Airline hubs

MISSING IMAGE

Material Information

Title:
Airline hubs changes in urban employment structure and network connectivity
Physical Description:
x, 186 leaves : ill. ; 29 cm.
Language:
English
Creator:
Ivy, Russell L., 1962-
Publication Date:

Subjects

Subjects / Keywords:
Access to airports   ( lcsh )
Industrial location   ( lcsh )
Geography thesis Ph. D
Dissertations, Academic -- Geography -- UF
Genre:
bibliography   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1992.
Bibliography:
Includes bibliographical references (leaves 173-185).
Statement of Responsibility:
by Russell L. Ivy.
General Note:
Typescript.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 028699743
oclc - 27779136
System ID:
AA00013582:00001


This item is only available as the following downloads:


Full Text











AIRLINE HUBS: CHANGES IN URBAN EMPLOYMENT
STRUCTURE AND NETWORK CONNECTIVITY

















By

RUSSELL L. IVY














A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA

1992














ACKNOWLEDGEMENTS


I would like to thank Dr. Edward Malecki for his support,

encouragement and guidance throughout my years at the

University of Florida. I also wish to thank the other members

of my supervisory committee, especially Dr. Timothy Fik, whose

quantitative assistance was invaluable. A great deal of

computer assistance, including all graphics, and patience were

given by Jan Coyne.

I would also like to thank Dr. Jesse Wheeler, Dr. J.

Trenton Kostbade and Ann Wright at the University of Missouri

for sparking my interest in geography.

Lastly, I would like to thank my mother, who has always

supported and encouraged without question anything and

everything I have attempted.















TABLE OF CONTENTS

ACKNOWLEDGEMENTS ...................................... ii

LIST OF TABLES ................... ..................... vi

LIST OF FIGURES........................................ viii

ABSTRACT .................................... ........... ix


CHAPTERS

1 INTRODUCTION ................ ..................... 1

Hub-and-Spoke Structure....................... 1
Connectivity Change and the Location of
Economic Activities......................... 2
Objectives............ ......................... 3
Structure of Presentation...................... 3

2 INDUSTRIAL LOCATION DECISIONS .................... 5

Introduction................................... 5
Theory of Location of Organizations........... 5
Spatial Division of Labor..................... 7
Location Criteria of Various Parts of the
Firm ... ............... .................... 11
Regional Development and Industrial Location.. 18
Professional Workers.......................... 20
Labor Migration............ .................... 21
The Large Metropolitan Area.................... 21
Conclusion.................................... 24

3 THE DOMESTIC AIRLINE INDUSTRY.................... 25

Introduction..................................... 25
The Deregulated Airline Industry .............. 25
The Hub-and-Spoke System...................... 34
Conclusion.................................... 41

4 VARIATIONS IN HUB SERVICE IN THE DOMESTIC AIR
TRANSPORTATION NETWORK ......................... 42

Introduction................................... 42
The FAA Hub.................................... 43

iii









Deregulation of the U.S. Airline Industry..... 43
Hub-and-Spoke Growth and Development.......... 46
Service Variations............................ 47
Hub Connectivity................................ 51
Analyzing Hub Connectivity.................... 52
Results........................................ 55
Connectivity and Intensity Classification
Scheme .................... .................. 61
Conclusion.................................... 65

5 CHANGES IN CONNECTIVITY AND PROFESSIONAL EMPLOYMENT
LOCATION ........................................ 67

Introduction.................................. 67
Research Questions............................ 70
The Labor Data............. .................... 71
The Air Service Data.......................... 79
Research Methodology.......................... 85
Accessibility Indices......................... 86
Employment-Connectivity Relationships.......... 87
Regional Trends........ ........................ 96
Numerical Analysis............................ 107
Correlation Analysis....................... 110
Hierarchical Analysis of Hub Cities........ ll
Lag Structure............................... 115
Time Series.................... ............. 115
Conclusion.................................... 119

6 SUMMARY AND CONCLUSIONS.......................... 120

Introduction................................. .120
Summary of Results............................. 120
Guidelines for Future Research................ 122
Concluding Remarks and Contributions of the
Study ....................................... 125

APPENDICES

A CONNECTIVITY INDEX TOTALS........................ 126

B CMSA/MSA COUNTY COMPONENTS....................... 130

C ADMINISTRATIVE AND AUXILIARY EMPLOYMENT.......... 134

D TOTAL EMPLOYMENT BY MSA .......................... 138

E TOTAL ENPLANEMENTS BY MSA ........................ 143

F TOTAL NONSTOP DESTINATIONS REACHED FROM EACH
MSA........................................ .... 148

G ACCESSIBILITY INDICES: 1978-1988................ 152

iv









BIBLIOGRAPHY........................................ 173

BIOGRAPHICAL SKETCH.................................. 186














LIST OF TABLES


TABLE PAGE

2.1 Ten Most Important Factors in Selecting
Locations for Manufacturing/Processing
Plants................................ 12

2.2 Ten Most Important Factors in Selecting
Locations for Company Facilities...... 14

2.3 Five Most Important Factors in Selecting
Locations for Various Types of Company
Facilities............................. 16

2.4 Important Location Factors by Facility
Type................................... 19

3.1 Single Carrier Dominance at Hubs............ 40

4.1 Fall 1991 Hub Cities of the Major U.S.
Carriers............................... 45

4.2 Pre and Post-Deregulation Hub Cities
(1991)................................ 48

4.3 Nonstop Domestic Connections on all Carriers
by FAA Hub-Type (Fall 1991)............ 50

4.4 FAA Hubs ................................... 56

4.5 Accessibility Indices for Matrices Cl
and T................................. 58

4.6 Accessibility Indices of Matrix T With
Different Scalars..................... 60

4.7 Accessibility Indices of Individually
Weighted Nodes........................ 62

4.8 Hub Strength Classification Scheme.......... 63

5.1 60 Largest Metropolitan Areas of the U.S.
(1988-in thousands)................... 68








TABLE PAGE
5.2 Most Administrative/Auxiliary Employees:
1988................................... 73

5.3 Largest Growth in Administrative/Auxiliary
Employees: 1978-1988................. 74

5.4 Administrative/Auxiliary Employment
Loss: 1978-1988....................... 76

5.5 Administrative/Auxiliary Growth by Region:
1978-1988.............................. 76

5.6 Highest Percentage of Administrative/
Auxiliary as a Percentage of Total
Employment: 1988..................... 77

5.7 Multiple Airport Cities .................... 81

5.8 Leading Enplanement Cities of the U.S.:
1988................................... 82

5.9 Greatest Increases in Non-Stop Connections:
1978-1988.............................. 84

5.10 Study Set Cities With Highest Accessibility
Indices: 1988 ........................ 88

5.11 Mean Annual Rate of Change in Employment
and Connectivity: 1978-1988.......... 108

5.12 Hub Clusters Based on Population Totals.... 113

5.13 Mean Annual Rate of Change by Population
Cluster: 1978-1988................... 114

5.14 Time Series Analysis....................... 117















LIST OF FIGURES


FIGURE PAGE

3.1 Hub Cities of the Major U.S. Airlines:
1991................................... 36

5.1 Census Regions of the United States........ 78

5.2 Change in Professional Employment vs.
Connectivity.......................... 89

5.3 Nonhub Employment vs. Connectivity......... 91

5.4 Hub Employment vs. Connectivity............. 92

5.5 Employment Change: Hub vs. Nonhub.......... 93

5.6 Connectivity Change: Hub vs. Nonhub....... 95

5.7 Northeast Hub Rates of Change............... 97

5.8 Northeast Nonhub Rates of Change............ 98

5.9 North Central Hub Rates of Change......... 101

5.10 North Central Nonhub Rates of Change....... 102

5.11 Western Hub Rates of Change................ 103

5.12 Western Nonhub Rates of Change.............. 104

5.13 Southern Hub Rates of Change............... 105

5.14 Southern Nonhub Rates of Change............. 106


viii














Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree Doctor of Philosophy

AIRLINE HUBS: CHANGES IN URBAN EMPLOYMENT STRUCTURE
AND NETWORK CONNECTIVITY

By

Russell L. Ivy

August, 1992

Chairperson: Edward J. Malecki
Major Department: Geography

The deregulation of the domestic airline industry in the

U.S. brought about many changes in air transportation.

Carriers have come and gone, fares have fluctuated and

enplanement figures have risen tremendously. Fundamental

changes have also occurred in the geographic structure of the

air transport network. In particular, the hub-and-spoke

system has been adopted by major carriers and has led to

increases in flow efficiency and connectivity. One or more

cities are chosen by the airline as a regional collection

point for passengers. Passengers from many different origins

are funneled into the hub city to connect with flights to

their final destinations. Hub cities, therefore, are

characterized by having many nonstop destinations available

from their airports.








Living and working in hub cities can have tremendous

benefits for the time-sensitive traveler. Professional

employees such as research scientists and engineers, managers

and salesmen fall into this category. Face-to-face

communication is an important part of their job as they are

often required to travel quickly to a client, other facilities

of the firm or yet another city. Recent industrial location

literature and surveys of corporations imply that the high

quality air transportation that hub cities offer is extremely

desirable when locating or relocating these nonroutine

employees of the firm. The purpose of this study was to link

the changes in connectivity that occur as a city is chosen to

be an airline hub to changes in professional employment

growth.

The analysis of this study, however, fails to show a

significant relationship between changes in air service

connectivity and professional employment growth in the 60 MSAs

in the study set over the period from 1978 through 1988.














CHAPTER 1
INTRODUCTION


The deregulation of the domestic airline industry has

brought about many changes in air travel in the United States.

Both new and older, established carriers have disappeared in

bankruptcy courts or through mergers with other carriers.

Fares dropped tremendously at first, but have been on the rise

since the end of the 1980s, and passenger enplanements have

skyrocketed.

Another significant change has been in the geographic

structure of the air travel network itself. Fewer nonstop

flights are available to some cities than before, as service

is now fed into hub cities of specific carriers instead of

most or all larger cities within a region. This study will

examine changes in the professional employment structure of

those hub cities as their connectivity levels

rise.



Hub-and-Spoke Structure



Hub cities act as a collection point of air passengers in

a region and thus have a tremendous amount of flow in and out

of them. Passengers are flown into the city from many







2

different origins to connect with flights to their final

destinations. Hub cities, therefore, have a multitude of

nonstop alternatives available to local residents.



Connectivity Change and the Location of Economic Activities



It can be posited that the restructuring of air networks

can greatly influence the location of economic activities.

Because of the large number of nonstop flights and

destinations from hub airports, certain corporate activities

should find it advantageous to be located in a hub city.

Who is the corporate traveler? What types of positions

within the firm would require frequent air travel? Typically

such employees are those for whom a lot of face-to-face,

nonroutine communication both within and outside of the firm

is vital. Even though communication technology has made

phenomenal advancements in recent years, situations occur for

which there is no viable alternative or substitute for an on-

site visit. Firms desire close contact with their clients and

markets to respond quickly to their changing needs. Managers,

sales staff, scientists, engineers and other administrative

employees are often required to travel quickly to other parts

of the company, to clients, or to other locations. In order

to minimize travel time, it is preferable to fly nonstop to

one's final destination rather than to fly two or more flight

segments with layovers in between. The savings in total







3

travel time can be very significant to a profit-oriented

organization as well as the general public.



Objectives



How do hub cities measure up? Do they indeed have a

higher concentration of employees from management and research

and development labs? More importantly, has the change to a

hub-and-spoke network yet enhanced the attractiveness of newly

designated hub cities? Investigation of these issues is the

focus of this study.

Data will be collected and analyzed from various

publications of the Official Airline Guide, Inc., the U.S.

Census Bureau and the Federal Aviation Administration for each

year from 1978 through 1988 to examine the degree to which

changes in connectivity that occur when an airline chooses a

particular city as a transfer hub are related to changes in

employment growth of the above-mentioned type.



Structure of Presentation



Chapter 2 presents a review of the pertinent industrial

location literature, while Chapter 3 discusses the domestic

air transportation industry before and after deregulation.

These two chapters will build a framework within which to

place this dissertation in the broader perspective of







4

industrial location and air transportation research. In

Chapter 4, variations in hub service quality and connectivity

are discussed and a hub classification scheme (based on

connectivity) is derived using binary matrix analysis. The

main analysis (methodology and results) of this dissertation

is presented in Chapter 5. Chapter 6 is a summary of the

study and directions of future research.














CHAPTER 2
INDUSTRIAL LOCATION DECISIONS


Introduction


Different activities of the firm have different

locational requirements. Nonroutine activities, such as

management and research and development, are less influenced

by wage rates and site location costs than routine activities,

such as manufacturing. This chapter will look at the location

of organizations in the context of spatial division of labor,

and identify the specific locational needs of the various

activities of the firm. It also includes a discussion of

professional workers and the role they play in business

location decisions. Therefore, this chapter is an important

component of the study since it will help identify places

where professional workers are likely to be concentrated.



