate is n. The actual curve is constructed in a similar way, except in this case the flights are sorted in order of actual arrival time. Looking at Figure 5.6, one can observe that the two curves virtually overlap during the early part of the day. This means that at the time when n flights were scheduled to have arrived, very close to n flights had arrived, implying very little delay. Later on, the curves separate. For example, the 500th arrival was scheduled to occur around 21:20, but it was not until more than an hour later that the 500th flight actually did pull in. This implies that arriving flights at SFO were delayed during the latter part of the day. The total amount of this delay can be obtained by subtracting the sum of the scheduled arrival times from the sum of the actual arrival times. On June 5, it was 12,790 min, or an average of 24 min per flight. Figure 5.5. Operating cost per seat, fuel: $4.30/gal.

Figures 5.7–5.9 show the queuing diagrams for the remain- ing three days. Figure 5.7, for June 13, shows slight delays over much of the day, but no high-delay periods such as seen in the later part of June 5. June 18 is free of significant delays, except for a very small amount at the end of the day. Finally, June 25 is a very bad day, with substantial delays beginning at 8 AM in the morning. Average arrival delays on these three days are, respectively, 15, 8, and 45 min per flight. The research team wanted to estimate how arrival delays on these days would be different if certain flights were removed from the arrival traffic Figure 5.6. Queuing diagram, SFO arrivals (June 5, 2008). Figure 5.7. Queuing diagram, SFO arrivals (June 13, 2008). 95

96 and developed an algorithm for doing so. The details of the process are not important, but the basic principles are very straightforward: • If the actual arrival time of the removed flight was during a period with no delay, removing it will make no difference. • Removing a flight can never make the arrival time of another flight later. • If a delayed flight is removed, the delay incurred by that flight is (of course) eliminated. • If the actual arrival time of a flight is during a high-delay period, removing the flight enables subsequent flights to Figure 5.8. Queuing diagram, SFO arrivals (June 18, 2008). Figure 5.9. Queuing diagram, SFO arrivals (June 25, 2008).

move up and incur less delay, until there is a gap in the traffic stream large enough to make a trailing flight’s arrival time independent of the time of the flight in front of it. Among these principles, the last is the most ambiguous, as one must determine whether another flight could move up if a preceding flight were eliminated or landed earlier. If the time between the two successive arrivals is, say, 60 min, there is clearly no interaction between them, but if it is 1 min, there almost certainly is. The question is where to draw the line. The 90th percentile of the observed inter-arrival times was used (i.e., time between successive arrivals), conditional on airport capacity, in the data. This turned out to be 4 min for high- capacity conditions and 4.4 min for other conditions. With these assumptions, about one third of the total delay—or 8 min per flight—incurred by SFO arrivals can be attributed to arrival capacity constraints. The remainder is due to prob- lems at other airports and airline internal malfunctions such as maintenance problems. Using the ability to predict the delay impacts of removing flights from the arrival stream, the research team considered three up-gauging strategies. 5.2.4 Up-gauging Through Elimination of Short-haul Flights In the first strategy, short-haul flights are eliminated. As observed, short-haul flights generally use smaller aircraft, so this strategy implicitly involves up-gauging. In addition, short- haul flights are most easily substituted by surface transport. Thus, eliminating short-haul flights could be an efficient way to reduce congestion and delay at SFO. In assessing the operational impacts of eliminating short- haul flights or any other strategy, it is useful to quantify delay in seat-minutes rather than aircraft-minutes. The costs of delay to airlines increase with aircraft size, as do (on average) the num- ber of passengers affected by a delay. Thus, the operational impacts will be calculated in units of seat-minutes. Seat-minute delay can be calculated from a queuing diagram in which one counts seats on the vertical axis instead of counting flights. To predict the seat-delay impact of eliminating short-haul flights, a set of hypothetical “cut-off” distances (80, 150, 200, and 300 miles) was chosen. For a given distance, the research team predicted how seat-delay would change if all of the flights within that distance were removed from the arrival stream. Also, to put these results in perspective, the additional line-haul time was calculated if these aircraft seats were transformed into car seats—that is, the passenger on these flights drove to SFO instead of flying. This was done by comparing the scheduled flight time with the driving time estimated from Yahoo maps. This additional line-haul time is also expressed in seat-minutes, based on the sizes of the aircraft used for the eliminated flights. Although the units are the same, the unit values may be differ- ent, depending on the relative seat-minute cost of vehicle oper- ation, aircraft operation, and aircraft delay, as well as the fact that driving times are more predictable than flight delays. Results averaged over all four days are shown in Figure 5.10. Eliminating flights shorter than 150 miles saves more in delay Figure 5.10. Time impacts of eliminating short-haul flights, by cut-off distance, 4-day average. 97

98 than it costs in additional line-haul time. As the cut-off dis- tance increases, more flights are eliminated and the delay sav- ings increases, but the extra line-haul time increases much faster. The cross-over point, assuming equal valuation of the two forms of time, is somewhere between 150 and 200 miles, and probably closer to the former. If, as may well be the case, the unit cost of flight-delay seat-minutes is considerably greater than that of extra flight time, eliminating flights of 200 miles, or even 300 miles, or less may be cost-beneficial. The greatest operational benefit from eliminating short- haul flights occurs on highly congested days, such as June 25 in the sample. Figure 5.11 therefore shows results for that day only. Although it displays the same pattern as in Figure 5.10, the delay savings curve is shifted up, so that, for short distances, delay savings are double the line-haul time increase. Moreover, if the unit cost of flight delay were more than 1.6 times that of extra driving time, eliminating all flights less than 300 miles on such a highly congested day would be justified. The day-to- day differences found in comparing Figures 5.10 and 5.11 point to the promise of having a flexible strategy for serving short-haul trips, using flights on good days and surface modes on congested days. This strategy is referred to as real- time intermodalism. 5.2.5 Up-gauging Through Flight Consolidation A second approach to up-gauging is to encourage, when appropriate, the substitution of less frequent large jet service for more frequent commuter service. As discussed, there are situations in which segments of comparable length and total seat capacity vary differently—for example, Boise to SFO with five small jet flights a day on one airline versus Portland to SFO with two large jet flights on another. In the flight consol- idation approach, the migration of services from the former model to the latter one is encouraged. Like the short-haul flight elimination strategy, this one involves trade-offs. The cost of flight consolidation is less fre- quent service and diminished schedule convenience. To make flight consolidation as painless as possible, it is desirable to identify situations in which the elimination of a flight through consolidation has the least impact on convenience. To quan- tify the effect of consolidation, it is imagined that if a given flight is eliminated, the passengers on that flight would be forced to take the next earlier flight on the same airline from the same origin airport. This is somewhat arbitrary, as passen- gers could respond in other ways, such as taking the next later flight, switching airlines, or going to a different airport. How- ever, the assumption has the virtue of simplicity, and for most passengers, the assumed response is probably the least disrup- tive one. It respects customer brand (and airport) loyalty, and, because it assumes early arrival, does not disrupt passengers’ planned activities in the Bay Area.34 Figure 5.11. Time impacts of eliminating short-haul flights, by cut-off distance (June 25, 2008). 34 On the other hand, schedules at the origin may be disrupted because passen- gers must depart earlier. For this reason, some passengers would opt to take the next later flight, but modeling this mixed response is not complex and probably not worthwhile.

