Chapter 8 FORECASTING AND TRENDS

Chapter 8

FORECASTING AND TRENDS

Photo cradt: Dorn McGrath, Jr.

Contents

PageAviation Demand Forecasting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159

Methods of Forecasting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159The FAA Aviation Forecasting System. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163Limitations of Aviation Demand Models.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 166

Recent Trends in the Airline Industry. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169Changes in Airline Industry Composition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 170Changes in Route Networks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172Formation of New Hubs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174Aircraft Equipment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 178

Implications for Airport Development . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184

List of Tables

Table No. Page38. FAA Forecasts of Aviation Activity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16439. Summary of FAA Forecasts, 1959-83 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16540. Domestic Airlines by Class of Carrier . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17141. Changes in Stations Served by Air Carriers, 1978-81 . . . . . . . . . . . . . . . . . . . . 17342. Comparison of Hubs Served by Selected Carriers, 1978 v. 1982 . . . . . . . . . . 17343. Aircraft Departures by Hub Size. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17444. Weekly Departures and Seats, by Hub Size, 1978-83 . . . . . . . . . . . . . . . . . . . . 17545. Changes in Weekly Departures Between Major Categories of Airports . . . . . 17646. Present and Future Commercial Jet Aircraft . . . . . . . . . . . . . . . . . . . . . . . . . . . 18247. Aircraft Operated by Commuter and Regional Airlines, 1970-81 . . . . . . . . . . 184

List of Figures

Figure No. Page19. Hub and Spoke Route System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17720. Western Airlines Activity at Salt Lake City . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17921. Composition of U.S. Commercial Jet Fleet, 1982 . . . . . . . . . . . . . . . . . . . . . . . . 18022. Age of Aircraft in Service in U.S. Commercial Jet Fleet . . . . . . . . . . . . . . . . . 18123. Composition of Commuter Aircraft Fleet, 1981 . . . . . . . . . . . . . . . . . . . . . . . . . 183

Chapter 8

FORECASTING AND TRENDS

Prudent management must take into accountfuture events and conditions. Often their naturecan be anticipated by analyzing events of the re-cent past and applying techniques to project theeffects of these trends into the future. The firstpart of this chapter reviews forecasting techniquescommonly used in aviation planning and describestheir use by airport operators, air carriers, and

AVIATION

Methods of Forecasting

government agencies. The second part discussesrecent events and emerging trends in the aviationindustry that will color future forecasts. These in-clude the effects of deregulation, changes in routeand service patterns, and the lingering effects ofthe air traffic controllers’ strike. The final part ofthe chapter speculates on how these trends mayaffect the future needs of airports.

DEMAND FORECASTING’

An aviation demand forecast is, in essence, acarefully formed opinion about future air traffic.Its primary use is in determining future needs orestimating when they must be met. Any of sev-eral methods may be used, with results that willvary widely in terms of scope, time scale, struc-ture, and detail; but they have certain commonfeatures. Chiefly, forecasts are derived from as-sumptions about the relationship of the past andthe future in that they postulate that certain meas-urable historical events or conditions have acausal or predictive relationship with events orconditions that will be of interest in the future,Analysis of these historical factors—usually bysome sort of mathematical manipulation of data—allows the forecaster to express expectations interms of some measure or index of aviation activ-ity. From this initial product (e. g., expected pas-senger travel, cargo volume, or aircraft opera-tions) the forecaster can derive further estimatesof the nature, magnitude and timing of futureneeds for equipment, facilities, manpower, fund-ing, and the like. Even though the method usedmay be quite rigorous and mathematically com-plex, forecasting is inherently a judgmental proc-ess where uncertainty abounds. The best that theforecaster can achieve is to be aware of his biases,to identify the sources of uncertainty, and to esti-mate the probable magnitude of error.

‘This section is based in part on a paper prepared for OTA byDavid W. Bluestone, John Glover, Dorn McGrath, Jr., and PeterSchauffler.

In setting out to prepare a forecast, the fore-caster has at his disposal two basic types of in-put data. He may choose data on aviation activ-ity itself and use historical performance trends toproject future activity. In effect, this approachassumes that the best predictor of future aviationdemand is past aviation demand. Alternatively,the forecaster may choose data related to underly-ing economic, social, and technological factorsthat are presumed to influence aviation demand,treating them as independent variables that canbe used to predict demand as a dependent vari-able. Among the factors that may be so used are:

●

●

●

basic quantitative indicators, such as popula-tion, gross national product (GNP), activityof certain sectors of the economy, personalconsumption expenditures, or retail sales;derived socioeconomic and psychological in-dexes, such as propensity to travel, incomeclassifications, employment categories, edu-cational levels, or family lifestyles; andsupply factors, such as fare levels, aircraftcharacteristics (size, speed, and operatingcosts), schedule frequency, or structure of theair carrier industry.

The outputs of the forecast are measures ofaviation activity—passenger enplanements, rev-enue passenger-miles, freight ton-miles, numberof aircraft in the fleet, or number of aircraft oper-ations. Other output measures, such as air car-rier revenue, air traffic control (ATC) workload,and demand for airport facilities can be derivedfrom these estimates.

159

I& I . Airport System Development

The range and scope of forecasts can varygreatly, depending on the purpose they are toserve. They might include all aviation or be lim-ited to a particular type of traffic (passenger orcargo) or a particular type of operator (scheduledair carrier, charter, or general aviation). The geo-graphical scope may be international, nationwide,regional, or limited to a particular market orairport.

The forecasting horizon may range from a fewmonths to 20 years, again depending on the pur-pose of the forecast. Airlines, for example, tendto use very short-term projections of traffic in or-der to estimate their financial or staffing needs ona quarterly or semiamual basis. Airport planners,on the other hand, use very long-range forecasts,on the order of 20 years, as a basis for major deci-sions relating to land acquisition and airport de-velopment. Between these extremes, forecastinghorizons of 1,5, or 10 years are common for plan-ning changes and improvements of airport facil-ities, estimating ATC workload, projecting airearner fleet requirements, and financial planning.

There are two basic approaches to aviation de-mand forecasting— “top-down” or “bottom-up.”The top-down approach begins with the largestaggregates of economic and statistical data (usu-ally national totals) and seeks to provide a gen-eral picture of aviation demand spanning thecountry and the entire system of air travel routesand facilities. Once the aggregate forecast has beendeveloped, portions of the total volume of traf-fic can be allocated to specific industry segmentsor geographical regions based on historical sharesor assumed growth rates.

The bottom-up approach, in contrast, beginswith data for a specific geographic area and de-velops a forecast of aviation demand at a particu-lar airport or in a metropolitan region, typicallyas an indicator of need for building or expandinglocal facilities. Where good data are available andthe economy of the region is developing in an or-derly way, this approach can closely approximatethe reality of the area under study. In some cases,a number of such bottom-up forecasts may becombined to make a composite forecast for alarger area, but this approach of building up aregional or national aggregate from many local

forecasts can lead to difficulties. For example,forecasts for some areas may be overly optimistic—often a defensive strategy designed to assureadequate future capacity. It is not unusual to findthat the sum of many such bottom-up forecastsexceeds the top-down forecast for the region bya wide margin.

Whether “top-down” or “bottom-up,” aviationdemand forecasting as practiced today uses a widevariety of methods. The attributes, limitations,and typical applications of these methods are dis-cussed below.

Time Trends

A simple forecasting method is the extrapola-tion from the past, where the forecaster assumesthat major trends, such as traffic growth or mar-ket share, will continue uninterrupted and thatthe future will be like the recent past. Historicaldata for some base period are gathered and ana-lyzed to determine a trend line, which is then ex-tended to some point in the future, using eithersophisticated mathematical procedures or simpleestimation of the most likely course. This methodis often used for short-term projections (1 or 2years) where basic conditions are unlikely tochange much. It is also better than no forecast atall in cases where a data base suitable for moresophisticated methods is not available. However,a basic shortcoming of trend extrapolation is thatit does not take into account underlying economic,social, and technological factors that affect avia-tion and that are themselves subject to change.

Econometric Models

The econometric model is by far the most fre-quently used method for forecasting aviation de-mand. It is a mathematical representation of airtraffic or its constituent parts and those independ-ent variables of the national economy which arethought to influence traffic growth. Econometricsis the statistical technique used to quantify theserelationships. The mathematical equations of themodel relate economic factors to the level of avia-tion activity, based on observation of past be-havior of both the economy and the aviation in-dustry. The equations may also be constructedso as to reflect the effects of specific factors within

Ch. 8—Forecasting and Trends “ 161

*4 - - . . , “ - ’

Photo credit: Federal Aviation Administration

Newark: the alternative to La Guardia and Kennedy

the air transportation industry itself, such as farelevels, route configurations, fuel costs, etc.