Theory of Location of Organizations



Traditionally, industrial location theory has been based

on profit maximization through the minimization of costs

(Alexander, Gibson, 1979; Boyce, 1978; Chapman, Walker, 1987;

De Souza, 1990; De Souza, Foust, 1979; Dicken, Lloyd, 1990;

Hoover, 1948; Wheeler, Muller, 1986). Transportation costs

5







6

and the locations of raw materials and markets were major

concerns. One of the main flaws in this approach, however, is

the assumption that labor quality, supply and cost are the

same at all locations (Moriarty, 1980; Watts, 1987).

Elaborate models to determine minimum-cost sites focused,

consequently, solely on transportation costs as spatially

variable (Dicken, Lloyd, 1990).

A broader view of industrial location theory, on the

other hand, holds that labor is the key to determining firm

location since, in reality, labor has a high degree of spatial

differentiation in terms of both supply and level of skill

and, therefore, cost (Storper, Walker, 1984). This

differentiation of labor also occurs according to both

industry (industries differ in their mix of labor) and

function within the firm, such as marketing, research and

development or routine manufacturing. Thus, the geographic

pattern of industry represents the spatial division of labor

as particular industries, as well as particular activities of

the firm, correspond to the geographic location of labor

(Blair, Premus, 1987; Czamanski, 1981; Massey, 1984; Storper,

Walker, 1983).

Spatial division of labor within the firm has become

quite commonplace and has been the subject of much academic

work (Schoenberger, 1986, 1987; Watts, 1987). Technological

improvements in communication and automation have increased

the spatial separability of parts of the production process







7

and of organizational functions (Storper, Walker, 1983). The

different parts of the organization can search for locations

that will more greatly benefit their specific tasks or roles

within the firm, without the necessity of being physically or

locationally tied to other functions of the organization that

might have entirely different locational requirements. As the

spatial variations in cost, quality and availability of

nonlabor inputs have largely diminished, the issue of labor

has risen greatly in importance (Moriarty, 1980; Storper,

Walker, 1984).



Spatial Division of Labor


Because a supply of unskilled labor for standardized,

routine production is virtually ubiquitous, labor-intensive

manufacturing and assembly operations can locate almost

anywhere (Blair, Premus, 1987; Clark, 1981; Czamanski, 1981;

Massey, 1984; Schmenner, 1982). When the manufacturing

procedure does become standardized or routine, firms will

establish branch plants--often in nonurban areas with

depressed local labor markets to reduce labor competition

(Erickson, Leinbach, 1979). Labor-intensive operations

primarily seek to reduce labor costs (to obtain higher profits

for the firm) and generally locate or relocate in low-wage

areas anywhere in the world. Such areas are typically located

outside of the traditional innovative centers of industry









since production need not be linked to innovative research and

development, which relies on highly skilled labor. Production

shifts to these peripheral regions from high-cost core regions

because the competition for labor in large urban areas, and

therefore the price paid for labor, has increased to the point

where profit margins are unacceptably low or nonexistent

(Barkley, 1988; Markusen, 1985). Thus, routine operations in

both manufacturing and services have largely decentralized to

peripheral areas within developed countries and to developing

countries (Clark, 1981; Schoenberger, 1986, 1987).

These industrial location shifts can be explained by

Vernon's product life cycle theory (1966). Industries, firms

and products move down the urban hierarchy as they go through

different stages (each with distinct locational requirements,

labor needs, growth rates and profitability) in their life

cycle (Barkley, 1988). Mobility of the various functions of

the firm is often a necessity for firm survival, influenced by

the potential for profit in each stage.

The profit cycle model complements the insights from the

project life cycle model. Markusen (1985) describes the long-

term profit cycle as having five characteristic stages. In

stage one (zero profit), output is very low (often done in

test runs) while a new product or new design is being

initiated. Costs per unit are high since the primary

workforce involved is composed of many nonroutine workers,

such as scientists, engineers and other technicians, who are







9

experimenting with the test runs. Production is usually

concentrated at or very near the firm's research and

development laboratory.

As development of the product or service becomes

successful, dramatic growth of production and profitability

occurs. The firm has a temporary monopoly in the market, and

demand-led prices for the new product or service will create

excess profits. In this stage (super profit), standardization

of production begins to take place, involving a greater number

of routine laborers than previously, and the average cost per

unit produced will begin to drop. As a result, production

begins to decentralize from the research and development

center.

In stage three (normal profit), other organizations enter

the market with similar products, and price competition can be

intense as the market approaches saturation. Mass

standardized production is now the norm as cost-cutting

becomes imperative. The trend toward decentralization becomes

accelerated and the role of the professional/technical worker

largely diminishes.

Stage four (normal-plus and normal-minus profit) and

stage five (negative profit) are both post-saturation stages.

In the former, profit levels can drop even further due to the

greatly increased competition or can rise slightly via

successful oligopolization. The latter stage is usually where







10
production will cease. The product becomes obsolete, and the

firm takes absolute losses on production.

Nonroutine and innovative activities of the firm, such as

research and development, corporate and regional headquarters

functions and marketing operations, need to be located in core

regions of developed nations as they have a high dependence on

skilled labor (Malecki, 1986). These divisions of the

organization cannot thrive in all locations in space as the

availability of labor skills varies by nation, region and city

(Kim, 1987; Walker, Storper, 1981).

These higher order functions often remain with or very

near the corporate headquarters in metropolitan areas or in

other metropolitan areas near other parts of the firm (Clark,

1981; Erickson and Leinbach, 1979). Due to rapid

technological and market changes, they require a high quality

and quantity of day-to-day information concerning competition,

suppliers and customer needs. While there is a trend toward

shorter production runs and more innovative products, which

requires interaction of research and development with

manufacturing, production and research activities remain on

the whole largely separate in purpose and location. This

"flexible" production, therefore, needs the same large-city

skilled labor as research and development (Arnold, Bernard,

1989; Malecki, 1991; Schoenberger, 1988). Thus, a

transportation and communication infrastructure is a

requirement for firms when locating nonroutine activities.









Location Criteria of Various Parts of the Firm


A recent survey of Fortune 500 firms clearly indicates

that the various parts of the firm have different locational

requirements (Corporate Site Selection for New Facilities,

1989). It examines location and relocation criteria of the

firms and separates the information into two general

categories: 1) manufacturing/processing plants and 2) company

facilities. The category of manufacturing/processing plants

includes assembly plants, processing plants, extraction plants

and warehouse/distribution centers, while the company

facilities category includes corporate headquarters, regional

headquarters, branch offices, sales offices and branch office

processing.

Table 2.1 shows the ten most important factors in

selecting locations for manufacturing/processing plants as

identified by the study. The criteria mostly relate to costs.

Reduction or minimization of labor costs, taxes,

transportation costs and land prices are high-priority

considerations for these facilities. It is evident that small

cities or, in some cases, rural areas in depressed economic

regions would be ideal choices (Goldfarb, Yezer, 1987; Heenan,

1991).

The study also found that the South Atlantic region of

the United States (North Carolina, South Carolina, Georgia and

Florida) was the most popular location choice for









TABLE 2.1
TEN MOST IMPORTANT FACTORS IN SELECTING LOCATIONS FOR
MANUFACTURING/PROCESSING PLANTS



Easy access to trucking services 79%

Easy access to domestic markets, customers
and clients 74%

Cost of labor 74%

Ample area for future expansion 73%

Easy availability of electricity 71%

Community receptivity to business and industry 70%

Reasonable government/state and or local
corporate tax structure 68%

Fair-market property costs 67%

Availability of skilled workers 64%

Extent of unionization 64%



Source: Corporate Site Selection For New Facilities, The
Time Inc. Magazine Company, 1989.







13

manufacturing/processing plants in the past five years and

will continue to be the favorite for the next five years.

Lower taxes, cost of living and relative absence of

unionization add to the attractiveness of this region to

firms. The South Atlantic was followed by the East South

Central region (Kentucky, Tennessee, Alabama and Mississippi)

and the West South Central region (Arkansas, Louisiana, Texas

and Oklahoma) in that order. Atlanta, Memphis and Dallas (in

rank order) were the most popular urban areas chosen for

location by those firms not desiring a small city or rural

location.

The attraction of the South has been evident for some

time. An earlier study of change in manufacturing employment

by state from 1967 to 1972 showed the greatest gains in

manufacturing employment in North Carolina (100,000), Texas

(78,000), Florida (58,000), Tennessee (49,000) and Georgia

(44,000) (Moriarty, 1980). Indeed the "Selling of the South"

for manufacturing operations began in the 1930s (Cobb, 1980;

1984).

The ten most important factors in selecting locations for

company facilities, a group that includes nonroutine corporate

functions (as determined by the Fortune 500 study), are listed

in Table 2.2. Here cost considerations largely give way to

access to markets and the supply of professional labor as well

as adequate transportation facilities.









TABLE 2.2
TEN MOST IMPORTANT FACTORS IN SELECTING LOCATIONS FOR
COMPANY FACILITIES



Easy access to domestic markets, customers
and clients 73%

Easy access to airport 62%

Efficient transportation facilities for
people/employees 59%

Availability of affordable housing 55%

Availability of technical or professional
workers 54%

Facilitates access to prospective clients 52%

Urban/metropolitan location 50%

Cost of living 48%

Reasonable government/state and/or local
corporate tax structure 48%

Fair-market property costs 48%




Source: Corporate Site Selection For New Facilities, The
Time Inc. Magazine Company, 1989.







15

Face-to-face, nonroutine communication both within and

outside of the firm is a vital part of the work of

professional employees. Even though communication technology

has made phenomenal advancements in recent years, there are

times when there is no viable alternative or substitute for an

on-site visit. Firms desire frequent contact with their

clients and markets to respond quickly to their changing

needs. Within the firm, management officials, research

scientists and engineers and other administrative service

professionals often need to interact with workers in other

facilities of the organization to deal with problems and new

innovations as quickly as possible. Consequently, frequent

air travel between the various facilities of the firm is quite

common for these employees (Coffey, Polese, 1989).

A comparison of the top location criteria determined by

the study for the various types of company facilities is

revealing (Table 2.3). Most of the lists are identical in

content with only slight changes in rank order of the

criteria.

The back office/processing category, however, is

noticeably different. This activity, traditionally associated

with the headquarters location, is now largely moving away

from central urban locations to, at the very least, suburban

locations where land costs are lower (Ady, 1986; Coffey,

Polese, 1987; Nelson, 1986). The nature of the work--

processing data and paperwork--can be classified as routine









TABLE 2.3
FIVE MOST IMPORTANT FACTORS IN SELECTING LOCATIONS FOR
VARIOUS TYPES OF COMPANY FACILITIES



Back Office/Processing Location

1) Availability of skilled workers
2) Availability of unskilled or semi-skilled workers
3) Efficient transportation facilities for people/employees
4) Easy access to domestic markets, customers and clients
5) Cost of labor


Branch Office/Recional Headauarters Location

1) Easy access to domestic markets, customers and clients
2) Facilitates access to prospective clients
3) Availability of technical or professional workers
4) Availability of skilled workers
5) Easy access to airport


Corporate Headquarters Location

1) Availability of technical or professional workers
2) Easy access to domestic markets, customers and clients
3) Easy access to airport
4) Availability of skilled workers
5) Efficient transportation facilities for people/employees


Sales Office Location

1) Easy access to domestic markets, customers and clients
2) Facilitates access to prospective clients
3) Easy access to airport
4) Efficient transportation facilities for people/employees
5) Availability of technical or professional workers




Source: Corporate Site Selection For New Facilities, The
Time Inc. Magazine Company, 1989.







17
even though it is unrelated to manufacturing in many cases,

such as in the finance and insurance sectors. A life cycle

from nonroutine to routine work also applies to service tasks.

The South Atlantic region of the nation proved to be the

most popular location choice for company facilities as well,

followed in rank order by the Pacific (California, Washington

and Oregon), the East North Central region (Ohio, Indiana,

Michigan, Illinois and Wisconsin) and the Middle Atlantic

region (New York, New Jersey and Pennsylvania). Because costs

are not the most important factor in locating these

facilities, the historically innovative industrial areas of

the United States still show up strongly. Atlanta, Chicago

and Dallas were the most popular urban areas chosen for

location or relocation of company facilities by the firms in

the survey.

One important firm facility not singled out by the

Fortune 500 study was research and development. Research and

development (R&D) employs mainly scientists, engineers and

other technicians who are involved in nonroutine, innovative

research and testing of new products or production processes

for the firm. R&D facilities also have their own set of

location requirements such as the availability of

professional/technical labor, good transportation

accessibility (particularly air) and access to scientific and

technical information (Ady, 1986; Browning, 1980; Lund, 1986;

Malecki, 1979a, 1979b, 1980a, 1980b, 1981). Most firms will







18
locate their R&D facility (or at least one such facility if

there are several) at or very near the corporate headquarters

(Malecki, 1980b).