With this assumption, one can evaluate the loss of conven- ience from eliminating a particular flight by finding the differ- ence between its scheduled arrival time and the scheduled arrival time of the previous flight from the same origin oper- ated by the same airline and multiplying this difference by size of the aircraft serving that flight. A metric is obtained with units of seat-minutes, which is traded against the seat-minutes of delay that would be saved if the flight did not take place. This metric is termed “schedule delay impact” (SDI). SDI was evaluated for each SFO arrival in the four June 2008 days in the sample. Figure 5.12 shows the cumulative distribu- tion of the SDI obtained, using a log scale. It is apparent from the figure that flights fall into three categories. First, there is a set of “one-off” passenger and cargo flights for which this met- ric is meaningless. These were all assigned an arbitrary, large SDI value and correspond to the vertical part of the distribu- tion on the right of the figure. Next, there is a set of flights that are the first flights of the day for a given airline and origin. Given assumptions, elimination of these flights would force passengers to travel on the previous day. For all intents and purposes, these flights are “off the table” as far as consolidation is concerned. Flights in this category appear in the s-shaped portion of the curve to the left of the vertical portion. The remaining flights—about 65% of the total—are the ones that can be considered for elimination through flight consolidation. Among these, the most promising are those with the lowest SDI values—say, 10,000 seat-min or less. (To put this figure in perspective, a flight using a 56-seat aircraft would have an SDI of 10,000 if the previous flight was sched- uled to arrive 180 min—or 3 hours—earlier.) The portion of the distribution corresponding to these flights is shown in Fig- ure 5.13. It shows that a small but non-zero fraction of flights have an SDI of zero. These are cases in which airlines inten- tionally schedule two arrivals from the same origin at exactly the same time. The data contain five such cases, all involving major carriers operating large equipment from distant hubs. Carriers do this presumably to provide sufficient capacity while maintaining the ability to cancel flights without disrupt- ing passengers when traffic or capacity is low. Aside from these cases, the lowest SDI values are on the order of 1,000-seat min- utes, the equivalent of a 33-seat flight whose predecessor is 30 min earlier. The research team used the SDI metric to identify the best flights to eliminate in pursuing the flight consolidation strat- egy. Analogous to the short-haul elimination strategy, a mini- mum SDI value was set and eliminated all flights below that value. The impact on queuing delay at SFO was then assessed. The procedure is somewhat complicated by the fact that when one removes a flight, the SDI values of other flights may change, as the removed flight is no longer available to receive passen- gers from some other flight. The team therefore updated the SDIs after each flight consolidation. The results for June 25, the worst day in the sample, appear in Figure 5.14. Queuing-delay savings of a magnitude greater than schedule- delay savings are obtained for SDI cut-offs up to 4,000 seat- min. Queuing delay is clearly more expensive than schedule delay, as it ties up aircraft and forces passengers to wait in planes and airport terminals, whereas schedule delay can be Figure 5.12. Cumulative distribution of SDI, SFO arrivals (June 2008). 99

100 anticipated and incorporated into passengers’ activity sched- ules. For these reasons, it is not unreasonable to assume a unit cost ratio of 2:1 or more. Figure 5.14 suggests that a consider- able number of flights could be eliminated through consolida- tion before the optimal trade-off point is reached. 5.2.6 Up-gauging by Diverting Very Small Aircraft Finally, the strategy of diverting small aircraft from SFO to some other local airport was considered. The research team set 15 seats as the threshold for small aircraft, which would eliminate all bizjets but no commuter flights. To assess the mobility impacts of this strategy, one must assume a time penalty for diverting a flight (or a seat) from SFO to some other alternative. That penalty reflects the additional travel time from being forced to fly into a less accessible airport. Depending on one’s point of view, it may also be increased to capture the greater value of time of bizjet travelers as com- pared to the rest of us. Figure 5.15 summarizes the impacts of this strategy. On June 25, diverting aircraft of 15 seats or fewer saves over Figure 5.13. Cumulative distribution of SDI, SFO arrivals (June 2008). Figure 5.14. Time impacts of eliminating short-average; June 25, 2008).

50,000 seat-min of delay on average and over 80,000 seat- min. The savings exceed the access time penalty unless diver- sion penalty exceeds 5 hours. As the actual extra time involved cannot be more than a hour or so, one would need to value biz jet passengers’ unit time cost at more than 5 times that of queuing delay to justify their presence at SFO. 5.3 Implications It has been shown that changing the schedule at SFO, whether by eliminating short-haul flights, consolidating flights, or diverting very small aircraft, can reduce delays and often does so at a reasonable cost in terms of the extra line-haul time, schedule delay, and access time that such changes require. There is strong evidence that the conclusion generalizes. For any airport with high delays due to inadequate operational capacity, eliminating flights during busy periods will reduce delays considerably. The quantity of this benefit, as well as the cost of losing any particular flight, will vary from flight to flight, time period to time period, day to day, and airport to airport. There is, however, a wide body of research and experience sug- gesting that, in many circumstances, the benefit greatly exceeds the cost, and that the cumulative gain from such changes would be impressive. What could be done to realize these gains? If the answer to this question were easy, it would have already been done. Broadly speaking, in the current system there is no actor who has both the authority to make the desired schedule changes and the ability to realize the gains from doing so. Airlines and other aircraft operators whose decisions determine the flight demand at any particular airport can realize benefit for some flights they control by eliminating or rescheduling other flights, but this is generally a small fraction of the total bene- fit (1). Moreover, in a competitive, unregulated, industry, the elimination of a flight by Airline A may be offset by initiation of a new service by Airline B. In this event, A has not only lost the operational benefit from its schedule change, but it also now faces stronger competition. The misalignment between authority and benefit realiza- tion is greatest in airports that have both substantial conges- tion and an unconcentrated distribution of flight traffic. In the United States, such airports are found primarily in the coastal mega-regions. By virtue of their high densities and land constraints, these are places with high levels of flight traf- fic relative to airport capacity. On the other hand, the periph- eral location of coastal airports discourages their use as hubs, which tend to be more concentrated. Thus, in addition to the other conditions—such as strong competition from other modes—that make coastal mega-region airports unique, such airports face unique challenges in encouraging their users to schedule flights that appropriately reflect the costs of congestion. 5.4 The Role of Airport Managers in Increasing Capacity Airport operators in the United States are only one element of a complex set of factors that affect the creation and resolu- tion of airport capacity issues. A major, yet highly constrained and limited, role is played by the airport managers. Figure 5.15. Time impacts of eliminating short-haul flights, by assumed access penalty, 4-day average. 101