Among Federal Government agencies, both theFederal Aviation Administration (FAA) and theCivil Aeronautics Board make extensive use ofeconometric forecasting methods. Econometricmodels are also used by airlines, industry asso-ciations, and aircraft manufacturers. TWA, forexample, employs a set of econometric models toforecast passenger travel industrywide and, fromthat, TWA’s prospective market share. The Asso-ciation of European Airlines uses a mathematicalmodel in which traffic varies directly with grossdomestic product and inversely with average rev-enue per passenger. McDonnell Douglas, Boeing,and Lockheed all have their own versions ofeconometric models to project future sales of air-craft. The equations for the McDonnell Douglasmodel, for instance, include the ratio of long- toshort-term interest rates since the cost of borrow-ing money has an effect on the ability of airlinesto purchase aircraft.

Gravity Models

The gravity model was first developed in thesociological and marketing fields to describe var-ious forms of human interaction. The techniquewas later adapted by traffic engineers to describetravel behavior. It is predicated on the assump-tion that travel behavior obeys a law analogousto the law of gravity, in that attraction betweencities varies directly with population and inverselywith distance. Thus, two large cities located nearone another have a strong mutual attraction andform a very dense transportation market; smallcities located far apart have little travel betweenthem. The gravity model uses socioeconomic datafor each pair of metropolitan areas to predict thelevel of transportation activity between them. Theequations often contain terms to describe thespecial attractiveness of each city for differenttypes of personal and business trips.

Although gravity models have been used ex-tensively in highway planning, their use for avia-tion forecasting is limited. The State of Califor-

— — —

162 . Airport System Development

nia uses a gravity model in its State AirportSystem Plan in an effort to give a statewide “sys-tem view” of air transportation. The Californiagravity model takes into account changes in popu-lation, employment level, and income of majormetropolitan areas to produce estimates of thetravel that will be generated between various partsof the State. To provide consistency among plansfor all transportation modes, a similar gravitymodel incorporating the same socioeconomicvariables is used for other transportation forecast-ing within the State.

Scenarios

The scenario method isstrate the variation due to

often used to demon-differing assumptions

about future conditions, thus bracketing the-rangeof uncertainty. The values of input variables inan econometric model, for example, are in them-selves simply guesses about the future behaviorof the economy. Rather than depend on a single“best” estimate of GNP in future years, the fore-caster may elect to construct several scenarios topredict the behavior of the aviation industry undera range of likely economic conditions. FAA beganusing this method in 1976 in an attempt to describeconditions that could affect the future of air trans-portation, and most FAA forecasts since that timehave included different scenarios incorporatingdivergent assumptions about the economy and theairline industry.

One of the drawbacks of the scenario methodis that the range between high and low estimatescan be so large that the forecast loses practicalvalue as a guide to planning. For example, in theinitial 1976 FAA study, where five scenarios wereused, the high estimate of revenue passenger-mileswas 2.3 times the low estimate, and the ratio ofhigh to low forecasts of aircraft operations was2.9.

Ratios

Some local aviation authorities and industrygroups make forecasts by the relatively simple ex-pedient of assuming a ratio between national “top-down” traffic forecasts and their own segment oftraffic. This method is often used by airports thatlack the funds or expertise to make independent

econometric forecasts. A notable application wasin 1969, when the major U.S. air carriers devel-oped a national forecast on a consensus basis andthen allocated portions of the traffic to each of22 major air transportation hubs. The allocation,based on the historical share of national trafficcaptured by each hub, was adjusted by expertjudgment to account for shifting patterns of air-port use.

Market Surveys

This method has been used extensively by thePort Authority of New York and New Jersey forthe past 25 years. The Port Authority uses in-flightpassenger surveys to gather information on pointof origin, choice of airport in the metropolitanarea, choice of ground access mode, ground ac-cess travel time, destination, purpose of trip, andother factors that can be used to predict travelbehavior and consequent demands on aviation fa-cilities. These data are classified in a travel mar-ket model made up of over 100 socioeconomic“cells” defined by age, occupation, income, andtrip purpose. The growth rate for each cell is pro-jected by straightforward econometric ”techniques.

The market survey method, while it producesa highly detailed forecast of travel, has some sig-nificant drawbacks. Data collection is compli-cated, time-consuming, and expensive. Since thesample is collected in a relatively brief period, itmay not be truly reflective of long-term travel pat-terns and preferences. Airlines, which serve as col-lectors of the data, are reluctant for competitivereasons to relinquish control of survey resultswhich they consider proprietary.

Judgment

To some degree, judgment enters into all fore-casting. Even the most formal and scientific fore-casting methods require that assumptions be madeabout future conditions and events. These as-sumptions, which represent the forecaster’s basicoutlook, are simply informed judgments, and theycan have a powerful effect on the outcomes. Judg-ment also enters into the forecasting process inother ways: on the methodology to be employed,on the trends to be assumed, on the selection ofyears to use as a base period, on the choice of data

Ch. 8—Forecasting and Trends ● 163

Photo credit: federal Aviation Administration

Commuter airline in Alaska

sources, and on likely changes in specific factorssuch as fuel availability, cost, and technology. Atthe completion of a forecast it is not uncommonto subject the results to the test of expert judg-ment and to adjust them in the light of what seems“reasonable. ”

Application of judgment has, in at least twocases, become institutionalized as part of the fore-casting process. U.S. airlines generally use econo-metric models for traffic forecasts and fleet plan-ning; but since they do not agree on method andinitial assumptions, the Air Transport Associa-tion (ATA) develops a consensus forecast basedon the judgment and practical experience of air-line personnel and the ATA forecasting workinggroup. The International Air TransportationAssociation (IATA) uses a modified “Delphi” tech-nique to produce forecasts for international pas-senger and freight traffic. Delphi is a method forattaining consensus among experts, in this casethe forecasters from participating IATA memberairlines. Using this technique, initial estimates areobtained from each expert. These estimates arearranged in a composite that shows each partici-pant how his forecast compares to the group asa whole, and each is invited to submit anotherforecast based on this information. After one ormore rounds of comparison and feedback, judg-ments begin to converge, and a consensus fore-cast is reached.

The FAA Aviation Forecasting System

The most elaborate aviation demand forecastsproduced in this country are those of the Federal

Aviation Administration. They consist of na-tional, regional, and individual airport forecaststhat typically cover a 12-year period, although20-year forecasts are sometimes prepared. Theseforecasts, updated and issued annually, providethe basic context for aviation demand forecast-ing in the United States. They are used, with avariety of specialized interpretations, by all ele-ments of the aviation community.

In addition to the basic annual forecasts, FAAalso publishes special studies and forecasts fromtime to time. Subjects covered recently have in-cluded air cargo activity (1979), commuter air-line activity (1977 and 1981), and forecastingneeds at the State level (1979). FAA has also pub-lished special “profile” reports on hourly airportactivity, air carrier operations, and internationalpassengers. In 1978, FAA began a series of in-dividual forecasts for 24 large hub airports. Theseare adaptations of other FAA forecasts, withspecial sections on local economic growth, pas-senger enplanements, cargo and mail enplaned,general aviation (GA) and air carrier aircraft oper-ations, and traffic handled by FAA towers.

FAA National Forecasts

Each year FAA publishes a national forecast en-titled FAA Aviation Forecasts. The most recentedition (released in February 1984) includes de-tailed year-by-year forecasts from 1984 to 1995for air carriers, air taxis and commuters, GA, andmilitary aviation. It also contains workload fore-casts for airports with FAA control towers, airroute control centers, and flight service stations.

The 1984 forecasts anticipate that enplanementsby major airlines will grow at an average annualrate of 4.6 percent. Larger aircraft and higher loadfactors will minimize actual increases in opera-tions to accommodate this growth, with the re-sult that FAA projects air carrier operations togrow by no more than 1.7 percent per year. Largergains are expected for commuter carriers, whoseenplanements are expected to increase by 7.4 per-cent per year and operations by 4.7 percent peryear. GA operations are expected to increase by6.0 percent annually. The current FAA forecastsare summarized in table 38.

164 Airport System Development

. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .

.

. .. .. .. .. .. .. .. .. .. .. .. .“ -“ ( n. -

.L . .a). .

x

Ch. 8—Forecasting and Trends 165

As part of the documentation for these annualforecasts, FAA sets forth the basic assumptionsconcerning the industry, government, and eco-nomic environment for the forecast period. Theprincipal indicators—gross national product,Consumer Price Index, and fuel price index—arecomposites of estimates obtained from four lead-ing nongovernmental economic forecasting orga-nizations: Chase Econometrics, Data Resources,Evans Economics, and Wharton EconometricsAssociates. FAA believes that this consensus ap-proach to formulating input assumptions lendsgreater credibility to the forecasts.