An earlier study conducted by Dow Jones, Inc. (Browning,

1980) confirms that the important criteria found in the

Fortune 500 survey for the various firm facilities have been

important for at least a decade. The results of this study

are given in Table 2.4.



Regional Development and Industrial Location


Regional development is greatly dependent on the location

decisions of firms (Knapp, Graves, 1989). As the number of

acceptable sites for the various firm facilities has

increased, competition among cities, states and regions to

attract industry has become particularly fierce (Cobb, 1980,

1984; Johnson, 1989). Most nonurban areas have focused on

attracting manufacturing branch plants, as have some

metropolitan areas in low-wage parts of the nation. Most

urban areas, however, are searching for the nonroutine firm

facilities to locate in their city (Thompson, 1987).

According to Jacobs (1984), these nonroutine activities of the

firm are perceived to be more immune to cyclical and

structural changes in the economy and are less likely to

experience severe job loss or closure in bad economic times.









TABLE 2.4
IMPORTANT LOCATION FACTORS BY FACILITY TYPE



Manufacturing Plant

1) labor availability
2) energy availability
3) good highway infrastructure


Distribution Center

1) good highway infrastructure
2) market accessibility
3) labor availability


Corporate Headauarters

1) good air transportation service
2) good highway infrastructure
3) professional talent availability


Regional Office

1) good air transportation service
2) good highway infrastructure
3) professional talent availability


R&D Facility

1) professional talent availability
2) good air transportation service
3) good highway infrastructure


Source: Browning, 1980.









Professional Workers


The professional labor force is indeed highly mobile and

exerts a great deal of influence on corporate location

decisions (Buswell, 1983; Knapp, Graves, 1989). Rapid

technological change has created a greater need for highly

skilled professional workers, thus increasing the significance

of professional labor as a location factor (Weiss, 1985). As

they are highly educated and career-oriented, they have a

greater degree of choice over where they live and work than do

unskilled laborers (Cooper, Makin, 1985; Massey, 1984;

Storper, Walker, 1983).

Research has shown that professional workers have a

strong preference for large urban areas or at least areas

within commuting distance of one (Bradbury, 1988; Malecki,

1987). Cities provide a greater choice of jobs, the ability

to change jobs without changing residences, employment

opportunities for the spouse and greater cultural and

recreational amenities (Herzog, Schlottmann, Johnson, 1986;

Noyelle, Stanback, 1984). According to a recent study, the

dual-career couple has started to exert a great deal of

influence over the location decisions of some firms (Bradbury,

1989). Because of professional workers' narrow range of

preferred locational characteristics, Buswell (1983) and

others state that professional workers are at the same time

geographically immobile (Business Week, 1981).









Labor Migration


Studies have shown that migration of laborers is affected

by changing economic opportunities (Gentile, Stave, 1988;

Greenwood, 1988). Workers leave an area when they feel their

economic prospects are greater elsewhere (Goldfarb, Yezer,

1987). Long and Hansen (1979) found that about half of all

moves in the U.S. were to take a job or seek out new

employment.

Other factors affecting migration are education,

occupation, sex and marital and family status (De Jong,

Gardner, 1981; Kaufman, 1982). Young, well-educated

professional workers are prime candidates for migration

(Greenwood, 1981; Gentile, Stave, 1988). They have fewer ties

to specific areas and are often required to relocate within

their firm to further their careers. Men migrate more often

than women, and single workers more often than married

(especially those with children) (Gentile, Stave, 1988).



The Large Metropolitan Area


For nonroutine and innovative divisions of the firm, the

firm's locational needs and the locational preferences of its

professional workers must be considered together. These

different sets of priorities both restrict and reinforce one

another with the large metropolitan area being the common







22

solution. Despite the fact that there are negative attributes

associated with them (such as higher land costs, taxes and

competition for professional labor for the firm, and higher

housing costs, congestion, pollution and increasing crime

rates for the professional worker), large urban areas are

highly attractive to both firms and workers because of

agglomerative advantages and high quality of life (Dahmann,

1983).

Agglomeration economies are advantages associated with

being in close proximity to markets and services needed by the

firm. Sometimes these advantages arise from the organization

locating in a specific area, such as Silicon Valley, where

their industry or similar industries tend to cluster

(localization economies). These advantages can also occur,

however, to all firms in all industries in larger metropolitan

areas as firms, suppliers and clients interact with one

another (urbanization economies) (Scott, 1988a, 1988b; Watts,

1987).

Large urban areas provide the firm with an ample supply

of professional workers, suppliers, services, information and

important infrastructure (such as airports with frequent air

service to a variety of destinations) and facilitate face-to-

face contact when necessary (Andersson, 1985; Dorfman, 1983;

Malecki, 1987; Oakey, 1985; Scott, 1983). Being located in a

large metropolitan area maximizes the opportunities for a firm

to find the industrial linkages it needs while minimizing the







23

cost of seeking out those linkages (Coffey, Polese, 1989;

Hoare, 1985).

Large urban areas are attractive to professional workers

because they provide a high level of amenities (such as

cultural and recreational) which positively affect their

quality of life. Cutter (1985) defines quality of life as

one's happiness with the physical and social environment based

on how well that environment fits one's personal needs and

desires. This, of course, is a subjective concept (making it

difficult to measure) and can vary greatly from one individual

to another (Canter, 1983; Cutter, 1985). Affecting the

quality of life image for professional workers are such

factors as climate, cultural and recreational amenities,

transportation accessibility, higher educational

opportunities, quality of health care, crime rates, pollution

levels and cost of living (Boyer, Savageau, 1989; Glasmeier,

Hall, Markusen, 1984; Herzog, Schlottmann, Johnson, 1986;

Hsieh, Liu, 1983; Pennings, 1982).

Professional workers can be more selective about where

they live, and do indeed evaluate quality of life factors when

making locational decisions (Markusen, Hall, Glasmeier, 1986;

Power, 1980). Location is often the most important factor in

the rejection of job offers by job-hunting professionals and

transfers within the firm (Collie, 1986; Pinder, 1977). For

these workers, large urban areas offer greater employment

opportunities for themselves and their spouses, the







24
possibility for changing jobs without changing residences and

a generally more satisfying lifestyle.



Conclusion


Professional workers are engaged in nonroutine activities

of the firm. These employees have more control over where

they live and work than do employees of routine activities.

Location of nonroutine activities, therefore, must be a

combination of the residential desires of professional workers

and the specific needs of the firm. These workers seek places

that provide a high quality of life for them and their

families. The firm's needs largely relate to good air and

ground transportation, access to supplies and markets and, of

course, adequate supply of professional labor. Large urban

areas provide the solution, however, not all cities are equal.

This is particularly the case, for example, in air

transportation, which ranked high on the list of locational

needs for all nonroutine activities of the firm. Airline hub

cities have a much greater variety of nonstop destinations and

more frequent departures from their airports than nonhubs.

What follows is a discussion of the air transportation

industry and network in the United States.














CHAPTER 3
THE DOMESTIC AIRLINE INDUSTRY



Introduction


Since the deregulation of domestic airlines in 1978, a

tremendous degree of change has occurred in almost every

aspect of the air travel industry in the United States. As

the industry became deregulated, network structures, fares,

enplanement levels and service quality were greatly affected.

Many new carriers have come and gone (such as Air Florida,

Midway and People's Express), and several older, established

carriers (such as Eastern, Pan Am and Braniff) have

disappeared as well. The first part of this chapter deals

with the deregulated industry and its evolution thus far. The

second part describes the hub-and-spoke network structure,

which is largely a post-regulation phenomenon, and looks at

changes in the location and use of hub cities by the air

travel industry.



The Deregulated Airline Industry



For 40 years the domestic airline industry in the United

States was regulated by the federal government under the Civil







26

Aeronautics Board (CAB). Beginning in 1938, regulation by the

CAB was considered necessary in order to protect and ensure

the success of the industry. The CAB originally also set

safety standards, a task which was later (in 1958) delegated

to the Federal Aviation Administration (FAA) (Bailey, Graham,

Kaplan, 1985; Brown, 1987; Meyer, Oster, Morgan, Berman,

Strassman, 1981). The CAB had the power to decide if new

carriers could enter the air transportation business, to tell

carriers what particular routes they could fly, to regulate

fares, to award government subsidies to carriers who were

forced to fly nonprofitable routes to smaller cities, and to

control mergers and acquisitions (Bailey, Graham, Kaplan,

1985; Brenner, 1988; Brown, 1987; McIntosh, Goeldner, 1990).

For example, if a carrier wanted to increase its fare between

Newark and St. Louis, to begin scheduled service between

Memphis and Charlotte or to discontinue service to Tulsa, it

was required to have CAB approval.

During the 1970s, many economists and politicians began

to argue that regulation was no longer necessary. It was

increasingly felt that the U.S. airline industry was mature

and that government intervention was creating a very

inefficient system (Bailey, Graham, Kaplan, 1985; Brown, 1987;

Cates, 1978; Cooper, Maynard, 1972; Snow, 1977). Removing

the regulatory barriers, it was further argued, would force

the industry to become competitive, bringing about a more

efficient and affordable transportation system to the American









traveler (Meyer, Oster, 1987). Proponents of deregulation

agreed that many changes would occur early in the era of

deregulation that would resemble chaotic instability of the

industry but, with time, things would stabilize and the cost

and quality of air service would greatly improve in the

competitive environment (Goetz, Dempsey, 1989).

The Airline Deregulation Act of 1978 was signed by

President Carter and called for a gradual removal of

government control over domestic air transportation, with the

exception of safety standards (under FAA supervision) and

merger and acquisition approval (monitored by the Department

of Transportation). A gradual phaseout was recommended to

ease the transition to perfect competition in the industry

and, by the end of 1984 (when the CAB was dissolved), the

phaseout was completed. Following the lead of the U.S.,

deregulation of air transportation (or at least partial

deregulation) has been put in action in a few other western

countries, such as Canada and the United Kingdom (Graham,

1990).

A major goal and promise of deregulation was that the

cost of domestic air travel in the United States was to become

more affordable to the average American. This was supposed to

occur because carriers would now compete with one another for

business. Two or more carriers would be flying nearly

identical routes as they could now choose where they would

fly. Also, new airlines were allowed to form at will, which







28

would eliminate the near monopolistic situation of the larger

carriers with respect to total market share of U.S.

passengers. The contestable market theory of air

transportation held that the mere threat of a new entrant in

a market would keep fares down (Leigh, 1990).

During the early 1980s, it appeared that these

predictions were realized (Graham, Kaplan, Sibley, 1983;

Maraffa, Finnerty, 1988; Rose, 1981). Fares dropped

dramatically, allowing more people to afford air travel than

ever before, and over 100 new airlines, mostly small

commuters, began operation (Goetz, Dempsey, 1989). The latter

part of that decade, however, saw unexpected changes that have

caused some to question the wisdom of deregulation (Bauer,

Zlatoper, 1989; Business Week, 1986, 1988b, 1988c; Leigh,

1990; Morrison, Winston, 1989). In fact, discussions of at

least partial reregulation are starting to grow (Kuttner,

1989; Rose, Dahl, 1989).

Many of the new carriers that were supposed to challenge

the dominant market shares of the major carriers,

unfortunately, were short-lived. Despite the fact that new

carriers were successful in the early years of deregulation in

reducing the share of revenue passenger miles of the larger

airlines, in the long run most all were squeezed out (Goetz,

Dempsey, 1989). It is unlikely that there will be a lot of

new airlines entering the industry to replace them, as







29

investor confidence in new airline ventures has lessened

considerably (Dempsey, 1987; Rose, Dahl, 1989).

These new smaller carriers had several strikes against

them from the beginning. For example, the essential

infrastructure of gates, terminal facilities and landing slots

has been controlled by the major carriers, making it harder to

enter new markets. Almost 70% of all U.S. airports have no

gates at all that could be leased to new carriers--although

space may sometimes be sublet from other carriers (Hardaway,

1986). Because of their greater buildup of capital, the major

carriers have a higher chance of surviving fare wars, and they

have used this power to drive both new airlines and old,

nearly bankrupt carriers out of business. For those small

carriers that did not arrange marketing agreements with the

large carriers, much of their business was taken away by the

latter due to frequent flyer program loyalty (particularly by

business travelers) (Toh, Hu, 1988) and computerized

reservation systems made available to travel agencies (Davis,

1982; Oster, Pickrell, 1988).