102 5.4.1 Financing Enhancements The most obvious, and best understood, role of airport operators is in providing the physical infrastructure that sup- ports increased airport traffic. The United States is almost unique in providing a system of federal grants and munici- pal debt financing that, when combined with local public ownership of airports, provides both the capital and man- agement resources needed for stable investments in infra- structure at the large airports that handle most passenger traffic. These annual investments typically range from $7 to $10 billion. The primary financing tools used to support airport capi- tal development are debt from the tax-exempt municipal bond market, Passenger Facility Charges (PFC), the federal Airport Improvement Program (AIP), and retained earnings. Together, these funding sources provide large airports with the financial capability for infrastructure development. Medium- sized airports, while having less financial capability, still retain access to these resources, with funding levels that generally increase with increases in passenger traffic. However, as is well known, adverse public reaction to air- craft noise and pollutant emissions at and near major airports continues to seriously impede development of new airport infrastructure. This resistance is unlikely to decrease at the study area airports, and major development in the form of new airports or new runways at existing major airports is unlikely, despite the ability to adequately fund such projects. Still, at the core, the public management teams who operate airports in the United States are deeply committed to expand- ing airport capacity through infrastructure development and the deployment of new technologies and procedures, wherever possible. 5.4.2 Market Factors A second major factor affecting airport capacity is the inter- action of airports, airlines, and market forces. The decisions airlines make about which markets to serve, which aircraft types to serve them with, and the frequency of service all have a large impact on airport capacity. Yet, because of federal laws and regulations, airports are extremely limited in affecting those airline decisions. As discussed in Chapter 1 (Section 1.4.2), airport managers are often disconnected from decisions made by the airlines concerning where to concentrate hubbing activities and where to eliminate them. As noted in that section, fluctuations in airline policies have significantly affected management of delay (congestion) at SFO, LGA, and JFK. The need to clarify the airport manager’s role in the determination of the num- ber of flights, and to forge a cooperative relationship with the airlines in the provision of those flights, has emerged as major themes of this research. 5.4.3 Restrictions Based in Law Despite the tremendous impact of airline scheduling deci- sions on airport capacity, airport operators are very limited by federal law, regulation, and policy in their ability to control scheduling practices or aircraft size. Although a full legal dis- cussion of these restrictions is beyond the scope of this study, understanding the issues affecting airport operators requires some discussion of the pertinent restrictions (2).35 In some cases, restrictive interpretation of legally enforce- able policies may act as a capacity constraint, or at least impede potential solutions. The most noteworthy is the assurance of AIP grant recipients to the secretary of Transportation. This assurance is threefold: that their airports will be available for public use on reasonable conditions and without unjust dis- crimination, that air carriers making similar use of an airport will be subject to substantially comparable charges, and that the airport operator will not withhold unreasonably tenant/ signatory status from an air carrier that assumes obligations substantially similar to those already imposed on air carriers of that classification or status. This assurance, required by Title 49 USC 47107, and the restrictions contained in Title 49 USC 40116 on state and local taxes, fees, and other charges on air travelers and air transportation, have proven to be significant issues with local efforts to allocate traffic among airports. In general, airports are prohibited from direct regulation of airline routes, rates, and charges. This prohibition has been determined to include direct regulation of equipment type, fre- quency of operation, time of day of operation, and aircraft environmental emissions. Airport proprietors’ rights have long included the right to establish discriminatory fees and charges for aeronautical use of the airfield. These rights have been rec- ognized as including the right to set minimum landing fees designed to affect various weight classes of aircraft differently, with the intent of providing incentives to reduce airfield delay during periods of congestion. In addition, the DOT 1996 Rates and Charges Policy (3) provides that an airport owner may impose a “properly struc- tured peak pricing system that allocates limited resources using price” and may “establish fees that enhance the efficient utilization of the airport.” Similarly, the DOT’s regulations on airport noise and access restrictions (14 CFR Part 161) provide that a peak-period pricing program with an objective “to align the number of aircraft operations with airport capacity” is not an “access restriction” (4). In a series of decisions in the Mass- port Program for Airfield Capacity Efficiency (PACE) pro- ceeding, where the airport operator sought to use landing fees to regulate airfield congestion, the DOT concluded that, “landing fee structures that vary from the traditional weight- based approach are permissible so long as the approach 35 A thorough legal discussion of these issues can be found in the DOT NPRM on its policy on airport rates and charges.

adopted reasonably allocates costs to the appropriate users on a rational and economically justified basis” (5). 5.4.4 FAA-Proposed Changes in Rules/Regulations Most recently, in the Notice of Proposed Rulemaking to amend its 1996 policy on rates and charges, the DOT pro- posed to explicitly allow airport proprietors to establish a two- part landing fee that recognizes both the number of operations and the weight of the aircraft, in order to provide incentives for airlines to modify aircraft gauge or frequency to reduce delays at congested airport. According to the U.S. Govern- ment Accountability Office (GAO), the 2008 Amendment to the Airport Rates and Charges policy allows the following: Announced in July 2008, the policy clarifies the ability of air- port operators to establish a two-part landing fee structure con- sisting of both an operation charge and a weight-based charge, giving airports the flexibility to vary charges based on the time of day and the volume of traffic. It also permits the operator of a congested airport to charge users a portion of the cost of airfield projects under construction and expands the authority of an operator of a congested airport to include in the airfield fees of congested airports a portion of the airfield fees of other under- utilized airports owned and operated by the same proprietor (6). The combination of these three new rules would give air- port managers more control over the efficient use of their runway and landside facilities, “These amendments are intended to provide greater flexibility to operators of con- gested airports to use landing fees to provide incentives to air carriers to use the airport at less congested times or to use alternate airports to meet regional air service needs” (6). Airport operators have essentially no direct control of airline activity at their airports, including whether the airline serves the airport at all, the frequency or time of day of service, or the aircraft type or size used to provide service. They do have pro- prietors’ rights to use rates and charges to influence airline ser- vice patterns, but those rights are still being refined. Building on the experience reviewed here, the following section will develop some potential demand management principles. 5.5 Guiding Principles for Demand Management 5.5.1 Legitimation In light of the potential to reduce delay with innovative freight management and the unclear role of aviation stake- holders in managing delay, a demand management approach could be tried, to better align flight scheduling decisions with the needs of society. The most fundamental, and therefore the most difficult, would be for all relevant parties to recognize demand management as a legitimate alternative to capacity expansion as a means of ameliorating airport congestion prob- lems. Some parties within the aviation community continue to believe that congestion and delay are required to spur develop- ment of the aviation system both nationally and locally. Demand management is perceived by such individuals as a palliative that inhibits the often difficult cure of capacity expan- sion. In such a view, only when the pain becomes unbearable— as in the New York airports or Chicago O’Hare—should demand management be undertaken. And even in these cases, the goal should be simply to reduce congestion to tolerable lev- els. Any further refinements, such as market-based slot alloca- tion, again raise the specter of demand management becoming a way of life, rather than a temporary expedient or last resort. There are certainly cases where capacity expansion is more desirable than demand management. But the time is long past when capacity expansion should be viewed as the inherently superior solution. At many airports, given current technology and regulations, it may no longer be feasible to expand run- way capacity. At others, the costs of expansion may simply be higher than those of demand management. And at yet others, agreement on an ultimate capacity may be the price for secur- ing approval for expanding the airport to reach that capacity. In all of these instances, demand management could be a legit- imate tool for preventing or alleviating excess airport delay. Beyond legitimizing demand management, the approach could be guided by two other principles. First, primary respon- sibility for demand management should be at the local level. Second, demand management should be anticipatory rather than reactive. 5.5.2 Localization There are a number of reasons why the primary locus of demand management responsibility and action would ideally be at the local level. First, recent federal efforts to innovate policy in this area have been met by strong resistance. The FAA’s attempt in the fall of 2008 to institute slot auctions at a modest scale at the New York airports was temporarily blocked by a federal court after an appeal by the PANYNJ. Demand management at the local level would be immune to legal or political challenges. It is likely that slot auctions would be opposed by airlines regardless of the body implementing the auctions. Political hurdles would also exist. In 2007, the FAA proposed a pilot program to give select airport authorities flex- ibility to impose market-based measures at the local level with guidance; this proposal did not gain traction and was not included in recent reauthorization bills that were introduced. Another challenge to demand management policy localization, airport monopoly power, is touched on in Section 5.7.2.2. There would likely be significant challenges to innovation in airport demand management whether it occurs nationally or 103