The air carrier portion of the forecast is devel-oped in several steps. For airline travel, FAA firstforecasts passenger yield (cost per passenger-mile)based on estimates of three independent variables:jet fuel prices, average airline wages, and avail-able seat-miles per aircraft. The next step is toforecast passenger demand, based on GNP andyield. Third, FAA develops forecasts of aircraftoperations based on load factor, average seats peraircraft, and passenger trip length, all of whichare estimated in consultation with industry ex-perts. Forecasts for itinerant operations, instru-ment operations, and other FAA workload meas-ures are developed from the basic forecast of totalaircraft operations, using empirically derived rela-tionships.

Past FAA forecasts have often been criticizedfor inaccuracy or unrealistic assumptions aboutfuture growth of aviation. An examination of thisquestion was made in a recent CongressionalBudget Office study of FAA forecasts since 1959.2

CBO divided the forecasts into three distinct peri-ods (see table 39). CBO found that the forecastsperformed between 1959 and 1965 were consist-ently low by an average of almost 19 percent.From 1966 to 1973 the forecasts swung sharplythe other way and consistently overestimated thegrowth of aviation activity by nearly a third.From 1974 on, which coincides with the time thatFAA has been using more sophisticated econo-metric modeling techniques, the results have beenmixed, sometimes too high and sometimes toolow. Overall error has averaged about 21 percent,somewhat smaller than in the previous period butstill rather large for this type of forecasting. FAAforecasts for the GA sector have been especiallyunreliable and consistently high, sometimes by asmuch as 50 percent. On balance, CBO concludesthat FAA’s forecasts have improved substantiallyin the past 10 years, showing a reasonably smallrandom error instead of the constant high or lowbias that characterized forecasts of the earlier twoperiods.

Table 39.–Summary of FAA Forecasts, 1959-83

Periods in whichforecasts made Method Performance 5 years ahead Market environment

1959-65 Trend forecasting: unspecified Average error – 18.7 percent Expanding, prosperous economy.links to economy, business Worst year –32.5 percent Rapidly growing population.cycle, population, fares, Declining first-class and coachcompetition from other modes fares, (declining unit costs

because of increasing use ofjets).

1966-73 Trend forecasting: unspecified Average error +32.5 percent Softening trends in aviationlinks to economy, business Worst year +58.4 percent activity. Increasing ticket taxes,cycle, population, fares, rising fares. Forecasts made incompetition from other modes 1969 (published January 1970)

assumed 4.25 percent growthrate in 1973, to continue at thatrate through decade. Inflation 2percent per year from 1973.

From 1974 Linear econometric models Average error +21.2 percent Airline deregulation, economicWorst year +34.7 percent recession, fare wars, and

166 ● Airport System Development

Terminal Area Forecasts

FAA Terminal Area Forecasts (TAFs), like FAAnational forecasts, are developed annually. TheTAF data base contains descriptive informationand forecasts for about 4,000 airports-the 3,200eligible for Federal aid and about 800 other public-use airports. Each airport record includes at least5 years of historical data and a 12-year forecastof aircraft operations, broken down into air car-rier, commuter, GA, and military categories. Theprojections for individual airports can also be ag-gregated to form State and regional forecasts.

The TAFs constitute a subroutine of the basicFAA “top-down” forecasts of national aviationdemand. The process for developing TAFs hasbeen refined in recent years, and they now serveas the basic frame of reference for other types of“bottom-up” forecasts undertaken by many localairport authorities and State and regional agen-cies. Not all local airport authorities, however,accept the validity of TAFs, which they believedo not adequately take into account the factorsaffecting aviation demand at the local and regionallevel and which are not, in their view, developedthrough appropriate consultation with local au-thorities.

The Special Problem of GA Forecasting

FAA forecasts GA demand in two segments—business aviation and personal flying. For busi-ness aircraft, the forecast is based on the real priceof aircraft, interest rates, and measures of busi-ness activity such as manufacturing and retailsales. For personal flying, the factors used are air-craft price, interest rates, and GNP as a measureof income. Itinerant operations are forecast as afunction of the size of the fleet and the real costof fuel. Instrument operations are a function offleet size. Local operations, predominantly train-ing operations, are a function of the number ofstudent pilots and the number of aircraft in thefleet. FAA has concluded that because of large andsomewhat unpredictable oscillations in all thesevariables, econometric models do not producereliable forecasts of general aviation. As a prag-matic approach, FAA uses modeling only as apoint of departure to produce first approximationsthat are subsequently adjusted with data from

periodic surveys and estimates from FU regionaloffices and industry representatives.

Some observers are of the opinion that FAA’sgeneral aviation forecasts are unrealistically high.For example, recent FAA forecasts estimate thatthe GA fleet will grow at a rate of 3.3 percent peryear for the remainder of this century.3 Thismeans an expansion of the GA fleet from about210,000 aircraft in 1983 to 269,000 in 1990 and385,000 in 2000. Such a growth rate is extremelyoptimistic, and realizing it would require an un-precedented level of manufacture and sale in theGA aircraft industry. An increase of 175,000 air-craft in the GA fleet by the end of the centuryis equivalent to adding 10,000 new aircraft peryear—13,000 if allowance is made for replacementof existing aircraft at the rate of 1 percent an-nually. When foreign aircraft sales are taken intoaccount, the FAl forecast implies that U.S. firmswould manufacture and sell about 16,000 aircraftper year between now and 2000. This seems unlikelyin light of performance over the past 15 years inthe GA aircraft manufacturing industry wheresales (including exports) have averaged only two-thirds of this amount. A

Limitations of AviationDemand Models

Aviation demand models can be very usefulforecasting tools, but it is important to recognizetheir limitations. First, all models are necessarily

incomplete. They attempt to reduce a large andcomplex system to a relatively few mathematicalequations that describe the most important in-teractions. Many factors must be left out, eitherbecause they are difficult to formulate mathe-matically or because including them would makethe model cumbersome and too complicated touse. There are other factors excluded not by de-sign but by inadvertence because, with the pres-ent state of knowledge about the relationship be-tween aviation and underlying economic andsocial forces, we are simply unaware of all thefactors that drive demand for air transportation.

3National Airspace System Plan (Washington, DC: Federal Avia-tion Administration, April 1983, revised edition), p. II-2.

4Aerospace Facts and Figures, 2983/84 (Washington, DC: Aero-space Industries Association, 1984), p. 33.


To guard against such structural weakness, themodel builder calibrates and tests the model byinserting data from past years to see if the his-torical record can be accurately reproduced. If themodel is well constructed and its mathematicalrelationships are a good representation of reality,using data from some past year in the equatiorlswill yield a forecast of the aviation activity thatactually occurred. Such testing gives the forecasterconfidence that for some future year the modelwill correctly predict aviation activity if the cor-rect values of input variables are used. Unfortu-nately, the correct values of input variables suchas GNP or interest rates for any future year arethemselves unknown, and uncertainty is simplytransferred from the behavior of the aviation in-dustry to behavior of the economy at large.

Models tend to assume that the future will bevery much like the past. If the real world situa-tion changes substantially, the model is corre-spondingly less accurate. Such changes might in-clude sudden economic perturbations, such as thefuel crisis of 1973, or longer term restructuringof the market or the economy. For example, avia-tion models constructed to predict the behaviorof a regulated industry with stable fares and routestend to be less accurate now that price competi-tion and freer entry into new markets are per-mitted. Further, because models are, at best, onlypartial representations of the world, it is not easyto predict which changes in travel behavior or eco-nomic conditions will be important in the futureor how they should be incorporated in a model.A major problem among forecasters is discrimi-nating among relationships that will persist andthose that will not.

From this, it is clear that aviation demandmodels are highly influenced by underlying as-sumptions about economic and social trends andfuture conditions. The model itself may accuratelydepict relationships between air transportationand the state of the economy or the structure ofthe aviation industry, but if the assumed statesof these variables at some future time are too op-timistic, too pessimistic, or simply inconsistentwith the course of events, the resulting forecastscan go far astray. It is probably fair to concludethat, even with the present limitations of themodel builder’s art, the inaccuracies due to the

structure of forecasting models are generallysmaller than those induced by erroneous inputassumptions. An aviation demand model is nomore robust than the assumptions on which itrests, and assumptions (usually a matter of ex-pert judgment) are the most fragile part of theprocess.