A glance at the market-share data before and after

deregulation shows the trend in market dominance by the

largest carriers. At the end of regulation in 1978, the top

six airlines controlled 71% of the domestic air traffic. In

1983, that figure dropped to 65% due largely to heavy influx

of new carriers, but by the end of 1987 it had risen to 79%

(Brenner, 1988) and is rising going into the 1990s (Fortune,









1990). Contributing to this rise was the wave of acquisitions

and mergers that occurred during the mid 1980s. The larger

carriers began to realize that the best way to survive was to

increase their market shares and to neutralize their

competition by merging with other carriers (Carlton, Kaplan,

Sibley, 1983).

The main, and largely unforeseen, problem that has arisen

out of the merging has been the increased control gained by

individual airlines over travel in and out of certain cities.

The merger of Northwest and Republic Airlines, for example,

and TWA with Ozark created market shares of over 80% at

Minneapolis/St. Paul and at St. Louis for Northwest and TWA,

respectively (McGinley, 1989). Between 1985 and 1988, the

number of airports served by at least four airlines dropped by

52%, while the number of cities served by only one carrier

increased by 25% (Kuttner, 1989). Nearly two-thirds of all

route destinations are now near-monopolies (Traffic World,

1988). This dominance has permitted fare increases, rather

than the lower fares predicted by deregulation enthusiasts

(Bauer, Zlatoper, 1989; Business Week, 1988b). For example,

fares out of St. Louis rose twice as fast as the national

average after TWA purchased Ozark (McGinley, 1989).

Another change has been the dramatic decline in service

to some urban areas. Some cities are simply harder to get to

since deregulation, prompting passenger complaints but little

response from either airlines or the federal government.









Small cities have been most adversely affected, both in number

of nonstop destinations reached from their facility and in

seating capacity (Chan, 1982; Ivy, 1991; Kiel, 1989; Maraffa,

Kiel, 1985; Warren, 1984). Airlines have often replaced

larger jet aircraft with commuter turboprop planes with less

seating capacity on existing routes. Out of 522 small cities

that received scheduled commercial air service in 1978, 62%

experienced a decline in flight frequency, and nearly 30% of

those experienced complete loss of service (Goetz, Dempsey,

1989).

Perhaps the most important issue to passengers is safety

of air travel. This has become a major public issue because

of the seemingly high number of accidents and mechanical

failures of aircraft reported during the past decade.

Government statistics lead one to believe that air travel is

actually safer in the deregulated skies. Compared with the

decade prior to deregulation, jet accidents through 1987

declined by 36%, while the number of fatalities from air

crashes declined by 40% (Moses, Savage, 1988). What such data

do not show, however, is that the number of near-accidents has

been on the rise because of delayed mechanical attention of

aircraft and increased congestion, both in the air and along

runways due to the competitive nature of the deregulated

industry. The number of near-accidents has risen from 568 in

1980 to 1056 in 1987 (Moses, Savage, 1988).







32

According to the Department of Transportation, the amount

spent on aircraft maintenance dropped 30% during the first six

years of deregulation (Valente, McGinley, 1988). Moreover, a

survey of commercial airplane pilots revealed that almost half

believe that their companies defer maintenance of their fleets

too long. In addition, the average age of the industry's jets

increased by 21% since 1979, with more than half of the jets

in service being 16 years or more old in 1988 (Valente,

McGinley, 1988). Older fleets typically require more

maintenance and repairs than do newer aircraft.

The general consensus of the flying public is that the

quality of air service has greatly declined since

deregulation. A recent survey of consumers found that 50%

felt that such service had declined significantly, while under

20% expressed feelings of improvement. The FAA has reported

a soaring number of consumer complaints against the airlines

(Consumer Reports, 1988). Deregulation and the resulting

network geography of the major carriers have not delivered the

promised benefits to all cities and passengers alike

(Anderson, Kraus, 1981; Ippolito, 1981). As already

mentioned, small cities, on the whole, were adversely affected

in terms of seating capacity, fewer non-stop choices and less

frequent service. However, many large cities that were not

selected as hubs by major airlines (such as New Orleans,

Buffalo, and Louisville) have experienced service decline--

particularly in the number of non-stop destination choices







33

reached from each city. In terms of flight frequency,

destination choice and ticket prices, residents of cities

chosen as a hub by more than one carrier (such as Chicago,

Dallas/Ft. Worth, and Atlanta) have probably benefited most

from deregulation. Because they are hub cities, they offer a

multitude of non-stop destination choices at various times

during the day. Having more than one carrier using an airport

as a major transfer hub creates competition along many non-

stop routes, therefore, keeping fares competitively lower.

For example, while the national average annual fare change at

hub airports from 1985 to 1988 was a growth of 4%, Atlanta

(which served as a hub for Eastern and Delta during that

period) experienced a growth of only 1.5% (McGinley, 1989).

The 1980's saw great instability in the industry, and the

1990's (thusfar) are showing no signs of stabilization. Many

airlines are still in financial trouble (Continental, America

West and TWA are operating under bankruptcy protection as of

early 1992). Soaring oil prices, the current recession and

intensifying competition from the three giants of the industry

(American, United and Delta) are the main sources of the

problem (Business Week, 1991a, 1991b, 1991c, 1991d, 1991e,

1991f; Fortune, 1990). Now that over a decade has passed

since deregulation, it is increasingly clear that air

travelers will likely pay much more for lower quality, less

convenient and perhaps more risky transportation by the U.S.

airlines.









The Hub-and-Spoke System



The hub-and-spoke system currently in place in the U.S.

airline industry is largely a product of deregulation. It is

the new structure adopted by most airlines to compete more

effectively. The major carriers created hubs at strategic

regional points in their air networks so travelers from

numerous origins (spokes) could be routed into the hub city

and then connected with a flight to their final destinations

(Business Week, 1988a). The hub-and-spoke structure cuts

costs by creating greater overall efficiency with more

occupied seats. Route planners, therefore, approach monthly

scheduling differently today than prior to deregulation, as

such planning is now more important to the profitability of

the company than previously (Pollack 1977, 1982).

What few hubs that existed in the regulated era had a

slightly different function than hubs today. Chicago, Denver

and St. Louis, for example, were used to serve long-haul,

east-west markets which used aircraft that did not have

transcontinental capabilities (Lopuszynski, 1986). Thus,

these cities were set up as transfer points for coast-to-coast

traffic. Because of CAB control over routes flown, it would

have been difficult for an airline to set up a hub-and-spoke

network prior to deregulation.

The deregulated era saw many true hubs develop that acted

as collection points for travelers from many origins, sending









them off to many destinations. This new system, along with

increased cooperation and code-sharing between major carriers

and commuter carriers, made a much greater on-line (same

carrier) city-pair matching available to air travelers (Oster,

Pickrell, 1988).

The first true hub in this sense was developed by Delta

Airlines in Atlanta, and actually was in place before

deregulation (Lewis, Newton, 1979). Delta dominated traffic

in and out of Atlanta for decades and gradually added more and

more service as the years passed. It was the model system

copied by other airlines as deregulation forced them towards

greater efficiency in order to compete. Figure 3.1 shows the

major hub cities (in mid-1991) of the largest carriers.

Much has been written on the positioning of hubs in the

air transportation network (Bauer, 1987; Grove, O'Kelly, 1986;

Kanafani, Ghobrial, 1985; Lopuszynski, 1986; O'Kelly, 1986a,

1986b; Toh, Higgins, 1985). What makes one city likely to be

chosen as a hub over another city? In general, the very

largest cities have been favored over medium-sized cities, and

eastern cities over western cities. In fact, several cities

have been chosen as hubs by more than one carrier, such as New

York City, Chicago, Denver and Dallas.

The early trend of choosing the largest cities of the

nation as hub cities eventually died as these airports became

overly congested. Not only was safety an issue, but delayed

flights, overuse of infrastructure, little room for expansion





























































FIGURE 3.1
HUB CITIES OF THE MAJOR U.S. AIRLINES: 1991







37

and oversaturated markets plagued some of the hubs. Carriers

began to look for medium-sized cities as their transfer

points. Often these were strategically located near one of

the large, congested hubs to give travelers a pleasant

alternative to places like Chicago O'Hare. Piedmont Airlines,

later acquired by USAir, made an early success by choosing

Charlotte as its southeastern hub, allowing passengers to

bypass chronically congested Atlanta's Hartsfield (Davis,

1982). Many other carriers followed Piedmont's lead and

expanded their air networks by adding new hubs. This

explains, for example, American's new hubs at Nashville and

Raleigh/Durham. Competition among airports for hub selection

by a major carrier has become intense. The expanded service

creates big business for the chosen facility and city, and

airport planning boards are more aggressive now than in the

past (Barrett, 1987; Butler, Kiernan, 1987; Insight, 1988).

Lopuszynski (1986) has identified some common

characteristics of hub cities. It is an ex post facto list

that looked at hub cities already in use to find similarities

among them. Most hubs were found to contain several of the

following characteristics: 1) a sizeable population force

with strong business and commercial opportunities 2) a good

geographic location with respect to other population centers,

physical terrain and weather patterns 3) good airport

facilities with adequate room for expansion of gates and

runways 4) a strong economy and balanced workforce 5) air







38

service competition at a minimum acceptable level 6)

avoidance of existing major hubs.

While the hub-and-spoke strategy has created greater

overall efficiency and cost-cutting benefits for the airlines,

the very essence of the system creates problems of congestion,

environmental problems such as noise and air pollution, strain

on infrastructure and overworked air traffic controllers.

The hub-and-spoke system requires that airlines

concentrate many incoming and outbound flights during a narrow

time frame to maximize the number of city-pair combinations

that can be served. The resulting congestion has been too

much for most airports to handle. For example, Atlanta's

Hartsfield International Airport has an optimum capacity of 21

arrivals every 15 minutes, as determined by the FAA; however,

32 arrivals were recently scheduled by the airlines between

9:00 a.m. and 9:15 a.m. (Morganthau, 1987). At Chicago's

O'Hare, 42 departures were scheduled between 7:00 a.m. and

7:15 a.m., despite a capability to deal safely and efficiently

with only 23 (Morganthau, 1987). Increases in the frequency

of flights since 1978 at several hubs have been quite large.

Baltimore (+94.2%), Dallas/Ft. Worth (+52.2%), Houston

(+85.1%), Minneapolis/St. Paul (+60.1%) and Salt Lake City

(+51.9%) grew by more than half between 1978 and 1984, while

flight frequencies more than doubled at Charlotte (+134.2%)

and Newark (+114.2%) (Report on Airline Service, 1984).









Another problem of the hub-and-spoke strategy has been

the reduction in the number of nonstop flights from cities

other than hub cities. For many city pairs it now takes much

longer en route because virtually all flights are routed

through at least one hub (Goetz, Dempsey, 1989). The

opportunity cost of time is a factor that many travelers

consider important. In general, small cities and some of the

large spoke cities have often ended up worse off than they

were prior to deregulation as they have lost some (and in a

few cases all) nonstop service to cities that now have to be

reached through hubs (Bauer, 1987; Ivy, 1991).

The advantage for hub-city residents of a large number of

nonstop destinations is counterbalanced with some

disadvantages. These passengers generally pay much higher

fares and have less choice of carriers for most destinations

than was the case prior to deregulation (Bauer, 1987;

Borenstein, 1989; McGinley, 1989; Toh, Higgins, 1985). Single

carrier dominance (Table 3.1) and higher fares are the price

paid by residents for greater flight frequency and nonstop

destination choice. Hub cities have been hit the hardest with

fare increases since their residents are, in effect,

subsidizing longer haul passenger fares. Passengers flying

from Norfolk to Kansas City through USAir's Charlotte hub, for

example, would likely be charged the same fare as passengers

travelling on USAir from Charlotte to Kansas City only.

Today, flights originating at or departing for hub cities are








TABLE 3.1
SINGLE CARRIER DOMINANCE AT HUBS


city airport 1978 1990
code


Atlanta
Baltimore
Charlotte
Chicago
Cincinnati
Cleveland
Dallas
Dayton
Denver
Detroit
Honolulu
Houston
Las Vegas
Los Angeles
Memphis
Miami
Minneapolis
Nashville
New York
Newark
Orlando
Philadelphia
Phoenix
Pittsburgh
Raleigh/Durham
St. Louis
Salt Lake City
San Francisco
Seattle
Washington, D.C.


40.8%
26.2%
66.1%
27.6%
34.9%
57.6%
36.3%
35.3%
25.9%
23.6%
37.5%
24.6%
34.9%
26.2%
33.5%
36.4%
30.1%
27.1%
20.4%
28.2%
43.4%
28.0%
26.6%
51.3%
59.4%
34.6%
37.4%
42.4%
31.8%
24.3%


53.9%
71.5%
91.1%
35.5%
77.9%
38.7%
48.8%
74.8%
42.2%
65.4%
35.5%
50.4%
48.6%
17.8%
53.6%
18.3%
75.3%
57.5%
18.0%
48.9%
30.5%
46.3%
43.5%
83.0%
65.1%
74.3%
76.4%
29.1%
20.6%
23.4%


Source: Calculated from
Certified Route
Transportation,
1978, 1990.