locally, subject to federal oversight. Nonetheless, the research team argues that the latter course would be the more promis- ing one. Second, it would be very difficult to craft a federal demand management program that would be effective across the wide variety of circumstances that exist at different airports. Impor- tant differences in this context include the following: • Airline/airport relations. Although in some cases airlines and airports maintain a straightforward landlord–tenant relationship, in other cases the relationship more closely resembles co-ownership. In the latter case, certain airlines have invested in both the airport itself and in developing the markets that the airport serves. In the context of demand management, this affects the manner in which available capacity could be allocated: through a market mechanism based on willingness to pay, or through a process that gives more consideration of established airport- airline relationships. • Financing mechanism. Airports are financed in one of two ways. In the residual approach, airlines agree to make up any shortfall in revenues in return for having a strong role, often including a veto, in airport capital expenditure decisions as well as the agreement that any airport’s non-airline revenue will go toward reducing the costs borne by airlines. In the compensatory approach, the airport assumes the risk, and in return can earn substantial surpluses that can finance future airport development, decisions about which it largely con- trols. Residual airports face unique constraints in employing market approaches to demand management because (a) any revenue from such charges is ultimately recaptured by the airlines in the form of reduced fees and (b) they typically have long-term usage agreements with airlines to which any demand management program must conform. • Variability in capacity. Some airports have fairly similar capacities under most weather conditions, whereas in others, capacity is highly variable. In the latter cases, a decision must be made about what capacity scenario to assume in formu- lating the demand management strategy. If the capacity is set too low, the airport will be underused much of the time; if set too high, there may be severe delays much of the time. There may also be cases where it is appropriate to assume dif- ferent capacities for different times of day or seasons of the year. Such trade-offs are best understood at the local level. • Expandability. The appropriate mix of demand manage- ment and demand accommodation depends on the cost and political difficulty of expanding an airport. Some fac- tors that determine expandability, such as the cost and availability of land and the sensitivity of surrounding land uses, can be assessed objectively, whereas others cannot. This is one reason why airport planning and expansion decisions have traditionally been made at the local level. Given the close coupling between such decisions and those related to demand management, it is appropriate for the same entity to make both. • Valuation of competing goals. Demand management involves complex trade-offs between competing goals, including delay reduction, schedule convenience, competi- tiveness, equity, and service stability. Different regions will place different values on these goals. Localizing demand management policy increases the opportunity to design programs that reflect these differences. • Competitiveness. Demand management policies can reduce competition between airlines serving a given airport as well as create entry barriers for airlines seeking to initiate service. Although such outcomes are rarely desirable, the severity of their consequences varies according to how competitive the airport is to begin with, the availability of alternative airports nearby, and, in some cases, the availability of competitive modes. It follows that the weight given to preserving com- petition in formulating demand management programs should vary from airport to airport. A third rationale for demand management being deter- mined at the local level is that, for the most part, delay is a local problem. It is the local population and economy that experi- ence the brunt of delay impacts. Although high delays at one airport can propagate throughout the system, most of the delay experienced in the United States is not propagated. Moreover, the airlines that operate at a high-delay airport recognize the system-wide impacts of the delays and will certainly express these—both explicitly and behaviorally—to local policy- makers. There is also anecdotal evidence from places such as San Francisco, New York, and Boston that if demand manage- ment were made a local responsibility, it would be embraced by many of the localities where it may be needed. To ensure that a solution developed to solve a local delay problem does not have the effect of making the situation worse downstream at other airport(s), the delay modeling used to develop the delay triggers at an airport would account for the impact on other airports. Fourth, local responsibility would result in a variety of approaches being tried. Much can be learned from this process. Just as states are the laboratories of democracy, airports could become laboratories for demand management. Our limited experience with airport demand management in this country, as well as the limited success of attempts at it to date, suggests that there is much to learn. Finally, as the research team elaborates below, making demand management primarily a local responsibility does not mean that the federal government would have no recourse when that responsibility is performed improperly or not at all. Airports’ increased latitude in developing demand manage- ment programs could be accompanied by clear guidance and principles of accountability. 104

5.5.3 Anticipation As currently practiced in the United States, air demand management is a reactive strategy that is performed after delays have reached unacceptable levels. For example, there is legislation authorizing meetings with airlines to discuss sched- ule reductions at severely congested airports. The authority appears to be restricted to cases where the airport is already severely congested. In contrast to this, the demand manage- ment programs can be implemented most effectively prior to the advent of severe congestion. Such pro-activity could take two forms. The demand man- agement programs could be formulated while the airport is relatively uncongested and prior to the time when severe con- gestion is clearly foreseeable. This would allow a deliberative approach. Moreover, it would require stakeholders, lacking reliable information about when and where congestion will occur, to participate in the process without clear knowledge about how it will affect their own interests. Second, the program itself could be proactive, with actions that are triggered when unacceptable levels of congestion are foreseeable, rather than when they actually occur. It is possi- ble to foresee congestion because airline schedules are avail- able several months ahead of time. Capacity, the other key determinant on congestion, can also be confidently charac- terized within this time frame, at least in a probabilistic sense. This demand and capacity information would be used to determine what, if any, demand management actions are needed. Airlines could then make adjustments to their sched- ules accordingly, thereby alleviating or preventing the con- gestion that would otherwise have occurred. This is the basic logic of the Massport demand management plan put in place several years ago at BOS. Such an anticipatory approach offers great advantages for both airlines and passengers. Airlines are given relatively long lead times to adjust their schedules. This expands the range of possible responses. Carriers can adjust flight schedules, shift operations to other airports, up-gauge, and increase fares in congested periods. Passenger dislocation is kept to a minimum, as most bookings are made just a few weeks in advance, well after the schedule adjustments would have taken place. Such an approach would certainly arouse concern for those who believe that severe congestion is the only reliable means to secure consensus to expand airport capacity. In this research, severe capacity shortages manifest primarily in the form of demand management actions and air carrier responses to them, rather than long delays. But these actions and responses will themselves be clearly visible to policymakers, stakeholders, and the public at large. Local politics could be relied upon to bal- ance the costs of demand management against those of increas- ing airport capacity. 5.6 Guidance and Accountability This research has revealed the benefit and need for multiple parties to be at the table when considering airport demand management. At a recent panel discussion, participating air- port operators expressed their support for multistakeholder involvement stating that the airport operator should be a strong player in capacity and demand management. To work together effectively, all parties would need to be clear on expec- tations and roles. Clear guidance could help airports manage congestion through the entire process. Guidance can be thought of as the boundaries or constraints of operation. Within these boundaries, the airport would have latitude in how it decides to manage congestion. The following two exam- ples will shed light on how guidance for overall management of capacity and delay might occur. 5.6.1 Existing Examples of Guidance Although local airport managers enjoy a unique perspec- tive on their airports, they also are heavily involved with the air carriers at the airport. Federally mandated rules, such as the development and adherence to airport competition plans and providing service to small communities, provide the air- port with a clear set of rules for airport management. This can prove to be useful, as explained in the following examples. Airport competition plans are developed by an airport fol- lowing guidelines set out by the FAA. The plans are to show how an airport is open to opportunities for carrier relations (7). Although it is an extra task for an airport to develop such a plan, airports welcome these plans because they provide operating guidelines for the airport that are agreed upon with the FAA. Having federal policy that supports new entries allows airports to uphold competition. The competition plan allows the airport to have guidelines against which airline service decisions can be made; the federal government helps preserve competition. Defining small communities for designated air service and allocating seat capacity to these small communities are tradi- tionally federal roles. Definition of these small communities is often politically charged. Instead of the airport getting involved in both politics and controversial carrier relations, the federal government can define small communities and necessary capacities to service these communities. The airport avoids discussing with communities why they were denied access and avoids any difficult carrier relations. 5.6.2 Example: Developing a Framework for Demand Management In consideration of Section 5.2 which displayed how delay can be reduced with innovative aircraft management, the 105