The limitations, biases, and characteristics ofa forecast may depend as much on who is doingthe forecast as on the particular method beingused. Airport authorities, aviation agencies, air-lines, and industry associations all make forecaststo help them plan for the future and to help themjustify plans and programs. There is, in manycases, a natural inclination to err on the side ofoptimism in order to protect the future interestsof the agency producing the forecast. Thus, localairport authorities planning an expansion projectmay tend toward an unduly high appraisal of theoverall growth prospects in the local economyand, hence, future passenger and cargo traffic. Asa consequence, basing decisions to construct orexpand facilities on these forecasts may lead toexcess capacity or premature investment of capi-tal. However, this may be less detrimental thanrelying on a forecast which is too low. Slight over-capacity and anticipation of demand is viewed bymost airport planners as preferable to congestion,delay, and perhaps deterioration of service andsafety.

Local forecasting may not take into accountbroader regional trends and conditions. The cur-rent shifts in population and economic activity

Photo credit: Federal Express

Air cargo hub

25-420 0 - 84 - 12


from the “Frost Belt” to the “Sun Belt” area casein point. States in the South and West are expectedto experience much greater growth in aviationactivity in the next few decades while communi-ties in the North and East will tend to decline.However, this trend—with its consequences forairport planning and development—is not yet gen-erally recognized in the forecasts prepared by“Frost Belt” communities, perhaps because theyare unwilling (or find it politically unwise) to goon public record as predicting their own decline.

The tendency of the aviation community toorganize categorically—air carriers, commuters,general aviation, helicopters, etc.—also influencesaviation demand forecasting, both the forecastsprepared by such groups themselves and thoseprepared by others who seek to anticipate thedemands that each sector will place on the air-port and air traffic control system. All of thesegroups compete in some measure for the scarceresources of airspace and airport facilities, andnone is prepared to concede that its own require-ments are less pressing than those of others or toadmit any scenario other than continued growth.As a result, each sector tends to publicize its ownaspirations lest it lose out in the general competi-tion and find its access foreclosed. With sufficientrepetition and vigorous advocacy, such declara-tions become accepted as the reality of futuredemand.

For example, the President of the National Busi-ness Aircraft Association, Inc., asserted in mid-1982 that “a great pent up demand for aircraft isbuilding all the while the recession continues.Once the general economic climate begins to im-prove and the price of borrowing money declineseven modestly on a long-term basis, we willwitness marketplace activity on a scale neverbefore known.”s While this bold prediction maybe arguable, it reflects a natural defensive strat-egy. If the forecast turns out to be true, generalaviation, especially business aviation, will soonbecome the majority user of the airport and ATC

system. In this event, business aircraft operatorswould be unwise to allow air carriers to preemptlanding slots or to accept restriction or diversionof their activities to accommodate air carriers. Onthe other hand, if the forecast proves too op-timistic, decisions based on it could result in seri-ous overcapacity or misallocation of resources.

No forecast is any better than its input data,and a little foreseen consequence of deregulationis that forecasting may become more difficult andless accurate in the future because of the lack ofa detailed and adequately maintained data base.The Airline Deregulation Act of 1978 calls for agradual phaseout of the Civil Aeronautics Board(CAB) by 1985. At that time, all functions of theCAB will either be eliminated or transferred toother agencies. It is still not clear how much ofCAB’s extensive data collecting activities will betransferred and continued. Since nearly everyonein the aviation community makes use of CAB datafor forecasting purposes, the prospective loss ofthis resource is now the focus of extreme concernin industry and government. Not all parties agreeon which parts of the CAB data base should bemaintained and who should be responsible, butthere is apparent consensus on at least four ma-jor points:

●

●

●

●

CAB data have become a crucial part of theaviation forecasting process.Continued collection of at least the basic dataon air carrier and commuter operations isvital to intelligent analysis, interpretation,and forecasting of regional and local avia-tion demand.Without such data, the reliability of forecast-ing for air carrier and commuter activitycould decline, perhaps to a level no betterthan current general aviation forecasting.Forecasting aviation demand could becomemuch more difficult after 1985.

The prospective loss or drastic reduction of thewidely used and respected CAB data base loomsas only the latest example of an unexpected ma-jor event in the uncertain world of aviation de-mand forecasting.


RECENT TRENDS IN THE AIRLINE INDUSTRY6

Of all the forces acting on civil aviation at thepresent time, that which has the most profoundeffect—and which is perhaps the least understoodin all its ramifications—is the recent deregulationof the airline industry. For 40 years, from 1938to 1978, CAB exercised broad powers over theairline industry, controlling entry and exit of car-riers, markets served, and fare structures. Al-though this was a period of great growth in theairline industry, the pace of change in patternsof service and airline route structure was slow be-cause it was tempered by the regulatory process.CAB proceedings on route awards or fare changestook months or years to complete. The CAB ofteninterpreted “public convenience and necessity” inlight of the need to maintain financial stabilityamong existing carriers. Although CAB progres-sively increased the level of competition, this proc-ess was usually accomplished by extending theoverlap of routes and services among existingcarriers.

The Airline Deregulation Act set a timetable forphasing out CAB statutory authority over a 4-yearperiod. The major provisions of the act, effectiveimmediately, relaxed CAB authority over routes andfares and made it easier for carriers to enter newmarkets or to reactivate dormant routes. Except inlocalities qualifying for essential air service, airlinesbecame free to terminate service to a communityby means of a simplified notice procedure. In addi-tion, carriers were allowed to change fares withina broad “zone of reasonableness” determined by air-line costs. Deregulation thus set the stage for a widevariety of changes in the way air services are offered.

While some of the effects of deregulation arenow apparent, the full impact of these regulatorychanges on the market cannot yet be evaluated.Air carriers have been operating in a deregulatedenvironment for only a little over 5 years, andthey are still in a “shake-down” period of adjustingto new freedoms and competitive pressures. Atthe same time, other major factors—escalatingfuel prices, the recession, and the Professional AirTraffic Controllers Organization (PATCO) strike–

bThis section is based in part on a report prepared for OTA bySimat, Helliesen & Eichner.

have had their own effects on the airline indus-try, thus distorting the view of what deregulationhas actually produced.

In general, the period 1978-81 marked a sharpincrease in all airline costs, but no other costescalated as much as fuel, which rose from $0.39per gallon in 1978 to $1.04 in 1981. Fare flexibilityprovided by deregulation enabled some carriersto blunt the effects of increased fuel costs byquickly passing on the resulting higher costs toconsumers. However, the deep recession, begin-ning with the first quarter of 1980, compoundedthe problem as the industry was caught in theposition of needing to raise fares at a time of gen-eral economic decline.

The full effects of airline deregulation weremuted if not altered by the recession, which beganabout 15 months after deregulation and has con-tinued to disturb the U.S. economy. The reces-sion has had a major effect on the airline indus-try, whose health has always been closely linkedto GNP. In the period 1976-78, before the reces-sion, real GNP increased by 6.7 percent per an-num. During the period immediately followingderegulation, GNP grew at less than 1 percent and

Photo credit: Federal Aviation Administration

Controller’s view of a hub


actually declined during 1982. The overall eco-nomic performance of the airlines showed a cor-responding slump, although some airlines faredbetter than others and a few even managed to in-crease their profitability.

Some analysts argue that service and fare stim-ulation following deregulation, combined with theability to drop unprofitable routes, cushioned theeffects of the recession on the air carriers. Highinterest rates and poor business conditions clearlyblocked the start of several new carriers and in-directly provided a measure of protection for ex-isting carriers.

In effect, the air traffic controllers’ strike inAugust 1981 reintroduced regulatory limits on theindustry. The strike led FAA to close 58 ATCtowers at small airports and to impose a cap onoperations at the 22 busiest airports. To allocatethe hardship equitably, FAA required each air-line to reduce operations proportionately at those22 airports. An airline could exercise full latitudeto select routes, so long as it had operating rights(slots) at the capped airports. Slots could be tradedand, for a limited time, even sold. But if an air-line did not use a slot, it was forfeited. Thesestrike-related restrictions have since been lifted,but their effects linger. Because operating rightshad considerable value in a depressed, strike-ridden market, FAA restrictions forced many car-riers to hold on to slots and postpone or cancelplanned route changes. Equally important, the capon operations limited opportunities for new car-riers to enter many of the major markets.

In many respects, the strike slowed the proc-ess of deregulation, but it also stimulated forma-tion of new hub-and-spoke patterns of operationas air carriers sought to increase the number ofpassengers handled by using larger aircraft at theconstrained airports or by transferring operationsto new regional hubs at less crowded airports.

The ultimate outcome of deregulation is notyet clear. The effects apparent so far—althoughblurred by the impacts of increased airline costs,the recession, and the air traffic controllers’ strike—suggest that profound changes are taking placein the structure and economics of the airline in-dustry, with repercussions that may persist forseveral years.

Changes in Airline Industry Composition

Since deregulation, the number of airlines hold-ing CAB operating certificates has increased ten-fold. At the end of 1978, there were only 36certificated carriers; now there are 355. To accom-modate these changes, CAB has devised a newclassification system that categorizes airlines asmajor carriers, national carriers, and large ormedium regional carriers on the basis of theiroperating revenue. This system is not wholly com-patible with two-way classification of trunk andlocal service carrier in use before 1978. (Table 40compares the new and old classification systems. )For purposes of this discussion, it is more con-venient to use the pre-1978 categories, augmen-ted by two additional groups (startup jet carriersand commuter airlines) in order to show thechanges that have taken place in air carrier routesand services since deregulation.

Before deregulation, the 10 trunk carriers dom-inated major domestic U.S. air service markets.Even today, these airlines account for about 60percent of total domestic enplanements and aneven greater proportion of revenue passenger-miles, revenue, and fleet capacity. The trunk air-lines operate large fleets of jet aircraft, of whichnearly 20 percent are widebody. Trunk carriersare equipped to serve primarily the long-haul mar-kets, emphasizing direct service between most ma-jor domestic markets.

Six local service airlines operate short-haul jetservice in domestic markets. In 1982, these car-riers enplaned 19 percent of all domestic air pas-sengers. Through expansion and mergers, localshave grown substantially in size and now threeof these—USAir, Republic, and Frontier—arelarger than some trunk carriers. Still, there aresubstantial differences between local service andtrunk carriers. Local service carriers operate pre-dominantly narrowbody jet aircraft, typicallywith 100 to 125 seats. The average stage lengthand passenger trip length for local service airlinesis less than half that of trunk carriers.

The “startup jet carriers” include former in-trastate carriers that have expanded to nationwideservice since deregulation, as well as new firmsthat began operation since 1978. Previously, therewere four intrastate carriers that provided short-


Table 40.—Domestic Airlines by Class of Carrier

M u s e -

N e w Y o r k A i r

A i r M i d w e s t

Air I l l inois

A i r W i s c o n s i n

P e o p l e E x p r e s s

Medium regionaisAspenCascadeEmpire

(sampie):

Golden WestMississippi ValleyWright

Charter carriers (sample):CapitolWorld

MidwayMuseNew York AirPacific SouthwestSouthwestPeople Express

Commuters (sampie):Air Midwestb

Aspenb

Wright b

EmpireAir “IllinoisAir Wisconsinb

Golden WestMississippi Valley

%Iassifications are based on carriers’ annual operating revenues.bceflificated before deregulation.

SOURCE: Federal Aviation Administration.

haul jet services entirely within the boundaries ofCalifornia, Texas, or Florida. After deregulation,these airlines obtained CAB certification and nowfly interstate routes. Even though they have ex-panded their operations to new out-of-State mar-kets, they retain certain pre-1978 characteristicsin that they continue to serve highly competitiveshort-haul markets with emphasis on frequencyof flights and low fares.

Several other new jet airlines have begun oper-ation since deregulation. Some—e. g., New YorkAir, Midway, People Express, and Muse—are

new ventures. One—Empire Airlines—is a formercommuter airline that has successfully introducedjet service. Capitol Air and World Airlines areformer charter operations that have inauguratedscheduled service. Typically, the startup carriersserve high-density, long-haul markets where theycompete with trunk carriers, principally on thebasis of low fares.

Commuter airlines usually do not operate jets,but propellor-driven aircraft with up to 30 seats.Before 1978, most commuter airlines were exemptfrom CAB fare and route regulation, but they


were restricted to the use of aircraft no larger than30 seats. A few—e.g., Air Midwest, Air NewEngland, and Wright-obtained CAB certificationin order to use larger aircraft or to receive sub-sidies for air service to small communities. Aslocal service and trunk carriers have withdrawnfrom smaller markets, commuters have moved into fill the gap. One of the earliest and most suc-cessful examples is Allegheny Airlines (now USAir),which transferred certain of its services to 12smaller independent airlines (known collectively

under the name “Allegheny Commuter”) thatoperate under contract to USAir.

Since 1978, over 300 commuters have obtainedCAB certification. They are becoming progres-sively integrated into the national air transpor-tation system, providing local point-to-point serv-ice and linking small communities with the largerairport hubs. Passenger enplanements on com-muters have grown rapidly, from 4.2 percent ofpassenger enplanements in 1979 to 6.3 percent in1983. 7

Changes in Route Networks

Airline deregulation has changed the nationalair service network from a stable system of routesserved by established carriers to a fluid market-place where carriers frequently adjust routes, levelof service, and fares. The older airlines have aban-doned some markets and begun service to others.New entrants have taken over some of these aban-doned routes and established themselves in thedense markets where they see a competitive op-portunity.

The course of the industry since deregulationhas been one of uneven expansion and contrac-tion. The principal effect on airports is that sud-den and less predictable changes have taken placein the air carriers serving the airport, the level ofservice provided, and the facilities needed to ac-commodate them. Some communities have ex-perienced a general improvement in air servicesince deregulation, but not all have benefited froman unregulated environment, and some have suf-fered almost complete loss of air service.

‘FAA Aviation Forecasts, Fiscal Years 1984-1995 (Washington,DC: Federal Aviation Administration, FAA-APO-84-1, February1984).

Table 41 shows changes in points served by asample of carriers between 1978 and 1981; table42 shows the changes in service by size of city.The designations large, medium, small, and non-hub are FAA and CAB classifications based onnumber of passengers enplaned. Large hubs arethe top 24 cities, medium hubs the next 39. Thereare 61 small hubs and 461 nonhubs. s

Most of the established carriers added anddeleted service points frequently during this period.Air California and Air Florida, no longer re-stricted to intrastate traffic, expanded the num-ber of points served. The Allegheny Commutersystem of USAir abandoned 23 stations and added30. American Airlines moved into Braniff’s mar-ket, adding 12 stations in the Texas, Louisiana,and Alabama markets. Frontier Airlines discon-tinued service to 28 nonhubs and small hubs andbranched out to other small cities and a few largehubs. Piedmont pursued the same strategy, ad-ding seven large cities and two high growth areasin Florida, while deleting 16 small hubs andnonhubs.

Competition has been intense in certain highgrowth markets. For example, cities like Phoenixwith substantial population growth have attractednew carriers. Traffic growth at Orlando, whichis now served by 10 carriers as opposed to 4 beforederegulation, is a product of both economic de-velopment (Disney World and Epcot) and a surgeof discretionary travel stimulated by lower fares.

For the most part, trunks and local service car-riers dropped short-haul markets. Between 1978and 1981, the number of trunk airline flights tomarkets with stage lengths under 200 miles droppedby 44 percent. Local carriers followed suit by re-ducing short-haul departures by 35 percent andmore than doubling flights in the range of 500 to1,000 miles. Some of the short-haul market hasbeen picked up by the commuter carriers.

The emerging pattern is one of increased activ-ity at large and medium airports and eroding serv-ice at the smallest nonhub airports, as shown intable 43. Confirmation of this pattern can befound in data on changes in aircraft departures

8FAA Statistical Handbook of Aviation, Calendar Year 1982(Washington, DC: Federal Aviation Administration, December1982).

Ch. 8–Forecasting and Trends ● 173

Table 41.—Changes in Stations Served by Air Carriers, 1978-81

Stations served in: Stations

Carrier 1978 1981 Added Deleted

Air California . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Air Florida . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Allegheny Commuter.. . . . . . . . . . . . . . . . . . . . . . . . . . . . .American Airlines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Braniff Airways . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Capitol International Airways. . . . . . . . . . . . . . . . . . . . . . .Continental Airlines. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Delta Air Lines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Eastern Airlines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Frontier Airlines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Jet America Airlines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Midway Airlines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Muse Air . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .New York Air . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Northwest Airlines. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Ozark Airlines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Pacific Southwest Airlines . . . . . . . . . . . . . . . . . . . . . . . . .Pan American . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .People Express Airlines . . . . . . . . . . . . . . . . . . . . . . . . . . .Piedmont Airlines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Republic Airlines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Republic Airlines West . . . . . . . . . . . . . . . . . . . . . . . . . . . .Southwest Airlines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Texas International Airlines . . . . . . . . . . . . . . . . . . . . . . . .Trans World Airlines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

United Airlines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .USAir . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Western Airlines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .World Airways . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1011494537

030716687

0000

31

471334

048

11446

93437

835431

0

1418566134

446717078

212

21336

4815211040

10437143250

775827

5

79

3024

5

417101519

212

2135

1152

109

1545

1217

101655

32

2388

01

101128

00000

103

150

17

2513

014

4

1612

90


Table 43.—Aircraft Departures by Hub Size (June 1, 1978, and June 1, 1982)

Number of Departures per weekb

Increase PercentHub sizea communities June 1978 June 1982 or decrease change

Departures by hub size:Large. . . . . . . . . . . . . . . . . . . . . . 1923 60,384 63,825 3,441 5.7Medium . . . . . . . . . . . . . . . . . . . 36 23,076 25,480 2,404 10.4Small. . . . . . . . . . . . . . . . . . . . . . 66 13,788 14,115 327Nonhub. . . . . . . . . . . . . . . . . . . . 504 28,575 25,239 (3,336) (1::;)

Total . . . . . . . . . . . . . . . . . . . . 629 125,823 128,659 2,836 2.3

Distribution of departure changes by hub size:Number of hubs

Change Large Medium Small Nonhub Total

Increase . . . . . . . . . . . . . . . . . . . 18 23 30 200 271No change . . . . . . . . . . . . . . . . . 0 0 1 12 13Decrease . . . . . . . . . . . . . . . . . . 5 13 35 292 345

for all commercial service airports, shown in table44. According to CAB, over 292 nonhub cities lostservice from 1978 to 1982.9 In some instances,localities dropped by larger carriers have receivedreplacement service from commuter airlines. Asa result, the average size of aircraft serving thesepoints is smaller, and the number of seats avail-able per departure has declined.

Medium hubs appear to be the main benefici-aries of increased air carrier activity. Two factorsappear to account for this trend: the air trafficcontrollers’ strike, which limited access to manylarge airports, and the growth of regional hubbing.As shown in table 45, departures from mediumhubs to other medium hubs increased by 30 per-cent between 1978 and 1982. Departures frommedium hubs to small or large hubs have also in-creased. By contrast, flights between large hubshave declined.

Formation of New Hubs

Air traffic has increased at airports in severalmedium-size cities as a result of moves by trunkairlines and local service carriers to consolidatetheir operations in hub-and-spoke route systems.(A diagram of a typical hub-and-spoke structureis shown in fig. 19. ) The basic strategy is to estab-

lish one airport as the hub into which traffic fromother cities is fed along radial routes. Flights bya carrier into and out of the hub are closely sched-uled in “complexes” or “connecting blocks” of ap-proximately 30 to 45 minutes so as to facilitatetransfers and minimize passengers’ waiting time.For the air carrier, the chief advantage is that serv-ice can be provided between smaller cities andalong thinly traveled routes more economicallythan if they were connected by direct flights.

Hubbing is not new. Delta established a hubat Atlanta long before deregulation, and hubbinghas been the core of Delta’s operating and mar-keting philosophy for years. The same principleof tight control of traffic feed through a singlepoint was applied over 20 years ago by Unitedin Chicago and USAir in Pittsburgh.

Deregulation, however, has added impetus tothe practice of hubbing by allowing airlines almostcomplete freedom to set up new routes and serv-ice points. Many local carriers, no longer satisfiedto feed traffic to the trunk carriers, have takenthe opportunity to establish their own regionalhubs—e.g., Piedmont at Charlotte, Dayton and,most recently, Baltimore; Republic at Memphis;and Western at Salt Lake City.

By setting up a hub at an underutilized airport(usually in a medium-size city), a carrier can avoidthe congestion encountered at major airports andreduce operating costs, while at the same time cre-ating markets in surrounding cities where it might


C q w mCuC60

l q c q q0 0 0

m-

‘%% I O(n I

1 - 1 -

%5 I 0 0 3 I

m t n u lI n m &t - - - m - hN WT oo-

r=

0 0 I

. . .. . .. . .. . .. . .: 2

. 3

. . .. . .. . .. . .. . .“ a l. &. 3

. . .. . .. . .. . .. . .

. . .. . .. . .. . .. . .“ a l. L. 3

. . .. , .. . .. . .. . .“a). >

176 “ Airport System Development

Table 45.—Changes in Weekly DeparturesBetween Major Categories of Airports

Percent changeItem (June 1978 to June 1982)

Between large hubs and:Large hubs . . . . . . . . . . . . . . . . . . . . . . (11.5)

Between medium hubs and:Large hubs . . . . . . . . . . . . . . . . . . . . . . 6.5Medium hubs . . . . . . . . . . . . . . . . . . . . 30.4

Between small hubs and:Large hubs . . . . . . . . . . . . . . . . . . . . . . (5.8)Medium hubs . . . . . . . . . . . . . . . . . . . . 11.2Small hubs . . . . . . . . . . . . . . . . . . . . . . (12.0)

Between nonhubs and:Large hubs . . . . . . . . . . . . . . . . . . . . . . (9.4)Medium hubs.. . . . . . . . . . . . . . . . . . . (12.1)Small hubs . . . . . . . . . . . . . . . . . . . . . . (25.5)Nonhubs . . . . . . . . . . . . . . . . . . . . . . . . (28.4)

not otherwise be practical to offer frequent airservice, or any service at all. Further, by choos-ing an airport that is not extensively served bycompeting carriers, the hubbing carrier maintainscontrol over traffic; passengers who arrive on thecarrier’s flights will depart on one of its connect-ing flights, not on a competitor’s.l0 In general, car-riers setting up new regional hubs have concen-trated on serving the traveler flying less than 1,000miles, a group which may represent as much astwo-thirds of the domestic market.

Hub-and-spoke operations seem to be increas-ing even at large airports, such as Chicago, At-lanta, Denver, and St. Louis. Since deregulation,most of the trunk carriers have increased the num-ber of markets served nonstop from large hubs.Ninety-nine percent of Continental’s departures(before the airline filed for bankruptcy) fed theHouston and Denver hubs. In 1979, Delta oper-ated nonstop flights from Atlanta to 59 cities;today it is serving 72, having added routes to theCaribbean as well as stations in the West andSouth. United’s schedule shows a net gain of 16points served nonstop from its Denver hub. In1981, TWA dropped direct service from its St.Louis hub to Atlanta, Toledo, and Knoxville but

added direct service to Des Moines, Houston, Lit-tle Rock, San Diego, San Antonio, and Chicago.However, some observers believe that this typeof route experimentation and readjustment maynot signal a long-term trend. They expect that,as air traffic increases generally, the disadvantagesof hub-and-spoke operation at the largest airportswill outweigh the advantages, and some air car-riers will divert their activity to medium hubs.

Air cargo carriers are also turning to hubbing.In 1972, Federal Express adopted the hub-and-spoke strategy for fast delivery of air cargo under70 pounds. The site selected, Memphis, is centrallylocated with a modern airport that is relativelyuncrowded and practically never shut down byweather. Between 11:00 p.m. and 4:00 a.m. eachday, Federal Express flights converge on Memphisfrom all over the country, unloading as many as80,000 packages, which are sorted and reloadedon flights to destination cities within a 4-hourperiod.

Other cargo carriers have likewise adopted thehubbing concept, especially for the overnight de-livery and small package segments of their busi-ness. For example, Flying Tigers has converted itsChicago storage facility into an overnight sortingcenter. Emery Air Freight, a large freight for-warder that now operates an aircraft fleet, hasbuilt an overnight sorting center in Dayton.

Selection as the site of a new hub can be a boonto the local airport in terms of growth in trafficand revenue. For example, since Piedmont estab-lished a hub at Charlotte, the increased traffic vol-ume has made it possible for the airport manage-ment to reduce landing fees for all airport usersand still enjoy higher revenues. Other aspects ofairport operation at Charlotte have benefited aswell. Income from parking and concessions is now20 percent ahead of forecasts, and the airport hasrecently built a new 10,000-ft runway and a $64million terminal. Similarly, Piedmont’s Midwest-ern hub in Dayton appears to be revitalizing anairport which had lost service after deregulation.

Western Airlines embarked on development ofa major hub in Salt Lake City in direct competi-tion with carriers serving regional and transcon-tinental markets through Denver and Dallas/FortWorth. Western currently offers nonstop service


‘ .

/

= 0u)

//

/ /\

.


to 33 cities, with approximately 86 flights perweekday. Largely as a result of Western’s hub-bing, traffic through Salt Lake City has increaseddramatically. Domestic enplanements are up 13percent from 1981, and departures have increasedby 25 percent. Salt Lake City Airport has re-.sponded by raising $30 million in bonds to ex-pand the terminal and to build Western a neweight-gate concourse.

On the other hand, becoming a new hub mayprove a mixed blessing for an airport. The pat-tern of operations associated with hubbing—many closely spaced arrivals and departures—means that the airport is extremely busy for briefperiods but practically idle the rest of the time.This peaking effect is illustrated in figure 20,which is a partial listing of Western Airlines’ ar-rivals and departures at Salt Lake City. In fact,peaking is even more severe than the figure showssince there are other airlines that rely on connec-tions with Western or that operate flights at thesame time as one of Western’s connecting banks.These extreme traffic surges strain airport capac-ity and may create the need for construction ofadditional facilities that are needed for only asmall part of the day.

Where the hubbing carrier dominates airportuse, the airport management may find itself at themercy of that carrier’s operating and expansionplans. For example, while Piedmont’s establish-ment of a hub at Dayton has led to a doublingof air carrier operations there, it also made theairport management dependent on one carrier forhalf its revenues. Construction of new facilities,needed solely to accommodate the highly peakedpattern of service of a single carrier, increases thatdependency and leaves the airport operator opento great financial risk should the carrier decideto move elsewhere.

Other adjustments may lie ahead for airportsthat serve as national or regional hubs. It is notyet known if there are minimum requirements fora hub to operate successfully or whether the oper-ation of several hubs in the same region will leadto an oversupply of air services in certain mar-kets. The pattern of hubbing will probably en-dure, but the fluidity of a deregulated market,especially one that continues to be distorted by

the recession and the lingering effects of the airtraffic controllers’ strike, makes the future diffi-cult to predict.

Aircraft Equipment

The key to economic efficiency for an aircraftis operation on a route to which it is well suitedin terms of size, range, and performance charac-teristics. Air carriers strive to use their equipmentin this way; but because aircraft are expensive andlong-lived investments, available equipment can-not always be matched to routes and traffic vol-ume as well as one would like. It is especially dif-ficult now, when service patterns and markets arein flux. The efficiencies attainable through the useof larger aircraft may not be available if theeconomics of hub-and-spoke operation and in-creased competition among carriers argue for theuse of smaller aircraft. As carriers begin to replacetheir present fleets with aircraft that are bettersuited to current market conditions, aircraft sizebecomes an important factor in assessing futureairport capacity needs.

The aircraft that now make up the U.S. jet fleetare of three major types. Narrowbody short- andmedium-range aircraft, like the DC-9, B-727, andB-737, constitute about 75 percent of the fleet.These aircraft, which seat 100 to 150 passengers,are used primarily for flights with stage lengthsunder 1,000 miles or those with several inter-mediate stops. Medium- and long-range widebodyaircraft such as the DC-10, L-1011, or B-747 makeup about 18 percent of the jet fleet. These aircraft,seating 250 to 400 passengers, are most efficientlyoperated on heavily traveled long routes of morethan 1,500 miles, but they can be (and have been)used effectively on shorter routes of sufficientlyhigh density. Long-range narrowbody aircraft likethe DC-8 and B-707, with 140 to 180 seats, makeup only about 7 percent of the fleet and are usu-ally operated on routes over 1,500 miles whichare not densely traveled enough to warrant useof a widebody (see fig. 21 and table 46).

The short- and medium-range aircraft whichmake up three-quarters of the fleet vary widelyin age (see fig. 22). About 60 percent are over 10years old, and 40 percent are over 15 years old.Many will be replaced in the next few years,

Ch. 8—Forecasting and Trends “ 179

;


Figure 21.–Composition of U.S. Commercial Jet Fleet, 1982

Medium/long rangenarrowbodv

7.OVO -

Short/medium range narrowbody aircraft:B-727 . . . . . . . . . . . . . . . . . . . . . . . . . 1,012B-737 . . . . . . . . . . . . . . . . . . . . . . . . . 238BAC-111 . . . . . . . . . . . . . . . . . . . . . . 26DC-9 . . . . . . . . . . . . . . . . . . . . . . . . . 473

1,749

Medium/long range widebody aircraft:A-300 . . . . . . . . . . . . . . . . . . . . . . . . . . 25B-747 . . . . . . . . . . . . . . . . . . . . . . . . . .128DC-10 . . . . . . . . . . . . . . . . . . . . . . . . . . 159L1011 . . . . . . . . . . . . . . . . . . . . . . . . . . ~

429Medium/long range narrowbody aircraft:6-707 . . . . . . . . . . . . . . . . . . . . . . . . ...71DC-8 . . . . . . . . . . . . . . . . . . . 90

161

Ch. 8—Forecasting and Trends . 181

Figure 22.—Age of Aircraft in Service in U.S.

600~

Commercial Jet Fleet (different scales)

Short/medium range

160F

80

60

40

2C

o

Long-range widebody

Long-range narrowbody

Before1963 1965-67 1968-70 1971-73 1974-76 1977-79 1980-82

Year placed in serviceSOURCE: Commercial Aircraft Fleet Databank, Lockheed-Georgia Co., July 1982.


Table 46.—Present and Future Commercial Jet Aircraft

Current generation Next Generation

Manufacturer Designation Seats a Manufacturer Designation Seats a

Short/medium-range narrowbodyBoeing B-727-1OOBoeing B-727-200Boeing B-737-200Fokker F-28McDonnell Douglas DC-9-1OMcDonnell Douglas DC-9-30McDonnell Douglas DC-9-50

Medium/long-range widebodyAribus industrie A-300Boeing B-747-1OOBoeing B-747-200Lockheed L-101 1McDonnell Douglas DC-10-IOMcDonnell Douglas DC-10-4O

Medium/longwange narrowbodyBoeing B-707-120Boeing B-707-320McDonnell Doualas DC-8-62

70-131120-175115-13065-85

90115139

220-300374-452374-500256-400250-380250-380

125120-189

189

Airbus IndustrieBoeingBoeingBritish AerospaceFokkerMcDonnell Douglas

Airbus IndustrieBoeingBoeingBoeing

A-320B-737-300B-757BAE-146-200P-332MD-80

A-31OB-747-300B-767-200B-767-300

150132-148186-22082-109110167

205-265530

210-255270

largely with aircraft like the B-757 (195 seats), DC-9-80 (155 seats), and the B-737-300 (140 seats)—all of which are bigger than equivalent models incommon use today (see table 46). Although somemanufacturers are considering production of newshort- and medium-range aircraft with about 100seats, they are generally viewed as replacementsfor aircraft of yet smaller size, like the DC-9-1O.

A somewhat opposite trend is expected in thelong-haul segment of the fleet. The long-rangenarrowbody jets are 13 to 20 years old. They aretechnologically obsolete, fuel-inefficient, andnoisy by present FAA standards. Most will beretired from service within the next 5 years, al-though it is possible that some could be refittedwith new engines to extend their economic lifetimea bit longer. There is no new narrowbody aircraftin design or production which is an exact equiv-alent. Instead they will be replaced either with cur-rent generation widebodies or new models like theB-767 or the A-31O.

Most long-range widebody aircraft now in thefleet are expected to remain in service for severalmore years, but they too will eventually be re-placed. Except for a stretched version of the B-747, which Boeing expects to produce for use onheavily traveled long-haul routes, most new wide-

bodies for domestic use will be smaller than ex-isting models. Aircraft such as the B-767 and A-310, which seat between 230 and 270, are typicalof this new generation of long-range aircraft.

Overall, the average size of the jet fleet mayremain about the same, or perhaps decrease slightly,but there will be a greater range of sizes. Today’sfleet averages 150 seats within a range of 85 to400; the future fleet might be as much as 10 per-cent smaller and made up mostly of aircraft inthe 100- to 250-seat range, but with some muchlarger aircraft with 500 seats (or even 700 to 800seats for a full upper deck 747) operating on denseroutes.

To the extent that the demand for air travel in-creases over the coming years, airports will clearlybear an additional burden. However, this burdenmay not be as onerous as it would first appear.For example, the average load factor in commer-cial aviation has been rising over the past 5 yearsin response to higher operating cost and lowerfares. To the extent that the average load factorcontinues to increase in the future, fewer aircraftoperations will be needed to handle a given num-ber of passengers. Or, viewed another way, thegrowth of aircraft operations will not be as rapidas the growth in passenger enplanements. If, how-


ever, the number of competitors increases in cer-tain high-density markets, the average number ofpassengers attracted by each competitor will therebybe reduced. In this case, the total number of flightoperations would be likely to increase as carriersfavored smaller aircraft and placed emphasis onfrequency of service and price competition.

The next generation of short- and medium-range aircraft—workhorses of the current jetfleet—are likely to have an additional 20 to 30seats per aircraft. This should translate into a re-duction of 20 percent or so in the number of oper-ations needed to carry a given volume of passen-gers. On the other hand, the average size ofaircraft on long-haul routes will probably decreaseover the next decade as more carriers competefor the market. While this might lead to an in-crease in the number of operations, there is rea-son to believe that reduction in aircraft size couldalso augur a change in the structure of airlinenetworks.

The present generation of widebodies provedeconomical only in service between major hubs.Many passengers had to be routed through thesehubs in order to take advantage of the attractive

Figure 23.—Composition of

economics afforded by widebodies. Bringing thesepeople into and out of large hubs often createdadditional and unnecessary aircraft operationsthat contributed to the delay and congestion atthese airports. If airlines choose to utilize the newgeneration of smaller (but economical) long-rangeaircraft to provide direct service in lower volumemarkets or in connecting flights to smaller hubs,the burden of increased passenger enplanementswill be shifted away from presently congested ma-jor hubs.

The aircraft used by commuter carriers arealmost entirely propellor-driven, with either pis-ton or turbine engines. Although they range insize from 6 to 60 seats, over 80 percent seat 19

or fewer passengers (fig. 23). Federal regulations—which limited the size of commuter aircraft to amaximum of 19 seats before 1972, 30 seats in 1972,and 60 seats since 1978—dampened the interestof U.S. manufacturers in building small transportaircraft.11 Consequently, few U.S. firms now pro-

Commuter Aircraft Fleet, 1981a

All single engine

Mul t iengine (seat ing capaci ty )

1-9

10-20

21-30

>31

Otherb

Percent

P 13.3

37.0

32.1

7.4

7.9

❑ Piston powered

Turbine powered

Other2.3

aPassenger aircraft onlybHellcopter and let.

SOURCE: Regional Airlines Association Annual Survey, 1981.

25-420 0 - 84 - 13


duce small transport aircraft, most of which areimported from Canada, Brazil, and the Nether-lands. With the progressive lifting of restrictionson the size of aircraft that commuters may oper-ate, the average size of aircraft in the commuterfleet has increased from about 10 seats in 1970to about 15 in 1980. This growth trend is likelyto continue, and many of the new commuter air-craft entering the fleet over the next decade willbe in the 30- to 60-seat range.

The growth of commuter airlines will have im-portant implications for future airport needs. Thecommuter airline fleet tripled between 1970 and1981, with much of the growth occurring sincederegulation (table 47). FAA forecasts that by2000 the commuter fleet will triple again, reachingabout 4,500 aircraft .12 While many of these willbe used in air service to small communities, wheremajor air carriers do not find it profitable to oper-ate, a large number will also converge at hub air-ports to feed passengers to larger carriers. Manyof these operations will be at busy airports whereit will be hardest to accommodate them at gatesand in terminal facilities. Further, the mixing of

12 National Airspace System Plan (Washington, DC: Federal Avia-tion Administration, April 1983, revised edition), p. II-2.

Table 47.—Aircraft Operated by Commuter andRegional Airlines, 1970-81

Aircraft PercentYeara in service annual change

1970 . . . . . . . . . . . . . .1971 . . . . . . . . . . . . . .1972 . . . . . . . . . . . . . .1973 . . . . . . . . . . . . . .1974 . . . . . . . . . . . . . .1975 . . . . . . . . . . . . . .1976 . . . . . . . . . . . . . .1977 . . . . . . . . . . . . . .1978 . . . . . . . . . . . . . .1979 (est) . . . . . . . . . .1980b . . . . . . . . . . . . . .1981b . . . . . . . . . . . . . .

687782791885997

1,0731,0091,1191,2001,3501,6061,743

141

12138

– 611

712198

large jet and small propellor-driven aircraft in thesame traffic stream will add a burden to air traf-fic control and could aggravate congestion anddelay both in the airspace and on the airport sur-face. A partially offsetting factor is that the aver-age size of commuter aircraft will probably in-crease. Thus, while commuter operations willgrow, they will not be as great as they would beif aircraft size were still restricted by a regulatoryceiling of 19 or 30 seats.

IMPLICATIONS FOR AIRPORT DEVELOPMENT

There is almost universal agreement that airtraffic will continue to increase throughout the restof this century and that some forms of growthcould lead to serious congestion and delay in thenational airport system. There is considerably lessagreement about how much traffic will grow andhow it will be distributed across airports andamong sectors of civil aviation. Of the two ques-tions, distribution is probably the more impor-tant, since it will determine which airports are af-fected and what measures can be taken to dealwith the congestion and delay that could result.If, for example, traffic continues to concentrateat airports that are already severely congested orat those now approaching capacity limits, thereis a legitimate fear that major hubs will be caughtin a form of “gridlock” that will spread through-out the national airport system. On the otherhand, if traffic patterns change and new growth

occurs not at the airports now saturated but atother airports with adequate reserve capacity, thesystem can absorb a large amount of new trafficwithout an appreciable increase in local or gen-eral congestion and delay. Between these extremesare other possibilities, each with differing conse-quences for various types of airports and classesof airport users.

If aviation demand grows at the rate foreseenby FAA, delay will increase at high-density air-ports—perhaps intolerably—and could even spreadto other airports that are now relatively free ofcongestion. Congestion would be especially severeif traffic were to concentrate at a few large hubs,so severe that air carriers would probably seekto avoid escalating delay costs by shifting theircenters of operation to other airports. Some car-riers might shift to other suitable large hubs, but


more likely the move would be to medium air-ports with adequate facilities and ample surpluscapacity—in effect breaking the pattern that nowexists at airports such as Atlanta or Chicago andselecting a site where the rehubbing carrier wouldbe the principal, if not the only, major airline atthe site. The current trend toward hub-and-spokeroute structure and the present inclination of air-lines to purchase short- and medium-range nar-rowbody aircraft seem to indicate the likelihoodof rehubbing. If so, the area of greatest interestover the next decade or so could be medium hubs,where the influx of traffic may necessitate rapid,but selective, expansion of airport facilities.

The prospect of extensive rehubbing at mediumairports does not necessarily imply that these air-ports will face a capacity crisis like that of somemajor hubs today. If the medium airports that areto serve as new hubs are chosen wisely and if thepractice does not lead to abuse,13 they offer a largecapacity reserve. The key is how the airlines re-spond in a deregulated environment where theymay compete in a variety of ways, some of whichmight offer a short-term advantage but at the costof an unbalanced use of the airport system as awhole. The full implications of rehubbing are byno means clear. On one hand, rehubbing seemsto be an attractive way to absorb growth in de-mand and provide air transportation service whileavoiding the delay costs of a large multi-airlinehub. On the other hand, we have too little ex-perience with the dynamics of an unregulated airtravel market to know how many new, smaller,single-airline hubs are economically practical.

h this setting, the reliever concept is of greatpotential value. By shifting GA traffic away fromcenters of air carrier operation while still allow-ing GA adequate access to major metropolitan

areas, relievers offer important advantages. Thereis the advantage of allowing each sector of civilaviation to grow without impediment to the other.This would not only reduce delay and its associ-ated costs for all users of metropolitan airports,it would also have important collateral effects onairport efficiency (through segregation of dis-similar types of traffic) and on airport expansioncosts. It may be less costly to upgrade one or tworeliever airports to handle more traffic than to ex-pand the major air carrier airport to absorb thesame amount. If general aviation—especially busi-ness aviation—grows as much as FAA projects,it is clear that some way will have to be foundto serve this segment of civil aviation.

For the reliever concept to work, however, itwill not be sufficient simply to push GA trafficoff to some other airport in the metropolitan area.For GA to accept this diversion and to embracethe reliever concept, the alternate airports mustprovide facilities and services appropriate to GAneeds and of quality comparable to that of themajor airport. This implies not only adequate run-ways, aprons, navigation aids, and ATC services,but also facilities for aircraft storage and mainte-nance and landside connections to activity centersin the metropolitan area. The reliever thus needsto be a mirror of the air carrier airport, not justanother place to land.

Both these observations suggest the need tothink of airports as a system, not as separate parts.The recently enacted Airport Improvement Pro-gram makes it clear that such a broad view isneeded in order to determine how to make gooduse of the infrastructure already in place and howto fit new demand into an existing system thathas large unused capacity overall, even thoughit is congested at a few points. This implies thata strategy of restructuring airport use through ad-justment of operational patterns and judicious im-provement of existing airports may be a less ex-pensive and more manageable alternative thancontinuing to build new facilities in response todemand wherever and whenever it occurs.

Documents

Chapter 8 FORECASTING AND TRENDS