Airport Activity


Statistics of


Air Carriers, U.S. Department of
Federal Aviation Administration,







priced up to 50% more than they would have been had

deregulation not occurred (Stockton, 1988). Comparison of

one-way fares on selected pairs indicate increases from 1978

to 1988 (in constant dollars) ranging from 28% to 495%, with

an average increase of well over 200% (Goetz, Dempsey, 1989).



Conclusion


Deregulation has had a much greater impact on the

domestic airline industry than anyone anticipated. Major

changes in service and connectivity have occurred in the U.S.

air transportation network. Small cities, and even large

cities not chosen as hubs, have experienced declines in flight

frequency and nonstop destination choice. Hubs have grown

tremendously in these areas, but fares have skyrocketed and

carrier choice along most routes has lessened. However, not

all hubs are alike. The next chapter will look at different

types of hubs and discuss hub intensity and connectivity.













CHAPTER 4
VARIATIONS IN HUB SERVICE IN THE DOMESTIC
AIR TRANSPORTATION NETWORK



Introduction



A hub is generally defined as a central collection point

or node in a transportation system or network. Usage of the

term, however, has become particularly applied in the air

transportation industry of the United States, largely since

deregulation and the advent of the hub-and-spoke system

discussed in the previous chapter. In reviewing air

transportation literature and data sources (both academic and

popular), one encounters the word frequently. Close scrutiny

of places labeled as air hubs, however, reveals that not all

hubs are equal in the service they offer, in either intensity

or connectivity. In fact, some hubs are vastly different from

others. Such variations can make the work of the air

transportation researcher problematic. This chapter will

identify different types of air hubs based on existing usage

of the term by the airlines and the Federal Aviation

Administration, and explore variations in hub intensity and

connectivity within the domestic air transportation network.









The FAA Hub



The initial usage of the word "hub" in the air

transportation industry was designated by the Civil

Aeronautics Board (now disbanded), and continued by the

Federal Aviation Administration (FAA). The FAA classifies all

communities with scheduled commercial air service as one of

four types of hub, and categorizes each hub based on its share

of the nation's annual enplanements. Data for multiple-

airport cities (like Chicago and Dallas) are usually summed to

represent a community total.

Large hubs, such as Atlanta and St. Louis, represent at

least 1.0% of the nation's total annual enplanements. New

Orleans and Norfolk, for example, are classified as medium

hubs since their share of total U.S. annual enplanements is

between .25% and .99%. Communities that represent between

.05% and .24% (like Richmond, VA) are classified as small

hubs, while non-hubs (such as Gainesville, FL) enplane less

than .05% of the nation's total passengers (Airport Activity

Statistics of Certified Route Air Carriers, 1990).



Deregulation of the U.S. Airline Industry


As discussed in the previous chapter, the Airline

Deregulation Act of 1978 forced the industry into a

competitive market situation. While airlines were adjusting







44

to the new environment, their network geography (node-linkage

association) was changing accordingly. Flow efficiency and

cost-reduction were made higher priorities. It became

increasingly clear that concentrating flights at one or more

key regional nodes in their air transportation networks would

raise the seat-occupancy levels, thus maximizing the usage of

aircraft. Such concentration would also maximize the number

of on-line (same carrier) city-pair matching available to

passengers. These central nodes typically offer non-stop

service to most every large and medium-sized city in the

nation and to smaller cities within the local region of the

node. Numerous arrivals and departures are scheduled within

a short time frame to allow the connections. This intense

type of system has become known as a "hub-and-spoke" network,

and now dominates U.S. air transportation. Table 4.1 lists

the U.S. hubs as designated by the major domestic carriers.

The list was compiled by consulting the fall 1991 schedules

and route maps of the carriers, and was confirmed by telephone

inquiry to each airline's public relations department. Not

included in Table 4.1 are a few cities labeled by some

carriers as mini-hubs. These are large cities that do not

serve as major transfer points within any airline's network,

but do offer some nonstop service (more than other spoke

cities) on an individual carrier (for example, Northwest

operates a mini-hub at Milwaukee and USAir uses Kansas City as










TABLE 4.1
FALL 1991 HUB CITIES OF THE MAJOR U.S. CARRIERS


city carriers) non-local
traffic


Atlanta
Baltimore
Charlotte
Chicago
Cincinnati
Cleveland
Dallas/Ft. Worth
Dayton
Denver
Detroit
Honolulu
Houston
Las Vegas
Los Angeles
Memphis
Miami
Minneapolis
Nashville
New York/Newark
Orlando
Philadelphia
Phoenix
Pittsburgh
Raleigh/Durham
St. Louis
Salt Lake City
San Francisco
Seattle
Washington


Delta
USAir
USAir
American, Midway, United
Delta
Continental
American, Southwest
USAir
Continental, United
Northwest
Continental, United
Continental, Southwest
America West
Delta
Northwest
American, Pan Am
Northwest
American
Continental, Delta, TWA
Delta, United
USAir, Midway
America West
USAir
American
TWA
Delta
United
Northwest
United


Note: Since the compilation of this list, Midway and Pan Am
Airlines have ceased operations, and USAir has planned the
closing of its Dayton hub in 1992.

Sources: Hub lists were obtained from the Fall 1991 schedules
and route maps of the major carriers. Transfer percentages
were supplied by the U.S. Department of Transportation for the
year ending 1990. Percentages were not separated in multiple-
airport communities.


67.60%
38.19%
74.19%
50.34%
53.86%
24.13%
60.04%
48.04%
50.19%
38.73%
31.72%
32.61%
22.27%
40.34%
62.92%
25.20%
47.76%
40.96%
21.87%
13.75%
22.04%
32.80%
60.64%
61.84%
55.33%
60.25%
24.40%
28.84%
20.03%







46

such). This service was built up to meet travel demand in

markets that some airlines considered to be underserved.



Hub-and-Spoke Growth and Development



The first domestic hub-and-spoke hubs chosen by

individual airlines were at airports that were already used by

carriers as connection or terminus points for long-haul, east-

west traffic using aircraft that did not have transcontinental

capabilities (Lopuszynski, 1986). These pre-deregulation

connecting airports were in the larger cities in the U.S., and

as such, generated high levels of local traffic. Individual

airlines tended to focus their early hub-and-spoke strategies

on those large-city connecting airports in which they already

controlled most of the scheduled departures. Most of these

airports were already well-connected (with non-stop service)

to other large cities in the nation. The hub-and-spoke hubs

were created as the carriers simply added many more flights to

many more destinations or spokes (particularly medium-sized

and small cities) at these facilities to build up connectivity

and create greater overall efficiency and market control at

the hub. In some cases, particular airports were selected by

more than one carrier as a hub-and-spoke hub. American and

United, for example, both created hubs at Chicago O'Hare,

while Atlanta became a major transfer city for Delta and the

now-defunct Eastern.







47

As these initial connecting hubs became saturated with

traffic, other (often medium-sized) cities were chosen as

transfer points, as new airlines developed and as older

airlines expanded their networks. Charlotte, Raleigh/Durham

and Nashville, for example, were developed as hubs around

Atlanta offering travelers a less-congested alternative in the

Southeast. Today, the largest carriers have as many as four

or five hubs scattered throughout the nation. One can divide

the list of hubs from Table 4.1 into two broad categories: 1)

hubs that were important connecting airports prior to

deregulation (although much less intensely connected than

today), and 2) hubs that achieved important transfer status

after deregulation (Table 4.2).



Service Variations



Closer examination of the cities on the hub list,

however, reveals vast differences among the airports' service

levels and functions. In fact, not all of the locations

listed in Table 4.1 really function as hub-and-spoke hubs. As

mentioned, hubs of this type are characterized by many non-

stop flights to cities of all sizes. Also, the very nature of

the network structure suggests that a good portion of the

traffic at these hubs should be nonlocal, if the hub is indeed

successful as a transfer point. Some of the airports labeled

by the major carriers as air hubs do not have true









TABLE 4.2
PRE AND POST-DEREGULATION HUB CITIES
(1991)


PRE-DEREGULATION HUBS:


Atlanta
Chicago
Dallas/Ft Worth
Denver
Honolulu
Houston
Los Angeles
Miami


Minneapolis
New York/Newark*
Philadelphia
Pittsburgh
St. Louis
San Francisco
Seattle


POST-DEREGULATION HUBS:


Baltimore (1983)
Charlotte (1981)
Cincinnati (1987)
Cleveland (1989)
Dayton (1982)
Detroit (1984)
Las Vegas (1985)


Memphis (1984)
Nashville (1986)
Orlando (1989)
Phoenix (1983)
Raleigh/Durham (1987)
Salt Lake City (1982)
Washington, D.C. (1986)


* Newark International Airport on its own would be classified
as a post- deregulation hub. People's Express (eventually
consumed by Continental Airlines) developed hub facilities
there in 1981.


Source: Hub information was obtained from a series of surveys
of airports and airlines by mail and telephone (in mid-1991).
The years next to the post-deregulation hubs indicate the year
of development as a major transfer hub as determined by the
airport in question.







49

hub-and-spoke functions, at least not at the same level of

intensity as others.

A few of the cities listed as hubs function as feeder

points for a specific carrier's international network. They

are usually well connected with the individual airline's other

domestic hub cities, and also offer nonstop service to and

from the largest cities (largest markets) in the nation.

Thus, they do act as transfer points for the carrier, but on

a less intense level. They are certainly not important

transfer points within the domestic air transportation

network. Los Angeles, Miami, New York (JFK), Seattle and San

Francisco are all international gateway hubs instead of hub-

and-spoke hubs for domestic air networks. These peripherally

located cities have a lower percentage of nonlocal traffic at

their airports than almost every other hub on the list (Table

4.1). This means that a smaller proportion of their traffic

is changing aircraft at their facility bound for a different

final destination. In addition, the particular carriers)

claiming hub status at each of the five facilities listed

above offer rather limited nonstop service from them,

especially in comparison to other domestic hub-and-spoke hubs.

Table 4.3 looks at nonstop connections between hubs listed in

Table 4.1 and FAA hub airports of various sizes. It is clear

that some hubs on the list (particularly the international

gateway hubs) are not as well connected to cities (spokes) of

all sizes as are many of the other hubs.










TABLE 4.3
NONSTOP DOMESTIC CONNECTIONS ON ALL CARRIERS BY FAA HUB-TYPE
(FALL 1991)



airports large medium small non total



ATL 27 23 33 24 107
BWI 22 16 10 19 67
CLT 24 18 27 25 94
CHI 27 29 35 27 119
CVG 27 17 19 7 70
CLE 22 13 7 3 45
DFW 27 26 25 23 102
DAY 17 6 7 6 36
DEN 25 19 16 24 84
DTW 25 16 14 14 70
HNL 11 2 3 5 21
HOU 26 16 10 9 61
LAS 22 12 3 1 38
LAX 27 17 5 14 63
MEM 23 10 14 20 67
MIA 20 10 4 4 38
MSP 26 13 11 30 81
BNA 24 14 18 14 70
NYC* 26 17 17 17 78
MCO 22 13 8 4 47
PHL 23 18 12 19 72
PHX 24 16 6 9 55
PIT 25 20 20 33 98
RDU 15 12 16 12 55
STL 27 22 16 19 83
SLC 18 12 7 17 55
SFO 26 14 8 14 62
SEA 22 7 4 14 47
DCA 25 17 23 19 84


*Includes Newark


Source: OAG Pocket Flight Guide, November 1991.







51

An emerging category is the destination hub. These are

cities that are popular travel destinations, and as such

offer great deal of nonstop service to and from large and

medium-sized cities around the nation to meet the high demand.

Orlando, Las Vegas and Honolulu fall into this classification

(although Honolulu also serves as an international connecting

point). All three generate comparatively low nonlocal traffic

rates (Table 4.1), and are not well connected to cities of

varying size (Table 4.3). Few passengers flying into these

cities are transferring to other domestic locations, and these

are mainly via commuter carriers from the local area. Some of

the international gateway hubs could be considered destination

hubs as well (like Los Angeles and New York) due to their

popularity as travel destinations.



Hub Connectivity


Because today's hubs (and the carriers operating them)

basically compete with one another for transfer traffic, the

success of a hub usually depends on how well connected it is

to other nodes in the U.S. air transportation network. The

hub airport (and carrier) that has a greater variety of

nonstop service to different sized cities in all parts of the

nation is in a better position to attract passengers and

control markets. Atlanta competes with Charlotte in the

Southeast, for example, and Dallas, Houston and Denver









basically compete in the West for the same transfer

passengers. What follows is a discussion of the measurement

of hub connectivity using a graph theoretical approach. Graph

theory allows one to view the network as a topological map

comprised of nodes (points of economic concentration) and

linkages (routes that connect two nodes). Thus, it shows

network connectivity in a relative sense. This technique

illustrates the strength (in terms of connectivity) and

attractive power of the hubs listed in Table 4.1.