following discussion centers on an example of how a frame- work for the airports to manage congestion might be devel- oped. A broad outline for how this could be accomplished is the following: setting mutually agreeable airport-delay targets, developing a detailed list of actions an airport can take to meet the delay level, and identifying incentives and penalties for not meeting such an action. The first step would be to define “critical-delay airports.” This definition could be tied with an existing program, such as the Operational Evolution Partnership (OEP) airports, which are the commercial U.S. airports with the most activ- ity. According to the FAA (8), more than 70% of passengers move through these airports, and delays at the OEP 35 air- ports have a ripple effect on other locations. Therefore, con- taining delays at these airports could be considered to be of national significance. Critical-delay airports could be given this designation to ensure that local decisions regarding a congested airport do not hinder the entire NAS. The goal would be for these air- ports to hold delays to a certain level. This level is called a “trigger” because any delay experienced beyond this level could set off a series of actions. The trigger would be deter- mined by using a combination of experience and economic modeling techniques. This delay trigger would be developed to ensure other airports in the NAS are not unduly affected by local decisions, particularly ones that result in high levels of delay. It is also expected that airlines using an airport would be well aware of the ripple effects to their own opera- tions caused by delays there and would make these known to the local airport operator. Critical-delay airports could be further divided based on their current levels of delay. Once the delay trigger is decided, each critical-delay airport could model the delay experienced on a fair-weather day with their existing schedules and an estimation of unscheduled traffic. Airports could be divided into those that exceed the trigger under existing conditions and those that will exceed it in the future. 5.6.2.1 Airports with Trigger Exceeded Airports that immediately exceed the delay trigger would update their master plan at once. This airport master plan update could have, at minimum, two new sections. One could address the potential of the airport to expand capacity in the long term to manage demand. The other section could be the development of a demand management plan. The capacity expansion plan could take the perspective of regional growth accommodation instead of airport-specific growth. The airport could study how it can provide its passen- gers service without being limited to runway development. The airport capacity expansion plan could incorporate many strategies, including multimodal solutions, regional airports, airports under the main airports purview, and HSR. An exam- ple is the Massport 1993 Strategic Assessment Report, which evaluated regional solutions for intercity travel demand. The demand management plan could outline the steps an airport would take to enforce the delay trigger. This could include the strategies to be employed to satisfy the trigger and also the details behind these strategies. For example, if an air- port planned to use peak-period pricing to reduce delay in its demand management plan, the airport would discuss the extra peak-period charge, the duration of the peak in which to charge, and other details. Other detailed strategies an air- port could consider are discussed in Section 5.7. 5.6.2.2 Airports with Trigger Not Exceeded Airports where the trigger is not exceeded could be further subdivided into two categories. Some airports would find through their modeling that traffic will exceed the trigger before their next scheduled master plan update. Such airports could update their master plan—potentially immediately. Airports where the trigger would be exceeded in over 5 years but before the next master plan could update their master plan in a 5-year period. 5.6.3 Airport Accountability Airports could have wide latitude to manage their own con- gestion and delay and could accept consequences for failing to meet the delay standard. To encourage airports to accept their designation of critical delay airports, incentives could be provided. 5.6.3.1 Accountability Incentives There are ways to encourage airports to embrace the critical- delay airport designation. As noted, airport managers tend to favor solutions beyond demand management; one way to encourage airport managers to see greater benefits of demand management could be to allow more flexibility in using rev- enue from their demand management plan. If some aspects of revenue neutrality (discussed in Section 5.7) were relaxed and the operator was allowed to have wide latitude to use the funds to make improvements, an airport operator might more read- ily embrace the critical-delay airport designation. There is general consensus among experts that operators of congested airports have to make money to run the airport as efficiently as possible. Any revenue raised could be considered funds for airport improvements. Certain airport demand patterns make revenue neutrality a challenge. For example, an airport with persistent all-day demand is unable to offer a discount in the off-peak, as the off-peak does not exist. Another example is an airport that 106

experiences extreme peaks and an entrenched hub carrier. An airport with extreme peaking could offer negative land- ing fees; however, if the airport is a hub airport, the hub carrier would almost exclusively benefit from such an organ- ization. Some airports could be able to remain revenue neutral with a demand management plan because they have a significantly long off-peak period in which they can offer discount landing fees. These airports could be given the option of remaining revenue neutral. These airports could use their revenue to offer off-peak discounts; they could also choose to operate like airports unable to remain rev- enue neutral. Revenues could be placed in an airport’s reserve fund and available for a wide range of purposes—from the capital pro- gram to the maintenance reserves. Although airports would have latitude in spending these funds, additional guidance would be necessary about how funds could be spent. 5.6.3.2 Passenger Accountability When airports accept public funds from the FAA, they agree under United States Code Title 49 (Section 5.4.3) to conditions of grant assurances. Agreeing to these assurances means that all aircraft that can safely land at that airport must be accommo- dated with no discrimination. Section 5.2 introduced the idea that carriers have fleet mix recipes with important ramifica- tions for the quality of service and the level of accessibility pro- vided by airports and the entire air transportation system. The guidance provided to airports for accepting their designation as critical-delay airports could involve a new way to envision aviation system accountability. Section 5.2 showed that a balance exists between providing customer service in the terms of flight frequency and reliability. It is possible that making an airport available to all aeronauti- cal users increases frequency and degrades customer service to a point where an airport is not “reasonably available” to pas- sengers. Similarly, nondiscrimination could lead to over sched- uling and, as shown in Section 5.2, to the over scheduling of small aircraft. This excessive frequency tips the balance so that reliability is significantly decreased. One of the many motivations behind demand manage- ment is to make air travel reliable for passengers. A way to consider airport accountability beyond making the airport available for all classes of aeronautical activities is to focus on the passenger experience. A congested airport does not neces- sarily make the airport reasonably available nor are delays arguably nondiscriminatory from the passenger perspective. When there is a delay, passengers experience a loss from as small as lost time to a missed connection. According to research performed by the research team, passenger delay at the 12 largest coastal mega-region airports in 2007 cost passen- gers $15.4 billion per year (as defined in Table 1.2, Section 1.2). Said another way, each passenger would be willing to pay an average of $48 per passenger trip to avoid delays. Coupled with the findings that many delays are related to airport congestion, a conclusion is that the 12 largest coastal mega-region airports are not available on reasonable terms because passengers have a large willingness to pay to avoid delays. For this reason, it is possible that although congested airports are following the classical definition of making an airport available on “reason- able terms,” passengers are not being served on reasonable terms. The delays incurred are also not equally distributed. Air- port passengers on high-frequency, low-capacity aircraft are experiencing less delay than passengers on low-frequency, high-capacity aircraft because passengers on frequent, low- capacity aircraft experience less schedule delay (the difference between desired departure time and actual departure time) than passengers on larger, less frequent aircraft. Again, although an airport may not be discriminating against classes of aeronauti- cal users, the delay distribution is not entirely equitable. This key guidance that an airport receives is crucial in motivating an airport to manage congestion and delays in a way that is agreeable to all parties. The following section dis- cusses an example of an airport receiving guidance from the FAA to develop a comprehensive demand management and capacity enhancing plan. The research team is not consider- ing this a “best practice” case study, but rather one instance of guidance in action. 5.6.4 Example: Guidance in Action The following example is based on Massport and the FAA’s relationship in managing BOS—a unique airport and histor- ically one of the nation’s most delayed airports. It is an OD airport as opposed to a transfer, hub airport. It is the biggest airport in the upper Northeast and therefore attracts a diversity of fleet mixes. Being in the Northeast, BOS deals with extreme weather patterns. Finally, BOS is constrained in expansion due to its location and is operationally challenged due to community pressure. In the mid-1980s, Massport worked to implement PACE. This program was implemented in response to a strong growth in regional operations; growth in this sector was threatening the capacity of the airport. The program changed the landing fee formula from weight-based to a hybrid-fixed and variable structure. However, owing to small aircraft being charged more and larger aircraft charged less under this fee structure, the DOT found that this scheme discriminated against an aero- nautical user group and therefore was in violation of the grant assurances. While PACE was closed, persistent delays continued at BOS, motivating further studies and investigations. Massport proactively sought out solutions that ranged from HSR to the increased use of regional airports. The purpose was to accom- 107

108 modate demand for intercity travel involving the Northeast region in the long run. However, the result of these studies pointed to BOS as the focus airport for the region. To this end, the airport proactively performed an environmental impact statement (EIS) in 1995. The goal of the EIS was a delay reduc- tion program at BOS. The EIS included a feasibility study, which looked at different delay reduction items and came up with the following list of potential strategies: • Demand management and peak-period pricing, • A new runway and new taxiway improvements, and • Using technologies to reduce spacing minimums on cer- tain runways. It is important to note from this list that demand manage- ment was one part of a larger list of delay reduction strategies. The analysis performed for the EIS showed that all these strate- gies were necessary for delay management. Furthermore, while a new runway was suggested, Massport was committed to stay- ing within their spatial footprint and therefore knew a demand management portion of their plans was necessary to comple- ment their capacity expansion. Because the strategies were complementary and Massport would remain in their existing footprint, the FAA Record of Decision and the state permit felt that all strategies should be implemented (5). Massport agreed to develop a demand management plan to complement their new runway development. This transparent, three-pronged delay reduction strategy also earned Massport stakeholders buy-in, as stakeholders were able to see the trade-offs among the strategies. Massport was obligated to develop a demand management program along with the plan for runway devel- opment, so as to both manage future demand and manage it in such a way that all stakeholders are prepared. The demand management program at Massport is out- lined as follows. Every 6 months, the airport collects the schedules given by the carriers. The airport then develops a monitoring report. This report involves the airport entering the collected schedules and non-scheduled traffic into a sim- ulation model to estimate whether the airport is oversub- scribed on a fair-weather day. Massport has a delay trigger of a 15-min average total delay per operation over a period of 3 consecutive hours. If the simulation report finds that this trigger is exceeded, the airport will tell the airlines that the congestion management program will go into place in the next schedule iteration. This action puts the airlines on notice that a peak fee will be implemented if the schedule does not change. The airport will recalculate the delay if the airlines update their schedules. Figure 5.16 displays the process for implementing the peak period pricing at BOS. The FAA provided clear guidance to Massport to develop a capacity enhancement and demand management proposal in response to the airport’s findings that both a runway for increased capacity and a demand management program would be necessary to balance future capacity and demand. There are many federal constraints with which Massport had to operate to develop their demand management plan. The plan refer- ences two decisions that, taken together, form the basis of the guidance for Massport’s demand management plan. The 2008 Record of Decision allows airports to implement peak pricing under certain conditions; the PACE decision separates the role of the airport and the role of the FAA. In addition to these rul- ings, the FAA provided Massport with “guidance in the form of a policy statement relating to the development of a reason- able fee structure and in two pending policy initiatives address- ing airport proprietor demand management programs” (5). It is this guidance from the FAA that enabled Massport to Figure 5.16. Time impacts of eliminating short-haul flights, by assumed access penalty, 4-day average (9).