Analyzing Hub Connectivity


One graph theoretic method of measuring the connectivity

or accessibility of a node begins with the construction of a

binary matrix that represents the network abstracted as a

graph (Lowe, Moryadas, 1975; Taaffe, Gauthier, 1973). It is

a square matrix in which the number of rows and columns each

represent the number of nodes in the transportation network.

The horizontal rows represent origin nodes, while the vertical

columns represent destination nodes. Both rows and columns,

of course, contain the same list of points. The cell entries

of the matrix are assigned a value of either one or zero. A

value of one shows the presence of a direct (nonstop) linkage

between specific nodal pairs, while a value of zero indicates

the absence of such a linkage. Nodes are not considered to be

connected to themselves; therefore, the principal diagonal of







53

the matrix (all i,i entries) contains zeroes as cell entries.

Thus, the connectivity matrix (Cl) shows first order

connections in a transportation network.

A vector of values that can be used as a crude measure of

nodal accessibility is obtained by summing the individual rows

or columns of the matrix. The higher the summed row or column

value of the node, the greater the accessibility of the point.

This accessibility index on its own is of limited usefulness,

however, because we are often interested in both direct and

indirect connections.

The number of indirect connections in a network can be

determined by powering the original connectivity matrix (Lowe,

Moryadas, 1975; Taaffe, Gauthier, 1973). The matrix (Cl) can

be multiplied by itself (resulting in matrix C2) to look at

second order or two-step connections (connections that pass

through an intermediate node) in the network. Likewise, the

third order connectivity matrix (C3) is obtained by

multiplying matrix Cl by matrix C2. To take all indirect

connections into account, the matrix should be powered to the

Nth order (CN), where N represents the diameter of the

transportation network. (The diameter is defined as the

shortest topologic distance between the two most distant nodes

in the network.) At this stage, all zero elements disappear

from the matrix indicating that all nodes are connected.

Summing the matrix Cl with the powered matrices shows the

total accessibility of each node within the network







54
(accessibility matrix). The individual rows or columns of

this accessibility matrix (T) can be summed to yield the gross

vertex connectivity for each node in the network.

Early researchers using the above-mentioned technique

(Garrison, 1960; Pitts, 1965), discovered that while powering

the matrix did give the maximum number of alternative paths in

a network, a number of redundancies (passing through the same

node more than once) were included in the final accessibility

matrix (T). This was particularly the case for nodes that

were directly connected in the original connectivity matrix

(Cl).

Another criticism by Garrison (1960) was that all

linkages should not be considered equal in importance. The

more indirect the linkage, the less it should add to the gross

vertex connectivity number. He introduced a procedure in

which a scalar number (s) that takes on a value between zero

and one is multiplied by the accessibility values in each

matrix powered according to the order of the matrix

[T=sC1+s2C2+s3C3+...+s"CN]. The real problem lies in assigning

the scalar value (Garrison's scalar of .3 was assigned

arbitrarily). The technique, however, does lessen the

importance of indirect connections relative to direct

connections and at the same time reducing the impact of

redundant paths. A new scalar method will be introduced in

the next section.









Results



The above-mentioned technique was applied to a study set

of 117 nodes to measure accessibility within the U.S. domestic

air transportation network. These nodes were all U.S. urban

areas (excluding 4 Hawaiian cities not well connected to the

U.S. mainland) that are classified by the FAA as a large

(total of 28), medium (total of 29) or small (total of 60) hub

(Table 4.4). The purpose was to find out how well connected

each of the airline hubs listed in Table 4.1 was to cities of

various sizes scattered around the nation. Due to the fact

that there are several hundred FAA non-hubs, they were

excluded to keep the size of the matrix manageable (117 x

117). The connectivity data for the original matrix (Cl) was

abstracted from the November 1991 issue of the OAG Pocket

Flight Guide.

The matrix was ordered to the third power (C3). At that

point, all of the non-zero elements disappeared from the

matrix cells. The diameter of this network is three because

one can fly between any domestic city pair in the study set in

three or fewer flight segments (due to the intense hub-and-

spoke structuring). The summed accessibility indices for each

metropolitan area for matrices Cl, C2, C3, and T are given in

Appendix A. The accessibility indices for each hub (from

matrix Cl and T) and their respective rankings are given in

Table 4.5. Note that the hub rankings for matrix Cl (direct









TABLE 4.4
FAA HUBS


Atlanta
Baltimore
Boston
Charlotte
Chicago
Dallas/Ft. Worth
Denver
Detroit
Honolulu
Houston
Kansas City
Las Vegas
Los Angeles
Memphis


Albuquerque
Austin
Buffalo
Cincinnati
Cleveland
Columbus
Dayton
El Paso
Ft. Myers
Hartford
Indianapolis
Jacksonville
Kahului
Lihue
Milwaukee
Nashville


FAA Large Hubs

Miami
Minneapolis/St. Paul
New York/Newark
Orlando
Philadelphia
Phoenix
Pittsburgh
St. Louis
Salt Lake City
San Diego
San Francisco
Seattle
Tampa
Washington, D.C.




FAA Medium Hubs

New Orleans
Norfolk
Oklahoma City
Ontario
Portland
Raleigh/Durham
Reno
Rochester
Sacramento
San Antonio
San Jose
Syracuse
Tucson
Tulsa
West Palm Beach










TABLE 4.4--continued



FAA Small Hubs


Akron/Canton
Albany
Allentown
Amarillo
Anchorage
Baton Rouge
Billings
Birmingham
Boise
Brownsville
Burlington
Cedar Rapids
Charleston (SC)
Charleston (WV)
Chattanooga
Colorado Springs
Columbia
Corpus Christi
Daytona Beach
Des Moines
Eugene
Fort Wayne
Fresno
Grand Rapids
Greensboro
Greenville
Harrisburg
Hilo
Huntsville
Islip
Jackson


Kailua-Kona
Knoxville
Lexington
Lincoln
Little Rock
Louisville
Lubbock
Madison
Melbourne
Midland
Mobile
Moline
Myrtle Beach
Omaha
Palm Springs
Pensacola
Portland
Providence
Richmond
Roanoke
Saginaw
Santa Barbara
Sarasota
Savannah
Shreveport
Sioux Falls
South Bend
Spokane
Tallahassee
Toledo
Witchita











ACCESSIBILITY


TABLE 4.5
INDICES FOR MATRICES C1 AND T


airports matrix Cl rank matrix T rank


ATL
BWI
CLT
CHI
CVG
CLE
DFW
DAY
DEN
DTW
HNL
HOU
LAS
LAX
MEM
MIA
MSP
BNA
NYC*
MCO
PHL
PHX
PIT
RDU
STL
SLC
SFO
SEA
DCA


93552
69753
83769
97950
82183
63749
88668
47457
71207
76515
20889
69597
53701
68678
63525
55436
70247
74299
81114
65272
74469
62542
83524
54992
81844
47692
64887
48591
82696


*Includes Newark







59
connections only) and matrix T (direct and indirect

connections) are slightly different (Spearman's rank order

coefficient of .979). Twelve of the cities (mostly ranked in

the top ten) remain at the same rank, but ten rise in the

rankings (Baltimore, Cincinnati, Cleveland, Detroit, Miami,

Minneapolis, Orlando, Philadelphia, San Francisco and Seattle)

while seven fall (Denver, Houston, Memphis, Phoenix,

Raleigh/Durham, St. Louis and Salt Lake City) when indirect as

well as direct connections are taken into account (matrix T).

The rankings show an eastern bias because more of the 117

nodes in the original matrix (Cl) were located in the East or

Midwest than in the western half of the United States.

Scalar multiplication was also performed on the original

and powered matrices. Table 4.6 shows the accessibility

matrix (T) indices using a variety of scalars. While the

magnitude of difference between the cities, of course, changes

as the scalar changes, the specific rank order of the cities

remains constant (concordant). It is also the exact rank

order of the unsealed accessibility matrix T (Table 4.5).

A more refined weighting technique assigns a different

scalar value to each of the 117 nodes in the network. This

weight is based on the individual node's share (from the last

column in Table 4.3) of total direct connections in the

network (1,969). One advantage of this weighting procedure is

that it allows the FAA nonhub connections to be taken into

account, albeit in an indirect way. The results of this











ACCESSIBILITY INDICES


TABLE 4.6
OF MATRIX T


WITH DIFFERENT SCALARS


airports s=.25 s=.3 s=.5 s=.75


CHI
ATL
DFW
CLT
PIT
DCA
CVG
STL
NYC*
DTW
PHL
BNA
DEN
MSP
BWI
HOU
LAX
MCO
SFO
CLE
MEM
PHX
MIA
RDU
LAS
SEA
SLC
DAY
HNL


1615
1542
1464
1380
1376
1362
1354
1350
1337
1260
1228
1225
1180
1161
1150
1149
1136
1077
1075
1051
1049
1036
916
906
889
807
791
783
347


2755
2630
2497
2354
2347
2324
2310
2303
2282
2150
2094
2090
2010
1980
1962
1961
1937
1838
1832
1793
1789
1765
1562
1546
1516
1374
1348
1336
592


12447
11886
11274
10642
10613
10508
10443
10405
10310
9723
9466
9444
9064
8938
8866
8854
8742
8302
8262
8104
8080
7963
7053
6989
6838
6192
6075
6035
2664


41543
39676
37614
35527
35424
35074
34856
34718
34407
32453
31589
31516
30219
29807
29588
29530
29146
27693
27539
27042
26951
26544
23523
23325
22793
20629
20244
20134
8871


*Includes Newark







61
weighting procedure are given in Table 4.7. The rankings are

slightly different from both the original connectivity matrix

Cl (Spearman's rank order coefficient of .979), and the

unweighted matrix T (Spearman's rank order coefficient of

.951). Cincinnati and St. Louis, for example, are closely

ranked in matrices Cl and T from Table 4.5, but the weighting

procedure requiring the calculation of a different scalar for

each individual node puts a greater difference in rankings

between these two cities (Table 4.7). This is because St.

Louis is connected to a greater total number of cities (once

nonhubs are included), and therefore, fares better in the

final ranking of accessibility numbers.



Connectivity and Intensity Classification Scheme


Using the accessibility information from Table 4.7 and

transfer traffic information from Table 4.1, the following

classification scheme measuring hub strength was derived.

Hubs are classified as super, maior--tyve A, maior--tpe B,

moderate--tye A, moderate--type B, minor--type A, minor--tye

B or non-hub (Table 4.8). The eight classifications were

made using a one-dimensional iterative partitioning clustering

method using the accessibility numbers given in Table 4.7

(Aldenderfer, Blashfield, 1984).

Chicago, Atlanta and Dallas are super hubs. They are the

top ranked for accessibility in all of the matrices that were










TABLE 4.7
ACCESSIBILITY INDICES OF INDIVIDUALLY


WEIGHTED NODES


airports weight accessibility rank
number


ATL
BWI
CLT
CHI
CVG
CLE
DFW
DAY
DEN
DTW
HNL
HOU
LAS
LAX
MEM
MIA
MSP
BNA
NYC*
MCO
PHL
PHX
PIT
RDU
STL
SLC
SFO
SEA
DCA


.054
.034
.048
.060
.036
.023
.052
.018
.043
.036
.011
.031
.019
.032
.034
.019
.041
.036
.040
.024
.037
.028
.050
.028
.042
.028
.031
.024
.043


22.76
5.53
15.13
31.24
7.57
2.24
19.87
1.05
10.23
6.96
0.25
4.69
1.39
4.89
5.18
1.34
8.70
6.84
9.45
2.50
7.18
3.44
16.49
3.02
10.78
2.72
4.36
1.94
11.44


*Includes Newark









TABLE 4.8
HUB STRENGTH CLASSIFICATION SCHEME


SUPER HUBS

MAJOR HUBS--TYPE A


MAJOR HUBS--TYPE B

MODERATE HUBS--TYPE A

MODERATE HUBS--TYPE B



MINOR HUBS--TYPE A

MINOR HUBS--TYPE B


NON-HUBS


Atlanta, Chicago, Dallas

Charlotte, Denver, Pittsburgh,
St. Louis

New York, Washington, D.C.

Cincinnati, Memphis

Baltimore, Detroit, Houston,
Los Angeles, Minneapolis,
Nashville, Philadelphia

Raleigh/Durham, Salt Lake City

Cleveland, Orlando, Phoenix,
San Francisco

Dayton, Honolulu, Las Vegas,
Miami, Seattle


Note: Type A hubs have a transfer (non-local) passenger
percentage of 50% or greater, while the non-local percentage
for type B hubs is less than 50% (table 4.1).







64
constructed, and are indeed in a class by themselves. Because

they are large in population, more centrally located in the

nation than many of the other U.S. mega cities, and each an

important transfer point for more than one carrier for several

years (until early 1991, the now defunct Eastern Airlines had

hub operations at Atlanta's Hartsfield), these cities have

been the major pivot centers for air transportation for more

than a decade. Each has a non-local (transfer) traffic base

of over 50% of its total enplanements. They have a clear

advantage over peripheral hubs like Miami, New York and Los

Angeles in that they are more proximal to a greater number

ofnodes in all parts of the country. These hubs are also

operated by one or more of the financially strongest carriers

in the nation (American, Delta, United) helping to make

possible more flights to more destinations.

Maior hubs--type A are cities with high accessibility

numbers and transfer passenger percentages of over 50%

(excluding the super hubs). Charlotte, Pittsburgh, St. Louis

and Denver make up this category. While the major hubs--type

B have high accessibility numbers as well, their non-local

traffic base is less than 50% of the total traffic at their

facility. For these hubs (New York and Washington, D.C.),

this is more a reflection of the city in question's

popularity as a final destination (due to sheer population

size), and the fact that both are multiple airport cities,

than a reflection on the service offered at the airports.







65

The moderate hubs have medium-ranging accessibility

numbers and include almost one-third of the hubs from table

4.1. Only Memphis and Cincinnati have non-local passenger

percentages of over 50% (type A), although some of the tyDe B

moderate hubs are near that proportion.

Minor hubs--type A and minor hubs--type B have low

accessibility numbers. Again, tvye A hubs have non-local

passenger percentages of over 50% (Salt Lake City and

Raleigh/Durham, for example), while type B hubs (like San

Francisco and Orlando) do not.

Miami, Honolulu, Seattle, Las Vegas and Dayton make up

the non-hub category. These are cities (largely destination

or international gateway hubs) with very low accessibility

numbers and transfer percentages below (in most cases well

below) 50%. Dayton, a USAir hub, is being pulled down from

that status in early 1992. These cities are in no way

comparable to the higher domestic connectivity and major

transfer role played by the other hubs from Table 4.1.



Conclusion


This chapter brings to light possible confusion in the

usage of the term hub, from the FAA definition (which has

nothing whatsoever to do with transfer or connecting status)

to the various service levels of the airline definition. The

hub-and-spoke phenomenon of the post-regulation period has,







66

without a doubt, drastically changed air transportion in the

United States. The accessibility matrix (T) gives us a good

indication of which hubs are the most highly connected to more

of the nation, and therefore, which are more powerful in

controlling market shares. This might be used to help

understand why some carriers are more financially successful

than others in an industry that has been very unstable for

more than a decade. The next chapter will investigate the

relationship between changes in professional employment growth

rates and changes in air transportation connectivity in the 60

largest MSAs in the United States.














CHAPTER 5
CHANGES IN CONNECTIVITY AND
PROFESSIONAL EMPLOYMENT LOCATION



Introduction


This chapter investigates a number of research questions

concerning changes in air service connectivity, professional

employment and corporate location. Data were collected on the

60 largest metropolitan areas in the United States in 1988,

listed in Table 5.1. The MSAs (metropolitan statistical

areas) include a mix of hubs and nonhubs from every region of

the country, although there is a bias toward the east. The

total number of employees, number of administrative and

research and development employees, airport enplanements and

nonstop destinations available (of the 60-city study set) were

collected for these cities in each year from 1978 through

1988. The starting point of the analysis is 1978 since that

was the last fully regulated and fairly stable year for the

airline industry. The analysis continues through 1988 to show

progressive changes and help identify lag effects in the

relationship between changes in connectivity and professional

employment.









TABLE 5.1
60 LARGEST METROPOLITAN AREAS OF THE U.S.
(1988-in thousands)


New York/Newark CMSA (NYC) 18,120
Los Angeles CMSA (LAX) 13,770
Chicago CMSA (CHI) 8,181
San Francisco CMSA (SFO) 6,042
Philadelphia CMSA (PHL) 5,963
Detroit CMSA (DTW) 4,352
Boston CMSA (BOS) 4,110
Dallas/Ft. Worth CMSA (DFW) 3,766
Washington, D.C. MSA (WAS) 3,734
Houston CMSA (HOU) 3,641
Miami/Ft. Lauderdale CMSA (MIA) 3,001
Cleveland CMSA (CLE) 2,769
Atlanta MSA (ATL) 2,737
St. Louis MSA (STL) 2,467
Seattle CMSA (SEA) 2,421
Minneapolis/St. Paul CMSA (MSP) 2,388
Baltimore MSA (BWI) 2,342
San Diego MSA (SAN) 2,370
Pittsburgh CMSA (PIT) 2,284
Phoenix MSA (PHX) 2,030
Tampa/St. Petersburg MSA (TPA) 1,995
Denver CMSA (DEN) 1,858
Milwaukee CMSA (MKE) 1,562
Kansas City MSA (MCI) 1,575
Cincinnati CMSA (CVG) 1,449
Portland CMSA (PDX) 1,414
Sacramento MSA (SMF) 1,385
Norfolk MSA (ORF) 1,380
Columbus MSA (CMS) 1,344
San Antonio MSA (SAT) 1,323
New Orleans MSA (MSY) 1,307
Indianapolis MSA (IND) 1,237
Buffalo CMSA (BUF) 1,176
Providence CMSA (PVD) 1,125
Charlotte MSA (CLT) 1,112
Hartford CMSA (BDL) 1,068
Salt Lake City MSA (SLC) 1,065
Rochester MSA (ROC) 980
Memphis MSA (MEM) 979
Nashville MSA (BNA) 972
Orlando MSA (MCO) 971
Louisville MSA (SDF) 967
Oklahoma City MSA (OKC) 964
Dayton MSA (DAY) 948
Greensboro MSA (GSO) 925
Birmingham MSA (BHM) 923









TABLE 5.1--continued



Jacksonville MSA (JAX) 898
Albany MSA (ALB) 851
Richmond MSA (RIC) 844
Honolulu MSA (HNL) 838
West Palm Beach MSA (PBI) 818
Austin MSA (AUS) 748
Scranton MSA (AVP) 731
Tulsa MSA (TUL) 728
Raleigh/Durham MSA (RDU) 683
Allentown MSA (ABE) 677
Grand Rapids MSA (GRR) 665
Syracuse MSA (SYR) 650
Tucson MSA (TUS) 636
Las Vegas MSA (LAS) 631



Note: The three-letter code listed after each city is the
airport code for the city as assigned by the FAA. Some cities
on the list are served by more than one airport. In these
cases, one code is chosen to represent the urban area as a
whole, but includes total destinations and flows from all
airports in the metropolitan area.

Source: Statistical Abstract of the United States, 1990.









Research Questions



This chapter investigates how changes in connectivity are

related to changes in employment growth of those highly active

in nonroutine communication within and outside of the firm.

It is assumed that hub cities have a much greater selection of

nonstop destinations from their airports plus a higher rate of

connectivity change as the hub develops, and as such, have a

greater (and growing) concentration of scientists, engineers

and administrators in their workforce than nonhubs.

The study will also determine which change occurs first.

The potential demand for air service could attract one or more

airlines to set up hub operations with abundant nonstop

service in the city, thus making it an attractive choice for

companies to locate or relocate such activities.

Alternatively, the demand could have already existed, with an

airline merely stepping in to fill the void in service. The

changes could also occur simultaneously as new economic growth

regions are identified by the airline industry and other

corporations.

If an airline's choice of a hub does make that city

likely to draw more company headquarters, regional offices and

research and development labs and their workers to the city

(or induce such facilities to leave nonhub cities), the study

will determine how long it takes for the flow to begin. If

the effect is in the other direction, the study will determine









how long it takes the airline to respond (with increased

service levels) to the potential market growth brought on by

the new addition to the urban area's employment structure.

Finally, the study will look for regional and

hierarchical biases. Hubs in some areas may show a stronger

relationship between the two variables than hubs in other

parts of the nation, and hubs with higher population totals

could show a stronger or weaker relationship between

connectivity change and employment change than smaller-sized

hubs.



The Labor Data


The professional labor data were collected from the

County Business Patterns series published by the U.S. Census

Bureau. It is an annual publication with volumes for all 50

states, the District of Columbia and Puerto Rico.

Unfortunately, this labor information is not published on a

consistent basis for metropolitan areas as a whole. The

series records the number of employees per economic sector

(based on SIC codes) by county in the states. Within each

sector, a separate classification is reserved for professional

workers. The publication refers to these workers as

administrative and auxiliary employees, defined as personnel

working in central administration offices and auxiliary

establishments such as research labs and financial services.







72

While the category also includes some warehouse and

distribution employees (nonprofessional labor), County

Business Patterns was found to be the best source available

for obtaining raw numbers of professional workers in a

specific area.

The number of professional workers for each of the 60

cities in the study set was obtained by summing the

administrative and auxiliary employees for each sector in each

county included in the MSA (as determined by the U.S. Census

Bureau) (Appendix B). These data were collected for all 60

cities from 1978 (the beginning of the study period) through

1988 (the most current data available at the time of the

collection process). The results are given in Appendix C.

Table 5.2 shows the 20 MSA's (from the study set) with

the greatest number of professionals in their workforce at the

end of that period (1988). It is not surprising that the

largest cities in the nation in population rank high on this

list. Most of these 20 MSA's lie in the traditional

manufacturing belt, although a few southern and western urban

areas made the list as well.

Table 5.3 indicates the flow pattern of administrative

and auxiliary jobs and workers. This table shows the U.S.

cities that experienced the greatest increase in the number of

professional workers from 1978 to 1988. While a few of the

large metropolitan areas of the Northeast made this list (such

as Boston, New York/Newark and Philadelphia), the southern and









TABLE 5.2
MOST ADMINISTRATIVE/AUXILIARY EMPLOYEES: 1988


MSA number of population
professionals rank



1) New York/Newark 462,305 1
2) Los Angeles 193,284 2
3) Chicago 183,688 3
4) Boston 139,258 7
5) Detroit 136,782 6
6) Philadelphia 136,055 5
7) San Francisco 120,469 4
8) Dallas/Ft. Worth 109,966 8
9) Houston 95,362 10
10) Minneapolis/St. Paul 83,141 16
11) Atlanta 81,958 13
12) Washington, D.C. 64,363 9
13) Cleveland 63,586 12
14) St. Louis 58,350 14
15) Providence 55,189 34
16) Seattle 43,944 15
17) Cincinnati 40,221 23
18) Pittsburgh 39,973 19
19) Miami/Ft. Lauderdale 36,602 11
20) Columbus 33,304 29


Source: Calculated from County Busi
Census Bureau, 1990.


Patterns, U.S.


nw=. ~










TABLE 5.3
LARGEST GROWTH IN ADMINISTRATIVE/AUXILIARY EMPLOYEES:
1978-1988




MSA + change



1) Boston 68,528
2) New York/Newark 55,239
3) Dallas/Ft. Worth 49,976
4) Atlanta 39,931
5) San Francisco 31,222
6) Washington, D.C. 28,310
7) Providence 28,292
8) Philadelphia 22,989
9) Minneapolis/St. Paul 21,459
10) Houston 21,236
11) Seattle 18,496
12) Miami/Ft. Lauderdale 17,697
13) Los Angeles 16,575
14) Charlotte 13,967
15) Orlando 13,965
16) Jacksonville 12,273
17) San Diego 10,488
18) Richmond 10,315
19) Columbus 9,843
20) Phoenix 9,821


Patterns, U.S.


Source: Calculated from County Business
Census Bureau, 1980, 1990.


~







75

western parts of the United States are strongly represented,

including cities such as Charlotte, Richmond, Jacksonville,

Orlando and Phoenix which did not make the list in Table 5.2.

Some of the cities in Table 5.2 experienced a loss of

administrative and auxiliary workers during the period in

question. These are listed in Table 5.4. Almost all of these

cities are in the traditional manufacturing belt. Tulsa and

Oklahoma City, however, are oil-producing cities whose

fortunes changed during the early and middle 1980s as the

price for oil dropped significantly, and the economy of oil

states (particularly Oklahoma and Texas) suffered.

Table 5.5 gives a regional analysis for the 60

metropolitan areas in the study set. It shows the total and

average administrative and auxiliary employment growth of the

study set cities within each region. These regions are the

traditional divisions of the nation as defined by the U.S.

Census Bureau (Figure 5.1). The Northeast still shows up

strongly with the South being a near second.

Perhaps of greater interest to a discussion of

professional workers in urban locations is Table 5.6. It

shows the 20 cities with the highest percentage of

administrative and auxiliary workers as a proportion of their

total employment for 1988 (constructed from Appendices C and

D). Many of the larger cities on the list in Table 5.2 either

drop off this list or appear at a lower ranking on the list.