develop their demand management plan with confidence after their previous attempt, PACE, was struck down. 5.7 Flexibility In Section 5.6, the research team described an example of how a framework might be developed for implementing demand management. In such a scenario, it would be impor- tant that airports have flexibility in how they could perform demand management. This section discusses that flexibility by examining the actions that an airport could take to meet their delay goal. Airports could have many options to man- age congestion in a way that fits the unique needs of each air- port. Some of these involve loosening current restrictions on how airports can charge for airport usage, while others entail active cooperation of interested parties. Three potential categories of actions an airport could take to manage their delays are introduced: capacity allocation, setting operational limits, and traffic flow management. Insti- tutional changes that could help an airport employ these strategies are also discussed. 5.7.1 Capacity Allocation 5.7.1.1 Pricing The idea that peak-period pricing could be used to help manage the demand/capacity balance for transportation ser- vices was first proposed formally in 1959 by William Vickrey, a Columbia University economist. That concept was advanced in 1989 in a text produced by a team at the Brookings Institu- tion and has since been widely promoted in the economics and transportation literature as well as in public policy forums (10). Testimony to the U.S. Congress JEC provided in 2003 by the GAO concludes that, “Congestion pricing—although only one of several approaches that can be used to reduce congestion on our nation’s roads, airways, and waterways—shows promise in reducing congestion and better ensuring that our existing transportation systems are used efficiently” (11). The first roadway area pricing project was implemented in Singapore in 1975, and there have since been numerous projects in which peak pricing has been used for U.S. roadways. The FHWA’s Value Pricing Pilot Program has funded numerous studies and implementations of congestion pricing approaches for road- way projects. The use of congestion pricing for managing airport capac- ity in the United States has been more limited. The PANYNJ implemented the first runway congestion pricing scheme in 1968, charging higher landing fees for peak-hour use by small aircraft at EWR, JFK, and LGA. As a result, general aviation activity during peak periods decreased by 30%. The peak- hour fees were discontinued after airline deregulation. In 1988, the Massachusetts Port Authority implemented higher landing fees for small aircraft at BOS as part of the PACE. With PACE, Massport experienced a significant drop in small regional aircraft. Although PACE was successful in promot- ing up-gauging, it was also found to be in violation of Title 49, U.S. Code, because the airport was not available to all aero- nautical users on “reasonable terms without unjust discrim- ination.” Those fees were also discontinued after a court order cited the lack of reasonable airport alternatives in that region. Although the PACE program was ultimately found to be in violation of U.S. Code, it provides a good example of two crucial components of charging policies: nondiscrimina- tion and revenue neutrality. Changing the way aircraft operations are charged allows for demand management. Air carriers would have an incen- tive to use scarce runway capacity more efficiently, through up-gauging or rescheduling flights to less congested periods. Furthermore, signals (in the form of prices) that capacity expansion is needed would be sent. A fee for use of the airfield could be levied in several ways. The most common approach is a peak-period surcharge that applies to all flights regardless of size. The permissibility of this option is explicitly mentioned in the Airport Noise and Capac- ity Act as well as the most recent FAA guidance on airport rates and charges. It is also the centerpiece of the Massport plan for BOS. Such a charge encourages aircraft operators to reduce flights in peak periods. Moreover, because the charge is size- independent, flights carrying smaller numbers of passengers are likely to be the most strongly affected. Depending on the circumstance, flights would be shifted to off-peak periods, combined through up-gauging, or eliminated altogether. Although such a surcharge seems to be the most logical price-based solution, it might not always be feasible. For example, airports that have long-term agreements on airfield charges might be unable to obtain approval to amend the agreement. Other pricing changes could be considered in such cases. One example is a peak-period PFC surcharge on passengers. PFCs are automatically added to the price of a passenger ticket, based on the airports included in the itiner- ary. There is no obvious reason why the same system could not be employed to vary the PFC based on the flight arrival and departure times. As compared with a peak-period flight surcharge, a disadvantage of the PFC approach is that it does not reward up-gauging. This problem could be addressed by differentiating the PFC by the size of the aircraft on which the passenger is flying. Parking and ground transportation charges could also be adjusted to encourage air travelers to fly in off-peak times. At OD airports, periods of high landside congestion are associ- ated with periods of high airside congestion. Peak surcharges on landside access and egress can address both the landside problem and the airside problem. An additional advantage of 109

this approach is that it could be done under current federal policy, which affords airport operators wide latitude to set landside charges. A disadvantage is that it would not be pos- sible to differentiate landside charges by aircraft size. Addi- tionally, it could be more difficult to make travelers aware of the price differentials when they make their travel decisions. 5.7.1.2 Capping The alternatives to peak surcharges involve setting opera- tional caps, as has been done in various fashions by the FAA—in cooperation with incumbent airlines—at severely capacity-constrained airports since the late 1960s. Airports could have discretion to manage demand through such slot controls. They may choose to do so for two main reasons. First, such controls are the most direct and reliable way to reach a desired level of demand and therefore operational performance. Responses to pricing solutions are difficult to predict, and the process of finding the right peak surcharge— however it is applied—would involve trial-and-error and, as demand conditions are ever-changing, continual adjustment. The slot-based solution “cuts to the quick,” in this respect. Second, slot control is the most obvious—if not the only— practicable method of controlling demand for airports that do not wish to employ market-based solutions. The primary disadvantages of slot limits are twofold. First, it is difficult for an airport to determine the appropriate num- ber of slots. The decision depends not only on capacity, but also on the optimum level of congestion. A price-based solu- tion will allow users to express, through their willingness to pay surcharges during peak times, what level of congestion is ideal. This is not possible with caps. Second, a means for allo- cating the slots must be found. There is considerable experi- ence in this area, but the methods employed to date all give a considerable advantage to incumbent airlines, resulting in entry barriers. The FAA has recently studied, and attempted to implement, auctions as a means of allocating slots at the New York air- ports. The idea proved unpopular with airlines and the PANYNJ, and it appears unlikely that it will go further as a fed- eral initiative. Airports, however, could be given considerable latitude to experiment with slot auctions. A key problem with the previous initiative is that the use of a runway slot entails the use of a host of other facilities (e.g., gates, baggage facilities, ticket counters, and ground access) that are the province of the airport. Designing auctions that take this into account, so that transfers of all of these resources from one airline to another can be successfully coordinated, is a challenging task in any case, but one that the local airport is best equipped to handle. A less adventurous alternative than auctions is an IATA-like slot procedure for negotiating slot allocation, combined with a secondary market. The IATA process is used at capacity- constrained airports in Europe. There is a considerable base of experience with this process, and most airlines are comfort- able with it. Many argue that this comfort derives from the fact that the process favors incumbent airlines. If, therefore, an air- line decides to adopt an IATA-like slot allocation procedure, it could be encouraged, or perhaps required, to take steps to cre- ate an active secondary market in which airlines can buy and sell slots. The market, if working properly, reduces barriers to entry without forcing incumbent airlines to give up their slots. Secondary markets have existed for high-density airports in the United States for more than 20 years. They have not operated effectively for several reasons. On the one hand, slot holders have been reluctant to sell slots to would-be competi- tors. On the other, buyers have been reluctant to pay for slots because it has been less expensive to secure access to the air- port through other means, such as making certain types of operations exempt from the slot limits. Under the concept presented here, airports could play an active role in lubricat- ing the secondary market, by always having control of an inventory of slots, which it could lease or sell to airlines. The airport could acquire these slots through purchase and could include the acquisition cost in its cost base. It is much easier to employ slot controls to limit growth in flight schedules than to reduce schedules already in place. This again points to the importance of performing demand management pro-actively, rather than waiting for high delays to occur before intervening. 5.7.2 Flexibility in Capacity Allocation Airports undertaking demand management could be able to experiment with any or all of the capacity allocation poli- cies. Implementing such an approach, however, would require changes and clarifications in policy. Three policies are of par- ticular note: nondiscrimination among aeronautical users, the definition of cost centers for purposes of setting airport aero- nautical fees, and airport profitability. 5.7.2.1 Nondiscrimination Across Aeronautical Users In the PACE proceedings, among others, it was argued that a flat fee “discriminates” against small aircraft. A clear policy would have to be articulated that under no circumstances can a flat (undifferentiated by aircraft size) fee for use of aeronau- tical facilities be considered discriminatory. Although the accountability discussion centered on the passenger perspec- tive related to nondiscrimination, the flexibility section would focus on aeronautical users. Aside from the pragmatic consideration of allowing air- ports to include flat fees as part of their demand management programs, there are several ways that this change might be jus- 110

tified. First, from an economics perspective, it is not flat fees that discriminate, but the current weight-based fees. Second, it should be recognized that congestion is a cost, and one that is higher for certain classes of users (e.g., large aircraft opera- tors and scheduled carriers) than others. A congested airport is therefore a discriminatory airport, and policies that mitigate this congestion have the effect of reducing discrimination. Third, when there is severe delay at an airport, the airport is no longer available to users “on reasonable terms.” Thus the principle that an airport should set fees so that “it is available to all users on reasonable terms” could be construed to mean that a capacity-constrained airport could set fees to manage demand and not to be attractive to all potential users. 5.7.2.2 Cost Recovery and Revenue Neutrality Federal law and policy require that airports use revenue generated from their operations for airport-related purposes. The use for other purposes is considered revenue diversion and is prohibited, with certain “grandfather” exceptions for airports with long-standing policies that included certain forms of diversion. Federal policy also requires that charges to aeronautical users reflect the historical costs of the airfield and airport roadways and be derived from some systematic accounting method. The latter policy is elaborated in the FAA Policy Regarding Airport Rates and Charges, first published in 1996. The Policy was revised to permit airports to include in the aeronautical cost base airfield projects that have yet to be completed but for which costs have been incurred and costs for projects at other airports that may be expected to divert traffic from the original one. The revised policy also explicitly authorizes the use of peak surcharges in order to manage demand. Under these policies, airports that employ pricing or auc- tions to manage demand must ensure that their programs are revenue neutral, or at least do not generate revenue in excess of the allowable costs. For example, a peak surcharge would be accompanied by a reduction in the weight-based landing fee, so that the total revenue generated remains the same. It could, however, become impossible to maintain revenue neu- trality while setting surcharges at the level required to induce behavior modification. This is particularly true for airports where demand exceeds capacity for many hours of the day, such as LGA. To address this problem, current constraints on airport rates and charges could be relaxed. This could be done in two differ- ent ways. The first would be to abandon the principle that aero- nautical revenues cannot exceed allowed costs. The policy against diverting revenue from the airport would be main- tained, but airports would be permitted to use aeronautical revenue to cross-subsidize other cost centers, such as terminal facilities, ground access, and security. This would be a dramatic reversal from current practice, where subsidies often go in the opposite direction, but it might be appropriate at a capacity- constrained airport. As already noted, most other airport charges, such as parking and facility leasing, cannot instantiate the principle of charging per flight rather than per unit of pay- load. To raise per-flight charges large enough to effectively manage demand, it might be necessary to increase revenue generated from the class of charges where this is most easily made—those for airfield use. A second alternative would be to expand the class of allow- able costs that airfield charges can be used to recover. The recent changes to the Policy on Rates and Charges point the way here. By allowing airfield costs at other airports to be included in the costs base, the Policy establishes the principle that allowable costs include not only those for supplying nec- essary facilities at a given airport but also those for curtailing demand at the airport. The present policy allows such curtail- ment costs only when they are spent on airfield facilities at substitute airports, but the fundamental principle has much wider application. For example, as discussed in Section 5.2, there are instances when it might be cost-effective to shift short-haul traffic to other modes. Encouraging such modal diversion serves the same end as encouraging intra-modal diversion to other airports. The costs of doing so—for exam- ple, by subsidizing luxury motor coach services from close-in airports—might be recoverable from aeronautical charges. The idea could be stretched further to include costs for main- taining a functioning secondary market for airports opting for operational caps and IATA-like slot allocation. In this case, the cost might be justified because the secondary market allows competition at the airport to be maintained without having to continually expand the airport to serve all comers. These potential options do not challenge the prohibition on revenue diversion, which would be maintained. Theoret- ically, any constraints on how airport revenue is spent can prevent those funds from being put to their highest and best use. And it is perhaps the case that applying airport revenue surpluses to improve urban schools or support other social programs would often serve society better than keeping them within the airport. The negative consequences of this prac- tice, however, could be severe because of airports’ monopoly power. In effect, airports could extract a profit by electing not to expand capacity even when it is cost-effective to do so. If an airport is really unable to find reasonable ways to spend revenue generated through congestion surcharges, it could be required to surrender that revenue into a fund that could be used for aviation projects in other locales. 5.7.3 Setting Operational Limits Most demand management schemes require that opera- tional caps be established. In some instances, the cap pertains 111

112 to the number of operations in an hour or some other time unit. In others, the cap pertains to delay. Depending on the scheme, the cap may be a hard limit or a trigger for some demand management action. In the case of the New York air- ports, for example, there is a hard cap on the number of oper- ations that can be scheduled in an hour. In the Massport plan for BOS, peak-pricing surcharges are triggered when the delay, under Visual Meteorological Conditions capacity, is predicted to exceed a certain threshold (an average 15 min per flight over a 2-hour period). Airports might be given flexibility in establishing these caps. There are many legitimate reasons why the caps could vary from airport to airport, even when they have identical capac- ities. These caps involve complex trade-offs between airport use and delay, and the “sweet spot” in this trade space can vary from place to place. A tourist destination, for example, might place a higher premium on handling large volumes of passen- gers, and therefore flights, and be willing to accept high levels of delay in order to do so. An airport serving short-duration business travel might prefer a more reliable service, even if this entails reduced volume. There is no justification to require a single standard that would apply to both of these cases. Of course, there are also situations where airports could abuse their discretion. On the one hand, they could use con- gestion as a reason to set caps that are really motivated by noise considerations, thereby undermining the compromise established in the Airport Noise and Capacity Act. On the other hand, they might set the caps too high, or not set them at all, because of the revenue that can be realized from the resulting traffic. It would therefore be necessary to determine how little or much traffic and/or delay could be tolerated. Air- ports with delays below a certain level might not use demand management to reduce flight volumes. Airports with delays exceeding a certain level, and who fail to act to address the sit- uation, could be subject to slot controls such as exist today at the New York airports. The range between the high and low values could be a “zone of indifference” in which airport demand management is an option. As previously suggested, this zone of indifference policy could be applied prospectively to address the case of an anticipated surge of demand at an airport that is currently relatively uncongested. There are some who believe that caps could be set too low. In certain European airports, pressure from controllers results in caps well below runway capacity being set. One possible solution would be for airports with slot controls to increase caps slightly on an experimental basis. If the resulting delays were within pre-established limits, the increased caps would become the baseline values; otherwise, the caps would return to their previous levels. 5.7.4 Traffic Flow Management A ground-delay program (GDP) is “implemented to control air traffic volume to airports where the projected traffic demand is expected to exceed the airport’s acceptance rate for a lengthy period of time” (12). Such a program is typically put into place when weather reduces the airport acceptance rate. However, the idea behind a GDP could be leveraged and used for chronically congested airports. It could be possible for a congested airport to request institution of a continuous GDP. With fair-weather capacity as the baseline capacity for the GDP, the GDP would impose ground delays when the demand for operations at the airport exceeded the fair-weather capacity. Imposition of the continuous GDP allows airlines to use tools such as Flight Schedule Monitor to obtain updated information on how their flights would be delayed. In some cases, this knowledge alone might lead to delay-reducing schedule adjustments. Moreover, it would provide enough lead time to adjust their published schedules so that the sched- uled departure time and arrival times would match the con- trolled times imposed by the GDP. In effect, the continuous GDP turns the process of smoothing schedules from a tactical, day-of-operation process to a strategic one. In principle, this could essentially eliminate delays from over-scheduling rela- tive to fair-weather capacity, while also reducing delays under reduced capacity.