Some new cities appear that were absent from Table 5.2, such









TABLE 5.4
ADMINISTRATIVE/AUXILIARY EMPLOYMENT LOSS: 1978-1988



MSA change



Detroit -27,340
Pittsburgh -12,179
Albany -4,094
Indianapolis -2,216
Baltimore -1,999
Tulsa -1,674
Oklahoma City -1,258
Chicago -592
Buffalo -367


Source: Calculated from County Business Patterns, U.S.
Census Bureau 1980, 1990.





TABLE 5.5
ADMINISTRATIVE/AUXILIARY GROWTH BY REGION: 1978-1988



region number of employee MSA
MSA's growth average



Northeast 12 170,993 14,249

North Central 12 36,919 3,077

South 24 243,392 10,141

West 12 100,311 8,359



Source: Calculated from County Business Patterns, U.S.
Census Bureau, 1980, 1990.









TABLE 5.6
HIGHEST PERCENTAGE OF ADMINISTRATIVE/AUXILIARY AS A PORTION
OF TOTAL EMPLOYMENT: 1988



MSA percentage of
professionals


1) Tulsa 8.76
2) Detroit 7.75
3) Allentown/Bethlehem/Easton 7.38
4) Minneapolis/St. Paul 7.16
5) Greensboro/Winston/Salem 7.14
6) Houston 7.08
7) Dallas/Ft. Worth 6.58
8) Richmond 6.53
9) Atlanta 6.44
10) Rochester 6.07
11) Columbus 5.96
12) New York/Newark 5.86
13) St. Louis 5.83
14) Cleveland 5.74
15) Dayton 5.73
16) Philadelphia 5.72
17) Cincinnati 5.58
18) Chicago 5.43
19) Boston 5.20
20) Charlotte 5.18


Source: Calculated from County Business Patterns, U.S.
Census Bureau, 1990.































































FIGURE 5.1
CENSUS REGIONS OF THE UNITED STATES









as the Greensboro/Winston-Salem and Tulsa metropolitan areas.

Because it readily identifies cities with strong

administrative and auxiliary functions, using the percentage

of professional employment in the total labor force isprobably

a better indicator of the true administrative and auxiliary

cities in the United States.


The Air Service Data


The sources for the air service data were the Official

Airline Guide (OAG) and Airport Activity Statistics of

Certified Route Air Carriers. Enplanement figures (Appendix

E) were extracted from the latter source (published yearly by

the Federal Aviation Administration) which records a variety

of information about passenger and cargo flow at all domestic

airports with scheduled commercial service. The Official

Airline Guide is published monthly and gives route schedules

of the commercial carriers in the United States. It was the

source for nonstop (Appendix F) and other city-pair connection

data. Unfortunately, this publication is not widely available

because of its cost and the sheer size of each issue (creating

storage problems). Back issues of the publication were found

at a library (University of North Carolina at Chapel Hill)

which saves only one issue from every year (usually, but not

always, July). Therefore, connection and schedule data for

the various years are from summer schedules of the major







80

carriers. It should be noted that summer schedules are

usually expanded to accommodate extensive vacation travel, and

are not representative of the average service offered during

the year as a whole.

Data were collected on all 60 cities in the study set for

each year from 1978 through 1988. More current air data were

available, but the collection process was stopped at 1988 to

keep the air data consistent with the available labor data (as

discussed in the previous section of this chapter). In

communities with multiple airports (e.g. Dallas and Chicago),

the individual airport information was combined to obtain a

community total. Multiple airport communities are listed in

Table 5.7.

The U.S. cities with the highest enplanement (boarded

passenger) levels in 1988 are given in Table 5.8 by rank order

along with their average number of daily departures and the

number of cities in the study set to which they are connected

with nonstop air service (59 is the maximum number of

nonstops). The list consists of the largest cities in the

nation, plus a few popular travel destinations such as

Honolulu, Las Vegas and Orlando. It is not surprising that

almost all (except Boston) are airline hub cities (Table 4.1).

Every city (of the 60 in the study set) grew in enplanement

levels from 1978 to 1988 except Louisville (which dropped by

38,163 passengers) and Scranton/Wilkes/Barre (a decline of

16,240) (Appendix E).









TABLE 5.7
MULTIPLE AIRPORT CITIES


Chicago

Dallas

Detroit

Greensboro

Houston

Los Angeles


Miami

New York City

San Francisco

Washington, D.C.


Midway, O'Hare International

Love Field, DFW International

Detroit City, Detroit Metro-Wayne County

Greensboro, Smith-Reynolds

Hobby, Houston Intercontinental

L.A. International, Hollywood/Burbank,
Orange County, Long Beach

Miami International, Ft. Lauderdale

JFK, LaGuardia, Newark International

San Francisco International, Oakland

Washington National, Dulles International









TABLE 5.8
LEADING ENPLANEMENT CITIES OF THE U.S.: 1988


city enplanements non/stops* daily
departures



1) New York/Newark 32,820,184 48 957
2) Chicago 29,770,857 56 1045
3) Dallas 23,488,986 50 765
4) Los Angeles 22,836,344 37 743
5) Atlanta 21,824,125 56 756
6) San Francisco 15,173,602 28 542
7) Denver 14,441,817 40 514
8) Miami 13,360,799 27 399
9) Washington 11,586,627 49 482
10) Houston 10,712,269 38 432
11) Boston 10,141,298 40 316
12) St. Louis 9,554,454 46 384
13) Phoenix 9,455,324 31 364
14) Detroit 9,343,770 43 352
15) Honolulu 8,396,313 11 230
16) Pittsburgh 8,378,639 49 345
17) Minneapolis 8,170,952 40 301
18) Orlando 7,473,086 38 259
19) Las Vegas 6,864,803 28 220
20) Seattle 6,825,513 22 296


Sources: Official Airline Guide,
Activity Statistics of
Carriers, 1988.


July 1988 and Airport
Certified Route Air


* number of different non-stop connections of the 60 in the
study set


Certified Routg Air







83

The average number of daily departures (Table 5.8) were

calculated by dividing the number of commercial departures

scheduled by the airlines during the calendar year 1988 (as

recorded by the Federal Aviation Adminstration) by 365.

Again, most of the metropolitan areas in the study set saw an

increase in the number of daily airplane departures from 1978

to 1988. Los Angeles, Chicago, San Francisco, Houston,

Phoenix and Charlotte all saw increases of over 200 daily

departures. Eight cities, however, experienced declines.

Milwaukee, Grand Rapids, Albany, Scranton/Wilkes/Barre and

Birmingham all showed decreases of less than 3 daily

departures, while Louisville lost an average of 8 departures,

New Orleans an average of 13 and Buffalo an average of 16

daily departures from 1978 to 1988.

Some cities with a higher number of nonstop connections

than a few of those listed in Table 5.8 are Cincinnati (43),

Charlotte (40), Philadelphia (43), Cleveland (37) and

Baltimore (38). They are all hub cities, and as such, are

highly connected. Twelve cities in the study set experienced

significant increases (10 or more) in the number of nonstop

connections available from their airports between 1978 and

1988 (Table 5.9). They are largely post-deregulation hubs

(Table 4.2). Fifteen cities, most of them spoke or nonhub

cities from the study set, however, experienced reductions in

the number of nonstop connections from their facilities. The









TABLE 5.9
GREATEST INCREASES IN NONSTOP CONNECTIONS
1978-1988



city increase



1) Orlando 20

2) Charlotte 18

3) Cincinnati 17

4) Minneapolis 16

5) Raleigh/Durham 16

6) Nashville 15

7) Memphis 14

8) Phoenix 12

9) Pittsburgh 12

10) Salt Lake City 11

11) Atlanta 10

12) Baltimore 10



Source: Official Airline Guide, July, 1978 and 1988

Note: of the 60 cities in the study set







85
most severe decline was felt by Buffalo (7), a city whose

economy suffered during the period.



Research Methodology



The analysis in this chapter investigates whether or not

changes in connectivity are related to changes in

professionalemployment growth or vice versa. To measure the

impact of hub selection on this growth, a general hub/nonhub

comparison was carried out using all 60 cities in the study

set over the 11 year period.

Changes in connectivity were monitored using a derived

accessibility index. The yearly connectivity indices for each

metropolitan area were compared to changes in professional

employment by looking at the average rate of change of each

variable from one year to the next. Graphing both rate-of-

change values for each hub city over time should give an

indication of the relationship between the two variables. The

means of the yearly average rates of change for connectivity

and professional employment were calculated for each MSA, and

a simple correlation analysis was done to determine the

statistical relationship between the two. In addition, by

separating the data for the MSA's into different groups,

regional and hierarchical information was obtained.

To test for a lag structure in the relationship between

the variables, times series regression analysis was performed







86

using the average rates of change for each variable for all

years in the study period.



Accessibility Indices



To obtain accessibility measurements to be used in the

analysis, binary matrices were constructed for each of the 11

years in the study period (1978-1988). Each 60 X 60 matrix

had cell values of either one or zero. A value of one was

assigned to the cell if a direct air service connection

existed between the particular cities in question. If such a

connection did not exist, then the cell was assigned a value

of zero.

As discussed in Chapter 4, this simple connectivity

matrix (Cl), can be used to measure accessibility from a graph

theoretic approach. The row values are summed to give the

nodality index or gross vertex connectivity number for each

node. The larger the nodality index, the greater the relative

connectivity. Higher order connections can be taken into

account by powering the original connectivity matrix (C1)

until all of the zero elements disappear from the resultant

matrix. Summing the matrix Cl with the powered matrices

(yielding matrix T) allows one to measure total accessibility

within the network.

Each matrix was multiplied to the third power. At this

point, all zero elements disappeared from the resultant







87
matrices. Thus, the diameter of each network (from 1978

through 1988) is three. Appendix G gives the yearly

accessibility indices for each metropolitan area from matrices

Cl, C2, C3 and T from 1978 through 1988. It is particularly

important in this study to include indirect connectivity

effects since business travelers often visit multiple

destinations in one trip without necessarily going back

through the same hub city between flight segments. The cities

(airports) with the highest accessibility indices in 1988 are

ranked in Table 5.10.



Employment-Connectivity Relationships


Figure 5.2 compares the average rate of change from one

year to the next (from 1978 through 1988) in the number of

professional workers (administrative and auxiliary employees)

in the labor force in each of the 60 metropolitan areas to the

average rate of change in air service connectivity. The rates

were calculated from the raw numbers of administrative and

auxiliary employees and connectivity indices given in

appendices D and H, respectively. A quick glance reveals that

the two graphs have the same general trend in the first half

of the study period, but no apparent similarity in the latter

years. If the two variables are related to (affected by) one

another, a change in one seems to bring a near instantaneous

change in the other during the early years of the period, with









TABLE 5.10
STUDY SET CITIES WITH HIGHEST ACCESSIBILITY INDICES: 1988



city index
number



1) Chicago 53,747
2) Atlanta 53,386
3) Dallas 50,158
4) Washington, D.C. 49,999
5) New York/Newark 49,887
6) St. Louis 49,839
7) Pittsburgh 49,297
8) Cincinnati 46,562
9) Detroit 46,193
10) Charlotte 45,955
11) Philadelphia 45,403
12) Minneapolis 43,304
13) Boston 43,239
14) Denver 42,651
15) Cleveland 42,371
16) Houston 41,649
17) Orlando 41,623
18) Los Angeles 41,250
19) Baltimore 41,109
20) Memphis 40,628


Source: Appendix G












































Years


FIGURE 5.2
CHANGE IN PROFESSIONAL EMPLOYMENT VS. CONNECTIVITY







90

the employment change occurring slightly ahead of the

connectivity change. In the latter years, either there is no

relationship between the two or possibly a complex lag

structure exists.

The data were divided between hubs and nonhubs to compare

the different rates of change for both air connectivity and

professional employment for each group. Post-deregulation hub

cities were considered part of the nonhub group until the year

that hub status was achieved (Table 4.2). For example,

Baltimore was classified as a nonhub until 1983, but was

switched to the hub category after 1983.

The nonhub connectivity rate of change (Figure 5.3) stays

very close to the rate of change for professional employment

(practically the same graph), while the hub city connectivity

rate (Figure 5.4) stays below the employment rate of change

for much of the study period (until the latter years).

Indeed, the two lines in Figure 5.4 appear to have little in

common. The graphs (Figures 5.3 and 5.4) seem to suggest that

a stronger and virtually instantaneous relationship exists

between connectivity change and professional employment change

for nonhubs than for hub cities.

Figure 5.5 breaks the employment data between hubs and

nonhubs. Throughout a fair portion of the study period,

professional employment at nonhubs grew by a higher rate than

hub cities. The mean of the yearly average rates of

professional employment change during the study period for




Full Text
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID EORIWOQZ6_FRI0OE INGEST_TIME 2013-04-08T20:10:43Z PACKAGE AA00013582_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES


xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID EZCPB2IP2_GZHALS INGEST_TIME 2013-03-12T13:29:46Z PACKAGE AA00013582_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES