COLUMBIA UNIVERSITY 2010

Boxed In: How Intermodalism Enabled Destructive Interport Competition

Cuz Potter

Professor Peter Marcuse

Submitted in partial fulfillment of theRequirements for the degree

of Doctor of Philosophyin the Graduate School of Arts and Sciences

COLUMBIA UNIVERSITY

2010

©2010

Cuz Potter

All Rights Reserved

Abstract

Boxed In: How Intermodalism Enabled Destructive Interport Competition

Cuz Potter

What is the appropriate scale for port governance in North America? By

standardizing freight technology, containerization has transformed freight trans-

portation from a segmented, mode-specific, and regional system into a seamless,

intermodal, and global system. The drive to provide global reach, deregula-

tion, and increasing capital costs have concentrated carrier ownership in ocean

shipping. These carriers have subsequently expanded the effective area of pro-

duction for freight delivery by adopting continental strategies for freight move-

ments. This scaling up of organization has permitted carriers to overcome ports’

historical geographical monopoly over their hinterlands and initiate competitive

bidding to host the carriers’ terminal operations. This dissertation examines

how this shift led to a particular competition between the ports of New York-

New Jersey, Baltimore, and Halifax to host a Maersk-Sea Land terminal in the

late 1990s. The case demonstrates that competition results in an unnecessary

and unrewarded transfer of wealth from local taxpayers and users to global

firms. Having illustrated the destructive nature of interport competition, the

paper turns back to containerization’s origins in the first half of the twentieth

century to identify strategies for countering global carriers’ power. The long-

shoremen’s successful struggle against exploitative employers to secure steady

and reasonable wages documents the importance of establishing a countervailing

geographical monopoly over the effective area of production by scaling organi-

zation up from individual ports to entire coasts. The paper thus concludes that

port governance must be scaled up to the national or continental scale to more

efficiently and equitably coordinate freight transportation in North America.

Contents

Contents i

List of Tables vii

List of Figures x

Acknowledgments xiii

1 Introduction 1

1.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

1.2 Technological change . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

1.3 Deregulation and concentration . . . . . . . . . . . . . . . . . . . . . 5

1.4 Port authorities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

1.5 Stakes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

1.6 The question . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12

1.7 Theoretical approach: Technology, territory, and terrain . . . . . . . . 13

1.7.1 Terrain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

1.7.2 Territory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16

1.7.3 Technology . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

1.8 Structure of dissertation . . . . . . . . . . . . . . . . . . . . . . . . . 19

i

2 Methodology 21

2.1 Quantitative Analyses . . . . . . . . . . . . . . . . . . . . . . . . . . 21

2.1.1 Computing and software . . . . . . . . . . . . . . . . . . . . . 22

2.1.2 Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22

2.1.3 Mapping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

2.1.4 Regressions . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24

2.1.5 Notes for interpreting plots and regressions . . . . . . . . . . . 26

2.2 Historical analyses . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27

2.3 Reproducibility and literate programming . . . . . . . . . . . . . . . 28

3 Technological and organizational change 31

3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31

3.2 Stage One: Invention, development, and innovation . . . . . . . . . . 37

3.2.1 Nineteenth century intermodalism . . . . . . . . . . . . . . . . 37

3.2.2 Early twentieth century . . . . . . . . . . . . . . . . . . . . . 39

3.2.3 Mid-century . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43

3.2.4 Early Containerization . . . . . . . . . . . . . . . . . . . . . . 49

3.3 Stage Two: Transfer . . . . . . . . . . . . . . . . . . . . . . . . . . . 51

3.3.1 Domestic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51

3.3.2 International . . . . . . . . . . . . . . . . . . . . . . . . . . . 54

3.4 Stage Three: Growth, competition, and consolidation . . . . . . . . . 56

3.4.1 Growth: The homogenization of space . . . . . . . . . . . . . 56

3.4.2 Competition: Deregulation and overcapacity . . . . . . . . . . 58

3.4.3 Consolidation: Mergers, acquisitions, and alliances . . . . . . . 61

3.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66

4 Spatial change 68

ii

4.1 Ports . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69

4.2 Rail . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73

4.3 Warehousing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74

4.3.1 Overall increase in warehousing employment . . . . . . . . . . 84

4.3.2 Deconcentration . . . . . . . . . . . . . . . . . . . . . . . . . . 84

4.3.3 Reconcentration . . . . . . . . . . . . . . . . . . . . . . . . . . 86

4.3.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87

4.4 Freight trucking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88

4.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90

5 Global capital and interport competition 93

5.1 Introduction: Territory and terrain . . . . . . . . . . . . . . . . . . . 93

5.2 Port authorities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96

5.2.1 Characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . 96

5.2.2 U.S. port authorities . . . . . . . . . . . . . . . . . . . . . . . 102

5.2.3 Canadian port authorities . . . . . . . . . . . . . . . . . . . . 102

5.3 Three ports . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103

5.3.1 Baltimore . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104

5.3.2 Halifax . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107

5.3.3 New York and New Jersey . . . . . . . . . . . . . . . . . . . . 109

5.4 Competition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115

5.4.1 Background: Easy marks . . . . . . . . . . . . . . . . . . . . . 115

5.4.2 The drudgery of dredging . . . . . . . . . . . . . . . . . . . . 117

5.4.3 The request for proposals . . . . . . . . . . . . . . . . . . . . 118

5.4.4 What was at risk? . . . . . . . . . . . . . . . . . . . . . . . . 121

5.4.5 Boxed in . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122

iii

5.4.6 Shunted away . . . . . . . . . . . . . . . . . . . . . . . . . . . 127

5.5 Territorial squabbles . . . . . . . . . . . . . . . . . . . . . . . . . . . 129

5.6 Costs of competition . . . . . . . . . . . . . . . . . . . . . . . . . . . 131

5.6.1 Port costs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131

5.6.2 Labor costs . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133

5.7 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135

6 Direct Employment 138

6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138

6.2 Economic impact analyses . . . . . . . . . . . . . . . . . . . . . . . . 139

6.3 Port operations and marine cargo handling . . . . . . . . . . . . . . . 143

6.4 Deep sea freight transportation . . . . . . . . . . . . . . . . . . . . . 146

6.5 Warehousing and storage . . . . . . . . . . . . . . . . . . . . . . . . . 149

6.6 Freight trucking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152

6.7 Freight transportation arrangement . . . . . . . . . . . . . . . . . . . 153

6.8 Good jobs or bad? . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157

6.9 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158

7 Decoupling 161

7.1 Port-related employment through localization economies . . . . . . . 161

7.1.1 Industry types and transportation . . . . . . . . . . . . . . . . 162

7.1.2 Dialectics of transportation . . . . . . . . . . . . . . . . . . . 164

7.2 Sector analyses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168

7.2.1 Natural resource extraction . . . . . . . . . . . . . . . . . . . 169

7.2.2 Heavy industry . . . . . . . . . . . . . . . . . . . . . . . . . . 171

7.2.3 Manufacturing . . . . . . . . . . . . . . . . . . . . . . . . . . 173

7.2.4 Retail . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179

iv

7.2.5 Information work . . . . . . . . . . . . . . . . . . . . . . . . . 182

7.3 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185

8 Territorial Monopoly 188

8.1 Introduction: Rationalizing congestion . . . . . . . . . . . . . . . . . 189

8.2 Before mechanization . . . . . . . . . . . . . . . . . . . . . . . . . . . 192

8.2.1 Meeting the hook . . . . . . . . . . . . . . . . . . . . . . . . . 193

8.2.2 Congestion and casualization . . . . . . . . . . . . . . . . . . 195

8.2.3 Casualization and organization . . . . . . . . . . . . . . . . . 201

8.3 Early mechanization . . . . . . . . . . . . . . . . . . . . . . . . . . . 206

8.3.1 Fear of mechanization . . . . . . . . . . . . . . . . . . . . . . 206

8.3.2 Decasualization and geographical expansion: The 1934 and 1936

West Coast strikes . . . . . . . . . . . . . . . . . . . . . . . . 208

8.4 Mid-century: Spheres of dominance affirmed . . . . . . . . . . . . . . 210

8.5 Mechanization and modernization . . . . . . . . . . . . . . . . . . . . 212

8.5.1 The Setting: West Coast weakness . . . . . . . . . . . . . . . 213

8.5.2 East Coast . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 218

8.6 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 225

9 Conclusion 229

9.1 Summary of findings . . . . . . . . . . . . . . . . . . . . . . . . . . . 229

9.2 Ongoing relevance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231

9.3 Policy recommendations . . . . . . . . . . . . . . . . . . . . . . . . . 232

9.4 Planning implications . . . . . . . . . . . . . . . . . . . . . . . . . . . 239

9.5 Concluding thought . . . . . . . . . . . . . . . . . . . . . . . . . . . . 240

A Total employment estimation 242

v

B NAICS and SIC code correspondence for selected industries 245

C Industrial sector regression analyses 252

D Additional industrial sector regression analyses 270

Bibliography 280

List of Interviews 308

vi

List of Tables

2.1 Regression variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

3.1 Percent of palletized cargo . . . . . . . . . . . . . . . . . . . . . . . . . . 45

4.1 Total North American container traffic by port range (1,000 TEUs) . . . 70

4.2 U.S. employment by industry as total and percent of U.S. total employment 92

5.1 Port Authority facility income (loss) in millions . . . . . . . . . . . . . . 114

6.1 Employment in port operations and marine cargo handling by state per

1,000 TEUs moved through ports in that state . . . . . . . . . . . . . . . 145

6.2 Employment in deep sea freight transportation by state per 1,000 TEUs

moved through ports in that state . . . . . . . . . . . . . . . . . . . . . . 148

6.3 Employment in general warehousing and storage by state per 1,000 TEUs

moved through ports in that state . . . . . . . . . . . . . . . . . . . . . . 151

6.4 Employment in general freight trucking by state per 1,000 TEUs moved

through ports in that state . . . . . . . . . . . . . . . . . . . . . . . . . . 153

6.5 Employment in freight transportation arrangement by state per 1,000

TEUs moved through ports in that state . . . . . . . . . . . . . . . . . . 155

6.6 Estimated ratio of jobs per 1,000 TEUs of container traffic by sector . . . 159

vii

7.1 Relationship of industrial sectors to transportation following Marshall and

Cooley . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167

7.2 Selected industries ranked by value of exports in 2000 . . . . . . . . . . . 170

7.3 Relationship of industrial sectors to transportation following Marshall and

Cooley . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 186

B.1 Port-related industries with SIC and NAICS codes and bridges indicated 246

B.2 Selected industries with SIC and NAICS codes and bridges indicated . . 250

C.1 General Warehousing and Storage (SIC 4225 and NAICS 493110 and

531130) (Log of employment) . . . . . . . . . . . . . . . . . . . . . . . . 253

C.2 General freight trucking (SIC 4210 and NAICS 484100 and 484200) (Log

of employment) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 254

C.3 Freight transportation arrangement (SIC 4710 and 4723 and NAICS 488510)

(Log of employment) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 255

C.4 Mining, iron ore (SIC 1010 and NAICS 212210) (Log of employment) . . 256

C.5 Petroleum Refineries (SIC 2911 and NAICS 324110) (Log of employment) 257

C.6 Glass container manufacturing (SIC 3221 and NAICS 327213) (Log of

employment) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 258

C.7 Book, periodical, and music stores (SIC 5942, 5994, and 5733/5735 and

NAICS 451211, 451212, and 451220) (Log of employment) . . . . . . . . 259

C.8 Adhesive manufacturing (SIC 2891 and NAICS 325520) (Log of employment)260

C.9 Gum and wood chemicals manufacturing (SIC 2860/2861 and NAICS


C.10 Pharmaceutical preparations (SIC 2834 and NAICS 325412) (Log of em-

ployment) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 262

viii

C.11 Electroplating, plating, polishing, anodizing, and coloring (SIC 3471 and

NAICS 332813) (Log of employment) . . . . . . . . . . . . . . . . . . . . 263

C.12 Turned product and screw, nut, and bolt manufacturing (SIC 3450 and

NAICS 332720) (Log of employment) . . . . . . . . . . . . . . . . . . . . 264

C.13 Machine tool (metal forming types) manufacturing (SIC 3542 and NAICS


C.14 Automatic environmental control manufacturing for residential, commer-

cial, and appliance use (SIC 3822 and NAICS 334512) (Log of employment)266

C.15 Semiconductors and related device manufacturing (SIC 3674 and NAICS


C.16 Periodical publishers (SIC 2720 and NAICS 511120) (Log of employment) 268

C.17 Motion picture and video production (SIC 7813 and 7814/7812 and NAICS


D.1 Deep Sea Freight Transportation (SIC 4410 and NAICS 483111) (Log of

employment) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 272

D.2 Marine cargo handling (SIC 4463 and 4491 and NAICS 488310 and 488320)

(Log of employment) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 273

D.3 Refrigerated warehousing and storage (SIC 4222 and NAICS 493120) (Log

of employment) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 274

D.4 Ready-mix concrete manufacturing (SIC 3273 and NAICS 327320) (Log

of employment) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 277

D.5 Cement, hydraulic (SIC 3241 and NAICS 327310) (Log of employment) . 278

D.6 Flat glass manufacturing (SIC 3211 and NAICS 327211) (Log of employ-

ment) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 279

ix

List of Figures

3.1 Total volume of containers handled by world ports . . . . . . . . . . . . 36

3.2 Total volume of containers handled by U.S. ports . . . . . . . . . . . . . 37

3.3 Containerization growth rate for the United States and the world . . . . 38

4.1 Container volume for major West Coast ports by year . . . . . . . . . . . 71

4.2 Container volume for major East Coast ports by year . . . . . . . . . . . 72

4.3 Container volume for Gulf Coast ports by year . . . . . . . . . . . . . . . 72

4.4 Map of total employment for public general warehousing in 1974 . . . . . 76




4.8 Map of location quotient for public general warehousing in 1974 . . . . . 80




4.12 General Warehousing and Storage . . . . . . . . . . . . . . . . . . . . . . 85

4.13 General freight trucking . . . . . . . . . . . . . . . . . . . . . . . . . . . 89

5.1 Annual container volume in millions of TEUs for three finalists . . . . . . 107

x

6.1 Total employment in marine cargo handling transportion by state . . . . 144

6.2 Total employment in deep sea freight transportion by state . . . . . . . . 147

6.3 Total employment in general warehousing and storage by state . . . . . . 150

6.4 Total employment in general freight trucking by state . . . . . . . . . . . 154

6.5 Total employment in freight transportation arrangement by state . . . . 155

6.6 Freight transportation arrangement . . . . . . . . . . . . . . . . . . . . . 156

7.1 Mining, iron ore . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171

7.2 Petroleum refineries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173

7.3 Glass container manufacturing . . . . . . . . . . . . . . . . . . . . . . . . 173

7.4 Pharmaceutical preparations . . . . . . . . . . . . . . . . . . . . . . . . . 179

7.5 Electroplating, plating, polishing, anodizing, and coloring . . . . . . . . . 179

7.6 Machine tool (metal forming types) manufacturing . . . . . . . . . . . . 180

7.7 Automatic environmental control manufacturing . . . . . . . . . . . . . . 180

7.8 Semiconductors and related device manufacturing . . . . . . . . . . . . . 180

7.9 Adhesive manufacturing . . . . . . . . . . . . . . . . . . . . . . . . . . . 181

7.10 Gum and wood chemicals manufacturing . . . . . . . . . . . . . . . . . . 181

7.11 Turned products and screw, nut and bolt manufacturing . . . . . . . . . 181

7.12 Book, periodical, and music stores . . . . . . . . . . . . . . . . . . . . . . 183

7.13 Periodical publishers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185

7.14 Motion picture and video production . . . . . . . . . . . . . . . . . . . . 185

8.1 Shape-up in New York City . . . . . . . . . . . . . . . . . . . . . . . . . 198

8.2 Weekly earnings for two gangs in 1928 . . . . . . . . . . . . . . . . . . . 202

8.3 Monthly earnings for two gangs in 1928 . . . . . . . . . . . . . . . . . . . 202

D.1 Deep sea freight transportation . . . . . . . . . . . . . . . . . . . . . . . 271

xi

D.2 Marine cargo handling . . . . . . . . . . . . . . . . . . . . . . . . . . . . 271

D.3 Refrigerated warehousing and storage . . . . . . . . . . . . . . . . . . . . 271

D.4 Ready-mix concrete manufacturing . . . . . . . . . . . . . . . . . . . . . 275

D.5 Cement, hydraulic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 275

D.6 Flat glass manufacturing . . . . . . . . . . . . . . . . . . . . . . . . . . . 276

xii

Acknowledgments

Like all dissertations, this one could not have been completed without a great deal of

help and history. I would be remiss if I did not recognize the contributions of some of

the people who aided my journey and apologize in advance to those who I neglect (but

appreciate nonetheless). I would especially like to thank my advisors, Peter Marcuse

and Susan Fainstein. Perhaps no other individual has shaped my thought as much

as Peter has. If there is another person who has, it would have to be Susan, who

also invested significant time in commenting on my drafts. I would also like to thank

the other members of my defense committee, Elliott Sclar, Stacey Sutton, and Jean-

Paul Rodrigue, for their consideration of and constructive comments on my ideas and

arguments. Jean-Paul was also particularly helpful in fleshing out my understanding

of global logistics. I also appreciate Bob Beauregard’s participation in my proposal

defense. I appropriated the concept of attenuation from his comments. Dave Madden

and Greg Smithsimon provided helpful comments on an early chapter. I am also

intellectually indebted to Josh Whitford and Neil Brenner for my understanding of

industrial organization and scale, respectively.

I would also like to express my gratitude to the technical assistance I received from

three individuals. Shafeek Fazal at the Stephen B. Luce Maritime Library on the

SUNY Maritime campus in the Bronx provided much useful guidance in identifying

data sources and pertinent periodicals. He and his colleagues were also very patient

xiii

with my use of the lovely facilities there. I received useful materials from David

Bensman at Rutgers University and Carmen Martino at New Labor. Jay Shuffield

also provided much useful information and many insightful comments throughout the

dissertation process. Without his assistance, my understanding of the Port Authority

of New York and New Jersey would have been much weaker.

Finally, I would like to acknowledge the many individuals and groups who have

challenged my ideas and contributed to my intellectual growth throughout my grad-

uate experience. In the urban planning program, I enjoyed many productive con-

versations in and out of class with Gabriella Carolini, John Powers, Erika Svedsen,

Johannes Novy, Matthew Gebhardt, James Connolly, Ingrid Olivo, Joyce Rosenthal,

Lei Wang, Shagun Mehrotra, and Shane Taylor. These acknowledgements would be

incomplete if I did not include those affiliated with the National Science Foundation-

funded Integrative Graduate Education and Research Traineeship International De-

velopment and Globalization Program under the direction of Jospeh Stiglitz, which

generously funded three years of my education and offered an intellectual space that—

for better or worse!—has rooted me in interdisciplinary approaches to development

and globalization. Though all members contributed to a lively experience, I would

particularly like to express my appreciation to Akbar Noman and those who actively

participated in the 1020 sessions: Gabriella Carolini, John Powers, Marissa King,

Matt Wai-Poi, Dan Neilson, Dan Choate, and Laura Paler.

xiv

For Yoonkyung and Sienna, who made this possible.

xv

1

Chapter 1

Introduction

Wherever the river of traffic slows down, it tends to deposit its

load: so it would be usually near the gates that the

storehouses would be built, and the inns and taverns

congregate, and in the adjoining streets the craftsmen and

merchants would set up their shops. . . . Thus the gate

produced. . . the economic quarters of the city. . . . The original

meaning of ’port’ derives from this portal.

(Lewis Mumford 1961, 305)

1.1 Introduction

Turgidly billowing clouds of mud marked the Regina Maersk’s July 1998 passage

under the Bayonne Bridge toward the Universal Marine terminal in Newark, where

its stacks of metal boxes would be lifted by cranes directly onto truck chassis and

railcars. Stretching four New York City blocks, the world’s largest container ship at

the time had already dropped off cargo in Halifax, Nova Scotia and had now lowered

its antenna and mast in order to squeeze under the bridge. The visit was Maersk’s

2

visceral declaration to the East Coast ports that they had better be able to handle

the largest, most advanced ships if they expected to host Maersk and Sea-Land’s new

consolidated hub terminal. The trail of mud reflected the turbulence in the shipping

and logistics industry in the late 1990s as the adoption of containerization and dereg-

ulation was transforming the relationship among shipping companies and between the

companies and workers and governments of port regions. This dissertation will exam-

ine the nature of this transformation as it relates to regional economic development,

particularly job creation, and port governance, and it will make recommendations for

U.S. port policy.

Three interrelated trends converged in the late 1990s to create an organizational

crisis in the industry. The first trend is evident in the preceding paragraph: techno-

logical change. Containerization, the shipment of goods in standardized metal boxes

(for legal definitions, see Rath [1975]), had facilitated intermodalism, the seamless

transfer of goods from one mode of transport to another. Inherent economies of scale

had encouraged the construction of ships too large for the Panama Canal and many

ports (earning them the moniker “Post-Panamax ships”), impelling ports toward ma-

jor infrastructural investments. Second, the deregulation of transportation was easing

logistics providers’ ability to concentrate ownership and control. Shipping companies

and terminal operators were taking advantage of these new freedoms to integrate

horizontally and vertically through mergers, acquisitions, and alliances with other

shipping companies, terminal operators, railroads, trucking companies, and other lo-

gistics providers. Third, ports, which have for most of the twentieth century been

operated by government entities, were moving increasingly towards landlord status

or complete privatization to accommodate technological and organizational change.

Taken together, these trends have dramatically transformed the playing field in

which port authorities and workers operate. The primary impact has been to elimi-

3

nate ports’ territorial monopolies over their hinterlands and to place them in direct

competition for the same markets. The second, related, impact has been to signifi-

cantly weaken their bargaining power vis-a-vis the shipping companies. And third,

the historical link between port infrastructure and economic activity, particularly

manufacturing activity, has attenuated, decoupling economic activity from freight

transportation infrastructure nodes in the U.S.

1.2 Technological change

Fifty years ago, prior to containerization, ports teemed with men struggling up and

down gangplanks weighted down by bags of coffee, bunches of bananas, and crates of

goods. Small gantry cranes hoisted painstakingly balanced palettes from ships’ holds

down to the docks, where they were transferred to handtrucks and ferried off into

a labyrinthine set of aisles or warehouses stacked high with goods. Nearby factories

would take immediate delivery of arriving goods, announcing their arrival with black,

transforming plumes of production. Other merchants would cut costs by delaying

pickup, letting their shipments clog the free flow of other loads. This activity could

go on eighteen hours a day for days at a time before a ship was emptied and free to

take on goods. When the unloading and loading were finally done, the longshoremen

would gather in their nearby bars and homes with the members of their community

to await the next ship.

According to Levinson (2006, ch. 3), the idea for containerization came in 1953

as Malcom McLean, owner of McLean Trucking, sought to counter a business threat

from domestic sea shipping. As increasing congestion in the nation’s road network re-

duced his trucks’ delivery times and his company’s profits, McLean began to perceive

domestic sea shipping companies, who had lower costs since domestic sea shipping’s

4

post-1930 decline and who were eligible for postwar discounts on military ships, as

direct competitors with ready access to his core markets. His bold strategy was to

combine shipping and trucking. While goods had been put in boxes to ease trans-

fer and minimize theft, and while some liner companies would haul the truck of any

driver who wanted to pay, McLean’s inspiration was to form one company that owned

both trucks and ships dedicated to hauling them along the coast.

Following the classic S-curve trajectory of technological adoption, containeriza-

tion growth progressed slowly at first, but its obvious economic benefits, primarily

derived from reduced labor costs, led eventually to explosive growth that may be

tapering off today. Today, an overwhelming proportion of international trade trav-

els by container ship. At today’s most advanced terminals, the longshoremen and

checkers who swarmed the docks are now invisible inside buildings behind computer-

ized control panels, and chassis drive themselves down the pier to accept one of the

standardized 40′ × 8′ × 812′

crenellated boxes being hoisted from the ship every two

minutes and lowered onto the chassis by cranes over 50 meters high. The containers

are then immediately transferred to trucks or trains and hauled directly through the

suburbs to inland, regional distribution centers, which serve as hubs for collecting and

distributing goods, much like ports in the previous era. The operations that concen-

trated around ports historically have now expanded into broader regions in a process

sometimes referred to as “port regionalization” (Notteboom and Rodrigue 2005).

In shipbuilding and shipping, there are immense economies of scale to be had, as

shipping costs decrease as ships increase in size (Cullinane and Khanna 2000, 186).

This has introduced a perverse incentive in shipping to compete in a glutted market

by building ships of even greater capacity at ever greater prices. The result has been

an exponential growth in container ship size. While the first cargo ship carried 58

containers, recent technology has permitted of ships capable of carrying 9,000. These

5

ships are too large to travel any but the deepest of waters and stretch more than 14

mile (0.4km).

In summary, technological change has contributed to larger ships, lower labor

requirements, greater capital investment, and a virtually seamless intermodal transfer

from sea to land. The transition from ports as labor-intensive central places in the

national economy to capital-intensive blips in a logistics operation reflects the eloquent

epigraph from Mumford that begins this chapter. When the river of traffic slowed

down as goods were transferred from sea to land, great port cities grew up around

them to coordinate the production and consumption of those goods. Today the river

need not slow down until it reaches distant, inland warehouses, decoupling economic

activity from port activity.

1.3 Deregulation and concentration

The history of transportation in the U.S. as told by Rose, Seely, and Barrett (2006)

describes a transition from modal separation to modal integration. Through trucking-

related legislation in the 1920s, modal fiefdoms were created within which firms com-

peted horizontally through licensing and cooperated through federally sanctioned or

operated rate-setting bureaus. Following the introduction of containerization in 1956,

political backing for deregulation slowly emerged under the sustained efforts of pres-

idents from Eisenhower to Carter, resulting in full deregulation of the airlines in

1978, railroads and trucking in 1980, and shipping in 1984 and 1998. New legislation

permitted cross-ownership among transportation modes and eliminated rate-setting

institutions. Firms, particularly shipping companies, have taken advantage of these

changes to consolidate ownership and control of entire global logistics networks.

Global shipping has long been a highly competitive business requiring some de-

6

gree of collective action for companies to remain viable. Since the late 1800s this

has taken the form of government-approved, geographically based “conferences” that

would collectively set terms for rates and services as well as punishment for shippers

who went outside the conference (Kendall and Buckley 2001). This continued even

as the disruptive new technology known as “containerization” transformed the indus-

try. Since the early 1990s, however, a qualitative change has taken place as cargo

carriers responded to the dramatic increase in international trade and competition

by vastly expanding capacity and entering mergers and alliances in order to achieve

economies of scale and to provide comprehensive global service (Blomme 2005; Slack

2004). Shipping companies have also sought to obtain more control over their product,

to distribute their commercial risk, and to improve customer relations by vertically

integrating through the acquisition of terminal operators, freight forwarders, land

haulage firms, and others (Blomme 2005, 164). In 2000, Heaver et al. summarized

the development as follows:

There is a trend unfolding in the maritime and port industries towards ever

greater control of the logistics chain through various forms of co-operation

(strategic alliances, mergers, etc.). They include both vertical agreements

along the logistics chain and horizontal agreements among suppliers of

similar services, particularly shipping companies. These developments

bring with them a danger of preferential treatment, conflicts of interests

and market dominance. (Heaver, Meersman, Moglia, and Voorde 2000,

372)

The parallel and intertwined processes of integration have increasingly steered firms,

governments, and academics to speak of “logistics chains” rather than separate types

of transport, reflecting the shift in discussion from multimodal to intermodal trans-

7

portation.

These shifts have entailed corresponding shifts in strategy. Notably there are two

complementary trends in carriers’ approach to the port hierarchy. First, hub-and-

spoke networks have emerged as a desired service configuration. In this arrangement,

cargo is transported rapidly between concentrated “load centers” where cargo is ag-

gregated or disaggregated and continuing cargo is transferred to smaller vessels for

shipment to regional ports. Second, load center “sites are being selected to serve

continents, not regions; for transhipping at the crossing points of trade lanes; and

for potential productivity and cost control” (Notteboom and Rodrigue 2005, 175).

Because the load center approach concentrates cargo throughput in a few ports and

reduces it in others that could serve the same wide geographical market, these de-

velopments generate subsidy competition among ports for selection by carriers and

terminal operators as their load centers.1

1.4 Port authorities

In most cases in North America, ports are administered by a regional port authority,

but they may exhibit a range of administrative forms. They may be landlord ports

that contract out all operations, operating ports that maintain ownership of all port

facilities and directly operate them, or some combination of the two (World Bank

2007). In either sense, port authorities function as the formal governance mechanism

for coordinating port functions and establishing and enforcing port regulations. While

some authorities are narrowly restricted to shipping concerns alone, others are per-

1Meersman, de Voorde, and Vanelslander (2005, 142) predict that the emergence of global ter-minal operators will undermine the power of carriers over port authorities. This implies not thatports will gain power over carriers but that some of carriers’ power will shift to terminal operators.Of course, some terminal operators are owned by shipping companies. In this case, carriers’ powerwould increase.

8

mitted to operate airports, commuter rail systems, bridges, tunnels, free trade zones,

and a variety of other pursuits.

Port authorities may also take a variety of institutional forms. They may be

responsible to government officials at the local, county, district, state, multi-state, or

(in Canada) federal levels. They may be independent organizations or they may be

incorporated into local, state, or national bureaucracies. Based on a small sample

of port authorities (60), Van der Lugt and De Langen (2007) identify three types

of authority goals: one oriented to the broadly defined success of the port cluster,

one oriented to profit maximization, and one a mixture of the two. These essentially

embody public, private, and public-private goals and reflect sources of financing.

While all port authorities collect users fees, their level of dependence on these for

operations and future developments varies. More publicly focused authorities have

greater access to government funds and thus more public goals, while more private

authorities depend on market mechanisms for provision and goal-setting.

As the rapid changes in logistics described above have shifted power toward the

large shipping companies and altered the terrain of economic activity, there has been

a great deal of debate over effective port strategy and an emphasis on port reform.

In these discussions, many writers conflate the wide variety of port stakeholders with

the port authorities, failing to distinguish the potential variance in goals (Fleming

and Baird 1999; Van der Lugt and De Langen 2007) (cf. Marcuse 2005). While the

orientation toward profit maximization does tend to align the interests of port author-

ities with some port stakeholders, notably firms, the willingness of port authorities to

subsidize private firms indicates that port authorities may have different or additional

goals. Thus, it is essential to distinguish port authorities from other stakeholders as

a unit of analysis. That said, the main emphasis of port reform has been a movement

away from goals other than commercial success. Proposals for port authority reform

9

generally prescribe privatization, corporatization, or commercialization. Privatization

involves selling the port facilities to a private entity. Corporatization is essentially

privatization with the government retaining a large share of the resulting company.

And commercialization indicates an increasing profit orientation (World Bank 2007).

Port consolidation and rationalization, i.e., the administration of ports at broader ad-

ministrative scales, like the national, are generally dismissed as nonstarters (Interview

with Richard Larrabee, March 25, 2008).

This trend toward commercialization has tended to shunt port authorities out of

direct port operations and into the status of landlord ports that primarily lease out

their facilities. This has led a number of scholars to investigate the repurposing of

port authorities, since their profit-orientation demands that they define their value-

adding activities. There is an increasing consensus that the coordination role of port

authorities reflects this landlord orientation. Authors like Chlomoudis, Karalis, and

Pallis (2003) have argued that port authorities emphasize their government-like roles

in setting targets for cooperation among stakeholders, directing port development

by defining the operational framework for regional port production, and forecasting.

These reflect some of the economic arguments for planning in a market society delin-

eated by Klosterman (2003): resolution of prisoner’s dilemma questions, provision of

public goods, and dissemination of information necessary for informed market choices.

Since this coordinating role cannot be limited to the immediate port cluster, we

must adopt the thinner conception of port clusters as incorporating related but geo-

graphically dispersed actors—both public and private (cf. Porter 1998; Whitford and

Potter 2007; Zeitlin 2005). The trend toward port regionalization implies such a ge-

ographical transformation. Port authorities, Notteboom and Rodrigue (2005, 307)

claim, must increasingly look to form associative relations with entities outside their

territorial jurisdiction to enhance their competitiveness. They list several major ar-

10

eas of possible cooperation between port authorities and inland distribution centers,

including traffic management, site issuing, hinterland connections and services, en-

vironmental protection, marketing, and research and development. For example the

PANYNJ has spearheaded efforts to share market information, has begun to collab-

orate with ports and governments throughout the Northeast Corridor, and has even

signed a memorandum of understanding with the Panama Canal Authority. This ex-

ample indicates three things: first, that port authorities are also beginning to head up

efforts to develop translocal networks to coordinate the provision of translocal public

goods; second, that the primary focus of translocal associations is vertical integra-

tion; and third, that the region itself may still be too small a geographical unit for

translocal port planning.

1.5 Stakes

Infrastructure, particularly freight transportation infrastructure, occupies the curious

position of serving as a vehicle for economic activity at large and as a means of

production for a particular industrial sector. Because freight movement is integral to

the production process, transportation supports virtually all economic activity and

thus serves the public interest. Insofar as private firms provide freight movement,

transportation infrastructure operates as a means of production for those firms, like

a machine in a factory. While both public and private interests stand to benefit from

more efficient, cost-effective transportation networks, their stakes differ. And these

stakes differ all the more where land meets the sea.

Freight shipping is an expensive, capital-intensive business. Containerships cost

hundreds of millions of dollars and take years to build. The docks, berths, and cranes

they require entail similar capital requirements. Generally, shipping companies have

11

borne the cost of ships and port authorities the cost of landside facilities, including the

highly valuable land itself, though private terminal operators have also often invested

in their own cranes. Additionally, the funding for dreding, which is required by most

ports, is shared by the federal government via the Army Corps of Engineers (ACE)

and local authorities.

For the private shipping companies, an efficient system offers access to large mar-

kets at low cost. As the dissertation will show, the size of this market area has changed

significantly since the advent of containerization, expanding from ports’ immediate

hinterlands to entire continents. Still, direct market access, by reducing transporta-

tion times and freight transfers, cuts costs and contributes to competitiveness. Thus,

the fundamental stake for shipping companies is their bottom line and the profits

they generate. Any arrangement that shifts the cost of infrastructure to other players

is likely to have a positive impact on the shipping companies’ profits.

The public sector, however, has interests in addition to the bottom line. Though

in many senses port authorities and government entities must adhere to their own

bottom lines (see Chapter 5), those bottom lines are often more broadly construed.

Ports can potentially generate profits through leases and user fees, but this is more

often the exception than the rule. Rather, the positive impact of port activities is

most often considered to be the jobs they create both directly and indirectly. The

industry at one time employed thousands of men in a single port to move goods

or provide services like ship repair and fueling. There are also jobs that thrive off of

supporting these workers. And perhaps most importantly, by reducing transportation

costs, ports and other freight transportation infrastructure reduce production costs

and increase the competitiveness of those firms that locate close to infrastructure

nodes. This fosters business growth and broader economic development.

Additionally, port regions sacrifice some of their local transportation networks’

12

efficiency to accommodate the movement of goods through the port and beyond the

region. The most desirable markets for shipping companies tend to be those with

the greatest number of consumers. Since large municipalities tend to face congestion

challenges simply from local traffic, adding large amounts of freight destined for other

regions exacerbates congestion for local residents and businesses.

1.6 The question

Given that logistics companies have increased their bargaining power by employing

containerization and intermodalism to scale up their activities beyond the reach of

any single port, what strategies can port authorities employ to counter these gains

and best pursue public goals, which are generally framed as direct and indirect job

creation? How can port authorities best address the impacts of competing for global

shipping firms’ business? Strategies suggested above fall into three basic categories:

commercialize, build translocal associations, and rationalize. Which will best serve

the public to which port authorities are responsible?

To get at this question, this dissertation explores the century-long history of inter-

modalism, the geographical relation between economic activity and freight transporta-

tion infrastructure, and the nature of interport competition. How did intermodalism

emerge? How has it transformed the economic terrain? Have logistics activities relo-

cated as a result of technological and organizational change in the industry? If so, to

where have these activities moved? How has intermodalism affected the location of

economic activity outside of transportation? How has intermodalism and industrial

change affected port actors, including employers and workers? How does interport

competition start and play out? How are participants impacted?

13

1.7 Theoretical approach: Technology, territory,

and terrain

Graham and Healey (1999) argue that there is an inherent tension between network

flows and territorial spaces. Network flows seek to flow past any boundary, while ter-

ritorial spaces function to contain flows (Taylor 2004). This distinction gains political

meaning from its extension from flows and territories respectively to Harvey’s (2006)

capitalistic and territorial logics of power. The capitalistic logic of power “focuses

on the ways in which economic power flows across and through continuous space, to-

wards or away from territorial entities,” borne by the flows of capital (in all its forms)

and labor. The territorial logic of power includes “the political, diplomatic and mili-

tary strategies invoked and used by a territorially defined entity such as a state as it

struggles to assert its interests and accumulate power in its own right” (Harvey 2006,

107). These two logics, while conceptually distinct, are deeply intertwined and often

interdependent.

1.7.1 Terrain

Within the capitalistic logic of power, the circulation of capital in the production

circuit is the network flow of primary importance. The circulation of capital can be

described by adapting Harvey’s (1982) elaboration of Marx’s (1977) classic formula:

M→ C → P→ C′ → M′. Money capital (M) is used to purchase commodities (C),

which are transformed through production processes (P) into new commodities (C′)

that are sold at a profit (M′). Thus, capital flows throughout the production process,

but undergoes continual transformation from one concrete form to another. The

longer this process takes, the more capital depreciates and the less profit accumulates.

14

The more expensive these transformations, e.g., through higher labor costs, the less

profitable the process is. Producers thus seek to increase profits in two basic ways.

First, they can reduce the time capital is instantiated as a commodity, that is, the

time goods are caught up in production. Second, they can reduce costs by redirecting

production to territorial entities with lower labor or input costs (Harvey 1982). These

two approaches tend toward the acceleration and regular relocation of production,

producing an ever-shifting terrain of capital accumulation.

Harvey, however, engages in a narrow focus on the capitalists’ perspective. As

Herod (2001) argues (citing Aronowitz), we must also consider the evolution of the

capitalist landscape from labor’s viewpoint. Labor, too, struggles to direct the cir-

culation of capital. From the perspective of the worker, capital is required for social

reproduction. Capital in the form of commodities (C) is consumed to generate labor

power (LP), which is combined with the means of production (MP) to earn wages

(W) that are spent on new commodities (C′) for consumption: C → LP + MP(=

P) → W → C′. This sequence defines the consumption circuit. Labor can accumu-

late greater capital within its circuit at least four ways. First, it can reduce the price

of commodities by reducing producers’ profits. Second, it can increase wages through

struggle at the point where labor power is combined with the means of production in

the production process, the workplace. Third, labor can engage in political struggle

to introduce or strengthen government regulation of prices and wages. Fourth, it may

also receive a social wage through government redistribution.

Though capitalists can have goals other than accumulation (Schumpeter 1983),

their bottom line is generally profits (Offe 1985). Thus, the goal of the capitalist is to

maximize the flow of capital in the first circuit. Workers, too, often have goals other

than accumulation, but in the context of capitalist society, their priority is also often

accumulation. Thus, the goal of the worker is to maximize the flow in the second.

15

Governments, to the extent that they operate as capitalist enterprises, also strive to

maximize flow within the first circuit. To the extent that they represent the workers

as their constituents, governments work to maximize the flow in the second.

There are thus two complementary types of flow: consumption and production.

The consumption flow is associated with workers, while the production flow is asso-

ciated with employers. Because these must manifest themselves physically (Harvey

1982), each is further associated with a concrete space. The means of production

and labor power must come together in concrete sites of production, e.g., factories

and offices, and commodities are consumed in concrete sites of consumption, e.g.,

homes and restaurants. Though these sites demonstrate a certain fixity, they too

are circulating capital, as some portion of their value is transferred to the goods or

services produced or consumed. The distinction, which sometimes puts them at odds

is the “different spatio-temporal horizon compared to the standard form of capital

circulation” (Harvey 2006, 101).

The material instantiation of these two forms of capital circulation constitute

what will be referred to here as the “economic terrain.” This dissertation will define

terrain as the momentum, or inertia, of capital flows associated with a particular

economic activity over a concrete space. Momentum is considered the quantity of

capital passing through a given space over a given period of time. For our purposes,

an increase in the volume or turnover of capital in a given area corresponds with

an increase in momentum, which is essentially economic growth in that space. For

example, keeping all else constant, if a business were to move its activities from

one city to another, the momentum would decrease in the first and increase in the

second, altering the flow of capital and hence the terrain. Or, if that business were

to simply produce larger quantities of a good each day, momentum would increase.

Note, however, that momentum has to be broken up into its constituent flows to

16

determine the actual beneficiaries of accumulation.

If we consider these two types of terrain with regard to a single product, they con-

stitute two categories that Levinson (1967) calls “product market areas” and “areas

of effective production.” The product market area is that concrete space over which

a product is sold and enters the consumption circuit, while the area of effective pro-

duction is that concrete space in which a good or service is produced. For instance,

prior to containerization the product market area for the delivery of freight2 was an

entire region, while the area of effective production was limited to piers. Similarly, in

mining, the area of effective production is limited by the source of minerals, but the

product market area is national or international in scope. In contrast construction

tends to have local areas of effective production and product market areas. In this

way, each product’s circulation is defined by and defines two types of terrain as their

momentum varies across concrete space.

1.7.2 Territory

Terrain is to be distinguished from territory. While terrain describes the countours

of flow, Jessop, Brenner, and Jones (2008) define territory as a form of sociospatial

structuration associated with efforts to bind, parcel, and enclose. The object of such

efforts is appropriately left open, but for the purpose at hand, we will focus simply

on flows of capital. Unlike Brenner and Elden (2009), who would like to constrain

“territory” to a specifically political context on the basis of presumed historical origins

and Harvey (2006), who is only considering such entities, this dissertation adopts a

broader conception based in White’s (1992) theory of identity and control. A territory

is here defined as a monopoly over some aspect of concrete space on the basis of

2Marx and Fernbach (1981) perceptively considered transportation to be a form of production,akin to workers moving goods around the shop floor.

17

conceived or perceived common interest. The territorial logic of power then refers to

an entity’s strategies for defining and sustaining a geographical monopoly in order

to accumulate power. This broader definition allows for the constant negotiation

and reproduction of overlapping territorial claims that may conflict, complement,

reinforce, or weaken each other. For instance, a gang may claim a territority in which

it possesses a monopoly over the sale of drugs that overlaps with one or more police

precincts and perhaps a merchants association, which has defined a territory defined

by the commercial activity of a particular street in which it would like to maintain

standards of appearance and behavior. Each organization makes territorial claims to

control specific socio-economic aspects of a concrete space that constantly interact to

define that space. From this perspective, organized labor is treated as a territorial

entity akin to organized states.

For our purposes, territory is oriented toward monopolizing (or at least containing)

the flow of capital and increasing its momentum within a given concrete space, that

is, it is oriented toward accumulation within a delineated concrete space. This is

achieved in three ways. First, a territorial entity can increase the quantity of capital

by keeping it from passing outside its boundaries. For instance, import substitution

keeps capital that would leave an administrative territory for processing elsewhere

from leaving the territory. This, in effect, lengthens the circuit of capital within that

territory. Second, a territorial entity can increase the volume of capital by diverting

it inside its territory from outside. An example of this would be convincing a new

factory to locate within the territorial jurisdiction. Third, by increasing the velocity

of capital, primarily through improved efficiency, a larger portion of capital is retained

within the territory. For example, a process improvement will at least for some time

increase a companies’ profits that would otherwise have passed out of the territory

with the product.

18

Territory is more or less easily defended as the common interest of the actors who

define it is stronger or weaker. That is, if the members of a social group share a

common interest, they are better able to band together in defense of that interest

than are the members of a social group with conflicting interests. This conception

is rooted in White’s (1992) idea that identity underlies control while at the same

time being an object of control. Its practical application derives from Offe’s (1985)

reaction to Olson’s (1971) argument on the logic of collective action. Offe argues that

it is easier for capitalists than workers to organize collectively not just because they

are fewer in number but also because they share a much narrower range of interests,

most particularly profits. Workers, on the other hand, have to collectively organize a

common interest not only in accumulation but also in a vast range of other dimensions

that infuse their lived experience, like obtaining job security, overcoming racism, and

enjoying work and home life. The difficulty in coordinating non-accumulation interests

undermines workers’ ability to maintain a common front in the face of workplace

challenges. The same argument could be made for government entities, particulary

subnational government entities. They may share a common interest in national

economic growth, but they may differ in their opinion on how it should be distributed

or in which policies would lead to local economic growth.

1.7.3 Technology

Technology is embodied in specific means of production, which dictates that it be

located at specific points in concrete space. As a means of production, technology

thus becomes the primary site for the negotiation of process and profits between

workers and employers. If workers can establish a territorial monopoly over the means

of production, they can divert a portion of capital (profits) into the consumption

19

circuit. If they cannot, employers can divert capital from the consumption circuit to

the production circuit.

The same holds true for governments. Because technology must be embodied

in specific locations, its form determines the area of effective production. Means of

production that require only a machine that can be operated virtually anywhere, e.g.,

a laptop computer, have an effectively unlimited area of effective production and thus

are more difficult to monopolize. A technology like container shipping, which requires

a pier with deep berths, has a much more limited area of effective production and is

thus easier to monopolize. If a government is able to monopolize the sites at which

a specific technological process can be employed, it can divert a greater portion of

capital generated by that production process into the circulation of capital within its

boundaries.

Technological innovation can be used as a tool to alter relations between labor and

capital. For instance, labor-reducing technologies cut down the total number of jobs,

creating a surplus of workers that allows employers to pay lower wages. Technological

innovation can also be used to alter the economic terrain, shifting some stages of

production to regions with more pliant workers (cf. Gordon 1978).

1.8 Structure of dissertation

The dissertation consists of nine chapters, including this introduction. It builds an ar-

gument for port rationalization as an effective counter to global shipping’s extractive

power by showing how containerization was employed to strengthen capital’s position

vis-a-vis labor and governments to the latters’ detriment. The various methodologies

employed in the volume—quantitative analysis of historical data, examination of his-

torical documents and secondary sources, and interviews—are described in Chapter 2.

20

Chapter 3 traces the development of containerization and intermodalism with a focus

on the twentieth century. The systems approach to technology studies pioneered by

Hughes and Luhmann is employed to illuminate the way in which containerization

emerged as a second-order system that tied previously existing logistics infrastruc-

ture into one global system and the attendant organizational changes. The spatial

impact of intermodalism on the economic terrain of the U.S. freight system, essen-

tially the restructuring of ports, is explored in Chapter 4, which shows how freight

traffic has shifted westward and how warehousing activities have moved away from the

coasts to a band a few hundred kilometers inland. Chapter 5 then illustrates how the

shift in terrain has weakened the bargaining power of ports vis-a-vis global shipping

companies, resulting in massive subsidies for these companies. The claim that these

subsidies are justified due to their role in job creation is examined in Chapters 6 and

7. These chapters fill a gap in efforts to evaluate the impact of ports on employment

by looking at time trends rather than static extrapolations of current employment

figures. Chapter 6 looks at direct employment in port industries, while Chapter 7

looks at the relation of a broader range of economic activities to freight infrastruc-

ture. They conclude that direct employment is declining and that economic activity

has decoupled from the nation’s freight infrastructure, indicating that governments

that host ports see little in the way of employment benefits. Having concluded that

ports have been losing money chasing non-existent job creation gains, Chapter 8 turns

back the clock to better understand how longshoremen were able to build a territorial

monopoly over a port range and obtain better working conditions. Finally, Chapter

9 argues that port authorities should follow the example of early twentieth century

longshoremen to build a territorial monopoly of their own through port rationaliza-

tion. It is suggested that this will not only stop destructive incentive competition but

also has the potential to more equitably distribute the gains of economic growth.

21

Chapter 2

Methodology

The research that contributes to this document involves mixed methods. Archival

research is employed to construct the historical narratives in the text, particularly

Chapters 3 and 8. This work relies primarily on contemporary sources and secondary

reports of historical developments in the labor and technology of transportation.

Quantitative analysis is incorporated to facilitate an understanding of the spatial

implications of sociotechnical change. Data sets for these analyses includes exist-

ing government-produced data and data compiled from industry publications by the

author.

2.1 Quantitative Analyses

The follow sections present the basic set up for the regression analyses and mapping

that underlie and inform several of the following chapters, particularly Chapters 4, 6,

and 7.

22

2.1.1 Computing and software

The data were processed through PostgreSQL 8.3 and 8.4, R 2.9.2 and 2.11.1, and

QGIS 1.0.1 on a computer with an x86 64 Ubuntu Jaunty 9.04 and subsequently

Lucid 10.04 operating system.

2.1.2 Data

The data for this study is drawn from three main sources. The first is County Business

Patterns (U.S. Dept. of Commerce, Bureau of the Census 1986, 1987, 1992, 1998,

2003, 2004, 2007, 2009), which is prepared annually by the U.S. Bureau of the Census

and contains information on total annual pay and total employment by firm size.

Data sets have been chosen in five year increments from 1974 to 2004, which avoids

all major recessions during the time period, coincides every second time with the

collection of census data, and is a sufficiently long period to observe statistical change.

Additionally, 1970 and 2007 data are appended. While these do not adhere to the five

year increment, they are respectively the earliest and latest data available at the time

and functional for the purposes of this analysis. The second data set consists of U.S.

Census data generated by the National Historical Geographic Information System

(Minnesota Population Center 2004) and linked to their historical geographical files,

which allows for the calculation of distances between points (e.g., county centroids,

which are the geographical center of a county, and international airports). Tax rate

figures are collected from the Council of State Governments (1978, 1984, 1988, 1994,

1998, 2004), which lists state level corporate tax rates for each state. Two notes are

in order with regard to the tax rate information. First, though there are often figures

for small businesses and large business, the rates for large businesses are used, as

these firms are the most likely to make decisions on the basis of tax rates. Second,

23

these numbers do not take into account any incentives offered to particular industries

and thus may underestimate the impact of these rates for those industries.

Data is all at the county level, as this is the only level at which sufficient and

intertemporally consistent data on employment by sector is available.

Despite the stability at the county level, it is important to bear in mind that

many industrial classifications are inconsistent over time as the government redefined

them to account for contemporary economic change. This is particularly so for the

1997 reclassification from the Standard Industrial Classification (SIC) to the North

American Industrial Classification System (NAICS) to unify the Canadian, Mexican,

and American systems in the wake of the North American Free Trade Agreement

(NAFTA). As a result of the reclassification, many industrial sectors cannot be com-

pared across this gap. To gain the most insight despite this gap, the following analysis

makes every effort to choose classifications that are consistent throughout the series.

As a consequence, industrial sectors are often narrower than ideal.

2.1.3 Mapping

The maps in this section are all based on the County Business Patterns data linked

to NHGIS map files via postGIS. Numbers for location quotients were generated in

PostgreSQL by dividing the ratio of county level employment in the given industry

(in this case, warehousing) by total employment in the county and then dividing this

by the same ratio for the country as a whole. Thus values greater than one indicate

an above average concentration of employment in the given sector, and values less

than one indicate below average concentrations. In this manner, it is possible to trace

the relative importance of a given industry to a county’s wellbeing.

24

2.1.4 Regressions

The regressions add geographical variables to analyses that generally rely solely on

socioeconomic indicators to predict industrial location. Most evaluations of industrial

location include variables such as education, race, foreign-born population, and wage

rates with politically determined tax rates as predictors. The regressions here add a

measure of distance from three types of transportation infrastructure nodes to county

centroids. The first of these is ports. Based on the latitude and longitude provided by

the UN location code system (United Nations Economic Commission for Europe 2007)

and augmented with information from World Port Source (2009), all major North

American ports listed in Containerisation International (2006) were included. This

necessarily excludes some small ports that handle containers, but none of those with

foreign calls. The second measure of distance is from county centroids to airports with

customs landing approval from the FAA as of 2002 (Federal Highway Administration

(FHWA) 2007). This excludes many local airports, but ensures that all airports

approved to serve as ports of entry for freight are included. The final geographical

variable is distance from county centroid to intermodal terminals in 2002, based on

data in Federal Highway Administration (FHWA) (2007). Intermodal terminals serve

as points for the transfer between rail and either ships or trucks. Thus these terminals

are located both on the coast and inland and capture the relative importance of rail to

the location of economic activity. One caveat must be added, however. As far as the

author has been able to determine, the only complete maps for these infrastructure

nodes date from 2002. Therefore, all of these nodes are treated as if they existed

throughout the time period studied. This is highly unlikely, but it may well be the

case that most of these facilities existed in one form or the other throughout the

period.

25

The operative hypothesis, as will be expanded in Chapter 4, is that

containerization has reduced the likelihood that economic activity will

locate near ports and will instead migrate toward the new break bulk

points determined by concentrations of warehouses. Prior to containerization,

most bulk was broken at the pier and then shipped off to factories and warehouses.

With the reduction of time and labor required to move freight through ports, goods are

now generally moved in containers to inland warehouses, where they are opened and

the goods redistributed. Thus, if economic activity is coupled to break bulk points,

whether for convenience or cost, then during the era of containerization, economic

activity should be moving away from ports.

Variables used in the regressions are listed and described in Table 2.1.

Table 2.1: Regression variables

Variables Descriptions

emp log Log of total employment in selected industry classificationpop density log Log of population density, which reflects the relative urbanization of

the countyinc per cap log Log of per capita income, which reflects wage levels and purchasing

powertaxrate Tax rate from previous period (except 1970 and 1974, which use rate

from 1974) in percentba Percent of population with bachelor’s degrees or higherhs Percent of population with a high school diplomanonwhite Percent of nonwhite populationforeign Percent of foreigne populationport 100km from county centroid to closest portairport 100km from county centroid to closest international airportintermodal 100km from county centroid to closest rail-based intermodal terminal

(as recorded in 2000)

26

2.1.5 Notes for interpreting plots and regressions

A few tips on interpreting the plots and regressions are in order to facilitate the

reader’s understanding of the results of the analysis.

Since the regressions are logarithmic (based on clear trends indicated in prelimi-

nary pairwise scatterplots), a one unit change in the coefficients of non-logged inde-

pendent variables can be interpreted roughly as the percent change in employment.

The further away from zero the values get the more important it is to exponentiate

the estimator, bearing in mind that the impact is multiplicative, not additive.

Second, the plots have vertical markers to represent reclassifications. The widest

is the 1997 transition to NAICS. This should help the reader avoid making too many

quick assumptions about the underlying numbers. There is also a horizontal grey line

at y = 0 (for those plots that include it). This line is provided as a reference not

just to the magnitude of effects but also to aid in evaluating the significance of an

estimator. The vertical lines through the estimator points represent the 50th (thick)

and 95th (thin) percentile confidence intervals. If these lines pass above or below

zero, the estimator is not significant at that level. This will hopefully save the reader

some unnecessary flipping back and forth between the plots and the regressions.

Finally, a good rule of thumb for interpreting the plots is that if the red trend

line is increasing, employment is moving away from ports, airports, or intermodal

terminals, as appropriate. And vice versa, if the line has a negative slope, the economic

activity is getting closer. Negative values imply a concentration that trails away,

while positive values indicate increasing employment the farther one gets. As values

approach zero, they indicate an increasing geographical deconcentration (an evening

out) of the economic activity with respect to the infrastructural node in question.

However, one should note that the confidence intervals, while often highly signifi-

27

cant in any given year’s regression, are wide enough that any interpretation of trend

is indicative at best. Most estimators in most regressions have confidence intervals

that overlap with each other, reducing the statistical possibility that the true value is

actually moving. That said, some trends seem apparent, but should interpreted with

great caution.

2.2 Historical analyses

The material for Chapters 3 and 8 is drawn from a mixture of primary and secondary

sources. Secondary sources are cited throughout the text, but particularly important

sources include Finlay (1988), Herod (2001), and Hobsbawm (1964). Primary sources

have provided invaluable information on working conditions, labor processes, and

technological trends, particularly for turn of the century dockers. Barnes’s (1915)

groundbreaking report for the Russell Sage Foundation, the first systematic study

of dock work and dockers in the US, proved invaluable. The U.S. Bureau of Ap-

plied Economics (1920) and National Industrial Conference Board (1921, 26) supply

information on standards of living at the time, from which the living conditions of

longshoremen can be deduced. Copies of The Waterfront Worker, the radical newspa-

per by and for longshoremen on the West Coast in the 1930s, offer labor’s perspective

on the most vital episode in longshore unionization. And an International Labour

Organisation working draft (International Labour Office 1973) helped identify the

core components of mechanization and modernization agreements and their relevance

and attractiveness around the globe.

Information on cargo handling efficiencies and port performance are drawn from

Bureau of Labor Statistics bulletins, especially U.S. Bureau of Labor Statistics (1932)

and Maritime Cargo Transportation Conference (1957), and an evaluation of Cana-

28

dian ports by Picard (1967) for the Minister of Labour. Contemporary journals,

including the Journal of Commerce and the compellingly titled Cargo Handling, also

supply information on cargo handling technologies and processes and insights into

the agreements and disagreements among employers on the future direction of those

technologies.

Material for the Chapter 5 on port competition was drawn primarily from the

Journal of Commerce and The New York Times. Efforts were made to contact actors

involved in the struggle for Maersk’s terminal, but they were universally rebuffed. The

Maryland Ports Authority (MPA) was unable to “find anyone at the MPA who was

close enough to the deal or has the institutional knowledge regarding the Maersk deal”

(Miller 2009). Maersk representatives refused to return my phone calls. And the Port

Authority of New York and New Jersey, after a preliminary interview, was unable to

cooperate with me because of a sensitive, ongoing lawsuit related to the deal, which

is touched upon in the chapter. However, the two sources provide sufficient numbers

of direct quotes and relevant analyses that I believe provide sufficient grounds for

the claims required by this document. Also, most of the information on PANYNJ

revenues, expenses, and volumes were taken directly from the Authority’s annual

reports (Port Authority of New York & New Jersey 2006).

The minutes of the North Atlantic Ports Association (North Atlantic Ports Associ-

ation 1999), housed at SUNY Maritime’s library in Bronx, NY were also investigated

in regard to the form and results of interport cooperation.

2.3 Reproducibility and literate programming

This project has adopted a foundational methodological approach by joining a con-

temporary movement that favors reproducible research and literate programming.

29

The ability to reproduce experimental results has long functioned as the cornerstone

of the natural sciences. As Popper (1959, 23) writes:

We do not take even our own observations quite seriously, or accept them

as scientific observations, until we have repeated and tested them. Only

by such repetitions can we convince ourselves that we are not dealing with

a mere isolated ‘coincidence’, but with events which, on account of their

regularity and reproducibility, are in principle inter-subjectively testable.

This ideal, of course, is not so simply implemented. In the natural sciences, the

particular equipment in a particular laboratory or prohibitively expensive equipment

can bar the reproduction of scientific results. In the social sciences, where researchers

examine complex interactions incapable of being isolated effectively from their envi-

ronments, reproducibility is an impossible goal. This is primarily because social in-

teraction takes place in perpetually evolving social situations (Sayer 1992) and there

is no way to turn back the clock on these changes (Prigogine 1997).

In one area of the social sciences, however, a limited form of reproducibility is

possible: statistical analysis. By opening up the datasets and the programming code

used to evaluate the data, other researchers can test the coding and the analysis by

running it on their own machines. In this way, problematic coding and, to a lesser ex-

tent, errant data can be more quickly identified and rectified. Additionally, it permits

other researchers more readily to consider alternative statistical approaches. While

opening up data and coding in this way unravels individual monopolies on producing

results (and may therefore contribute less to individual careers and encourage free

riding), it is clearly superior as a way of contributing to knowledge.

As a consequence, the coding itself should take on the characteristics of a well

written essay, so that it is readable to those unfamiliar with the coding itself. This

30

concept of “literate programming” was introduced by Don Knuth in 1984 and has

resulted in the use of clearly understandable variable names and copious commenting.

This approach was facilitated by his development of the WEB system, which wove

quality typesetting with actual programming code (Knuth 1984).

Fortunately, a contemporary approach for doing the same with statistical program-

ming came to my attention as I began preparing this document. It is now possible

to weave document coding with statistical coding through two open source programs:

LATEX, which provides sophisticated typesetting, and R, which is an open source sta-

tistical programming language in wide use. In a distant echo of Don Knuth’s original

excitement at developing the WEB system (“I enjoy the new methodology so much that

it is hard for me to refrain from going back to every program that I’ve ever written

and recasting it in ‘literate’ form.” [Knuth 1984, 1]), this document has embraced the

combination of LATEXand R using the pgfWeave and SWeave packages for R.

31

Chapter 3

Technological and organizational

change

3.1 Introduction

In his 1983 book Networks of Power, Thomas P. Hughes introduced the concept

of large technical systems (LTS) in his pathbreaking account of the invention and

evolution of electrification by integrating systems theory into historical narratives

(Hughes 1983). His framework proves a fertile ground for depicting the introduction

and expansion of containerization and its role in intermodalism.

Hughes (1983, 5–6) defines a system thus:

A system is constituted of related parts and components. These com-

ponents are connected by a network, or structure. . . . The interconnected

components of technical systems are often centrally controlled, and usually

the limits of the system are established by the extent of this control. Con-

trols are exercised in order to optimize the system’s performance and to

direct the system toward the achievement of goals. . . . Because the compo-

32

nents are related by the network of interconnections, the state, or activity,

of one component influences the state, or activity, of other components in

the system. . . . Those parts of the world that are not subject to a system’s

control, but that influence the system, are called the environment.

In the case of large technical systems, which rely on complex political and organiza-

tional arrangements for development and operation, like electrical systems and freight

systems, Hughes (1983, 465) labels these systems “sociotechnical systems.” His inten-

tion is to emphasize his claim that while individual and organizational actors play

primary roles in shaping and directing these complex systems, systems themselves

have “an internal drive and increasing momentum” (Hughes 1983, 462) as technical,

financial, and organizational resources invested on the basis of earlier decisions be-

come more difficult to alter as the systems evolve.1 The actors that do the work

of combining technological, political, and economic interests and resources into the

creation, development, and expansion of systems are dubbed “system builders.”

Following Joerges’s (1988) account of Hughes, the evolution of LTS proceeds

through three stages. The first stage spans invention, development, and innovation,

and the dominant system builders are technical inventor-entrepreneurs, who, like Edi-

son, are entrepreneurial in their development of both technology and customer base.

Invention ends in the introduction of a new technical system that must be developed

by embedding it within the political and economic context, after which innovation puts

it into proper practice. The second stage is the transfer of the system to new political

and economic contexts, which entails a myriad of organizational and often technical

adjustments. The dominant system builders at this point are engineer-entrepreneurs,

who adapt existing technologies to the local “technological style,” which refers to the

1A similar, though perhaps more extreme view, is expressed by actor-network theory (ANT)(Callon 1986; Latour 1996, 1997, 1999).

33

geographical, economic, political, and organizational factors pertinent to a given lo-

cation. The final stage spans growth, competition, and consolidation. The system

builders at this stage, manager-entrepreneurs and financier-entrepreneurs, focus their

efforts on rationalizing the system, making it more efficient, and concentrating capital.

The changes that occur during this stage are often driven by system builders’ atten-

tion to load factor, a term Hughes borrows from the power industry. “Load factor is

the ratio of the average load to the maximum load of a customer, group of customers,

or the entire system during a specified period” (Hughes 1983, 218). Maximizing load

factor is equivalent to maximizing efficiency and often entails expansions into new

customer bases. For example, one of the early pioneers of electrification, Samuel In-

sull, actively pursued transportation company contracts for street car power because

his analyses suggested that peak energy consumption during commuting hours would

complement that of lighting before and after commuting to and from work and that

of stationary motors in the work place.

Another useful term introduced by Hughes is reverse salient in place of “bottle-

neck” and “disequilibrium.” Drawn from military use, where it refers to a section

of a battle line or military front that remains connected to the rest but has fallen

behind, “the concept of a reverse salient refers to an extremely complex situation

in which individuals, groups, material forces, historical influences, and other factors

have idiosyncratic, causal roles, and in which accidents as well as trends play a part”

(Hughes 1983, 79). A component of a system that holds the other components back

on their way to achieve an actor-defined goal, some critical problem that stymies the

full functioning of the system, is considered a reverse salient. Hughes calls solutions

to reverse salients that sustain the current arrangement “conservative,” and solutions

that entail fundamental redesigns or the introduction of new, competing systems“rad-

ical” (Joerges 1988, 13). Summerton (1994a, 13) calls attention to the “particularly

34

striking” reverse salient of congestion in physical systems, for which the solution may

be not only increasing the physical size of a given network but also increasing the

regularity and predictability of movements.

The contributions to Changing Large Technical Systems place great emphasis on

boundary crossing by LTS (Summerton 1994b). In the introduction, Summerton

(1994a, 7–10) summarizes the contributions. She argues first that researchers have

conventionally treated LTS as individual systems, but that the contributions, par-

ticularly those of Braun and Joerges (1994), Bucholz (1994), Schneider (1994), and

Usselman (1994), demonstrate that LTS are essentially “second-order systems” that

integrate heterogeneous systems into a new system. For example, Bucholz (1994)

suggests that railroads and telecommunications are generally considered independent

LTS but were integrated into the traditional European armies prior to World War I

and transformed them into a second-order military system. The second observation

Summerton makes is that such processes of integration often cross both physical and

institutional territorial borders. For example, the takeover of an American railway by

a Canadian firm entails not only organizational crossing of national boundaries but

also institutional changes to legislation permitting such takeovers. (See also Egyedi

[1996] and Heins [2009].)

The LTS approach offers a useful framework for exploring the evolution of global

freight transportation and containerization in particular. Born of a trucker’s attempt

to overcome highway congestion in the Northeast, containerization was created as

a second-order system. Actual implementation varied in accordance with the spe-

cific needs and character of the major firms that first developed systems. Global

growth in container traffic faced the reverse salient of heterogeneous country- and

company-specific standards for rail gauges, truck chassis, and locking systems. The

shipping industry is plagued by global imbalances in the flow of trade as well as by

35

booms and busts in global trade that lead to cycles of undercapacity and overcapac-

ity. In response, shipping companies have competed through economies of scale in

ship size and through efforts to maximize the load factor for each journey. For exam-

ple, as manufacturing moved from North America and Europe to East Asia, loaded

containers fill ships travelling east to west, while empty containers dominate in the

opposite direction, leading shipping companies to offer significantly lower rates for

the latter journey. Fierce competition has led ultimately to organizational consoli-

dation, as the terms of survival demand the construction of fleets of expensive ships

and a global reach. Concentration of capital was also assisted by the deregulation of

the transportation industry, which relaxed prohibitions against crossing institutional

boundaries between transportation modes.

Hughes’ three stages can be summarized as an initial growth phase, an acceler-

ated growth phase, and a stabilization phase, which may be followed by a period of

decline (Gokalp 1992). Thus they can be loosely mapped onto the S -shaped curve in

measuring the growth of a new technology (e.g., the automobile) generally observed

in technology and transportation studies (Garrison and Levinson 2006, 46–50). In

this model, growth is slow when a technology is first introduced (stage one) and

transferred to new locations (stage two). This is represented by the bottom of the S.

Hughes’ growth and competition are reflected in rapid growth for some time as the

technology takes off (the middle of the curve). And finally, as the technology and

industry mature, growth slows down and levels out (the top of the S ).

Containerization, like other successful transportation technologies (Grubler 1990),

appears to be following suit. Figures 3.1 and 3.2 depict the growth of containerization

on the world stage and within the U.S., respectively. Both plots exhibit the same

pattern: comparatively slow growth until the early 1990s followed by a period of more

rapid growth. If the S -curve trajectory is to hold true for containerization, these

36

Con

tain

ervo

lum

e(m

illion

TE

Us)

1975 1980 1985 1990 1995 2000 2005

010

020

030

040

0

Largest

Top 5

Top 10

Top 20

Top 50

Top 100

All

Figure 3.1: Total volume of containers handled by world ports. Each line representsa different grouping of ports. Demonstrates the rapid increase in total volume fromthe early 1990s.Source: Containerisation International Yearbook

plots suggest that, as a technology, containerization has not yet reached maturity.

The recent economic crisis, which is not yet reflected in available data, may be a

turning point, but this will not be apparent for some years yet. The year-on-year

growth rates plotted in Figure 3.3 suggest a fairly even rate of growth in container

volume that averages about ten percent, which implies a doubling of volume every

seven years. The U.S. shows greater variability with a mean of roughly seven percent,

which implies a doubling of volume every ten years. Of particular note is the higher

rate of growth after the early 1990s. Two points are relevant here. First, growth was

more rapid than in previous decades. Second, the growth rate increased steadily while

the world growth rate remained steady, suggesting that there were changes unique to

37

Con

tain

ervo

lum

e(m

illion

TE

Us)

1975 1980 1985 1990 1995 2000 2005

010

2030

40

Largest

Top 5

Top 10

All

Figure 3.2: Total volume of containers handled by U.S. ports. Each line represents adifferent grouping of ports. Demonstrates a steady increase through the early 1990sfollowed by rapid growth.Source: Containerisation International Yearbook

the US responsible for more rapid growth. These points will be discussed below.

3.2 Stage One: Invention, development, and

innovation

3.2.1 Nineteenth century intermodalism

Intermodalism and multimodalism refer most generally to the transport of goods from

origin to destination via more than one mode of transportation. Efficiencies can be

gained either by increasing the average speed of the modes of transportation involved

or by reducing the time and cost of transferring goods from one mode to the other. It

38

1980 1985 1990 1995 2000 2005

-50

510

15

Per

cent

grow

thye

ar-o

n-y

ear

United StatesWorld

Figure 3.3: Containerization growth rate for the United States and the world. Notethat world growth rates during this period are consistently higher than rates forthe U.S. The dashed black line represents the U.S. locally weighted regression line foravailable data. The dashed grey line represents the world’s locally weighted regressionline. Grey background indicates periods of recession in the U.S.Source: Containerisation International Yearbook

is this latter concern of reducing or eliminating the obstacles to the smooth circulation

of goods into which containerization as a technology falls.

In a sense, the first bags and boxes employed to transport goods could be consid-

ered the precursors to containerization. They were employed with the same funda-

mental ends in mind: to simplify loading and unloading and to protect goods from

damage and theft. Containers, however, are products of rationalization and mecha-

nization under contemporary capitalism. It is thus best to go back no further than

Dr. James Anderson’s 1801 articulation of a container for transporting goods and his

1845 patent for a container that could be easily transfered between horse-drawn car-

39

riages and the expanding number of railcars. In the United States, precursors could

be seen in the “pantechnicon vans” and “lift vans” of the early twentieth century. Pan-

technicon vans were wooden crates approximately twelve to eighteen feet long and up

to seven feet wide that were drawn by horses and used to transport household effects

from one home to another. Lift vans isolated the crate so that it could be loaded

by cranes onto horse-drawn carriages or trolleys and were available for transatlantic

shipping (Palmer and DeGiulio 1989, 285–86).

Multimodal principles were also employed in moving commercial goods in the U.S.

at this time. As early as the 1830s, the Baltimore & Ohio Railroad offered “piggyback

service,” in which container vans were carried on flatcars. Similar efforts were made by

the Long Island Railroad after 1885, when it started “Farmers’ Trains” to transport

produce to New York City’s East River in produce wagons lashed to rail flatcars.

Also, from 1843 to 1857, the Pennsylvania Railroad Company carried sectionalized

canal boats on flatcars within Pennsylvania (Palmer and DeGiulio 1989, 285–86).

So clearly, the concepts and advantages underlying rapid transfer of “containerized”

cargo is a long-standing one in transportation.

3.2.2 Early twentieth century

A seamed system

Early transportation legislation embodied two seemingly contradictory approaches,

one privileging the market and the other depending on government regulation. Both,

however, led to decades of legally mandated modal separation (Palmer and DeGiulio

1989, 311–324). Based on the perception of railroads as monopolists (Rose et al. 2006,

3), the Interstate Commerce Act of 1887 embraced antitrust views that saw free mar-

ket competition as the best regulator of prices and rates. The second approach,

40

embodied in the Shipping Act of 1916, treats the transportation industry as a public

utility that exists to serve the public interest and must therefore be controlled to pre-

vent excessive rates, discrimination, and other abuses. Driving the Shipping Act is the

existence of shipping conferences over given routes, which are essentially rate-setting

cartels that were established to counter overcapacity and cutthroat competition. The

U.S. government felt that such protection was warranted as it saw the loss of domes-

tic maritime capacity through competition with foreign carriers as undesirable from

a military standpoint (Branch 1996; Kendall and Buckley 2001; Levinson 2006).

To effect these goals, both laws established government control over routes and

rates for interstate transportation. Originally only applicable to railroads, the In-

terstate Commerce Act was expanded in 1935 to cover motor carriers, in 1940 to

cover water carriers, and in 1942 to cover freight forwarders (Palmer and DeGiulio

1989, 311–313). Common carriers subject to the act were obliged to submit their

rates, routes, rules, and practices to the Interstate Commerce Commission (ICC) for

approval before they could be implemented. The rates had to be reasonable and

nondiscriminatory. While the ICC had jurisdiction over motor carriers, railroads,

and certain water carriers, the government also created the Federal Maritime Com-

mission (formerly the United States Shipping Board) to regulate both domestic and

international common carriers by water.

The impact of this organizational framework was to ensure decades of modal sep-

aration. Cross ownership of more than one mode of transportation was seen as anti-

competitive and actively blocked. Railroads in particular were forbidden from owning

shipping lines and trucking companies. The former prohibition was explicitly stated

in the Panama Canal Act of 1912 (Rose et al. 2006, 3) and the latter enforced by

the ICC as essential to land-based competition. Despite efforts by some transporta-

tion economists and and independent analysts to advocate for a coordinated national

41

transportation system that took into account all modes and encouraged intermix-

ing them, Congress passed further legistlation in the late 1930s that produced “a

‘national’ system composed of separate transportation industries and separate trans-

portation markets, each now defined variously as technology or mode and governed

by several equally disconnected policies and regulatory agencies” (Rose et al. 2006,

32–33). Legislation was used to ensure that each mode, in Hughes’ terms, functioned

as an independent, first-order system. The transportation system as a whole then

could be thought of as a system with clear seams between modes or as a collection of

independent systems.

One outcome of this separation, which becomes a major motivation for dereg-

ulation after the advent of containerization, is that freight transportation providers

could not offer joint through rates. Through rates are door-to-door prices for shipping

goods, and a joint through rate covers a journey that requires multiple modes to com-

plete. Thus, for instance, rather than quote a single price for pick up, transport, and

delivery, a water carrier could only offer a price from port to port; the other legs of

the journey would involve separate prices quoted by trucking companies or railroads.

This compelled shippers2 to spend additional effort and build additional expertise in

transport arrangement or depend on a single alternative quoted by an intermediary

(Palmer and DeGiulio 1989, 314–315).

On the docks

Mechanization crept slowly onto U.S. docks during the first two decades of the twen-

tieth century. Advances were generally limited to replacing onboard hand-operated

winches with steam-powered versions. Similarly, some winches were also installed on

2In keeping with the conventions of the literature on transportation, companies that send theirproducts to other destinations are referred to as “shippers,” while the companies that transportgoods are referred to variously as “shipping companies,”“liner companies,” and “logistics providers.”

42

the docks, but this was fairly rare, although the most advanced European shipyards

were installing their first large cranes for moving cargo from ship to shore (Barnes

1915). Additionally, small trucks came into use for hauling goods away from the pier,

and later fork-lift trucks were introduced to move goods around the docks, augmenting

the dollies that longshoremen traditionally used.

Intermodalism

While change was slow on the docks, it was advancing more rapidly on the rails.

Intermodal operations rapidly became commonplace over the first half of the twen-

tieth century, exploding at mid-century. Carfloat service (railcars on barges) from

Greenville, NJ to New York City began in 1904 and constituted ninety percent of the

city’s water-borne cargo by 1929. In the 1920s modern rail container services were

introduced between Cleveland and Chicago, trailer-on-flatcar (TOFC) service began

between Chicago and Milwaukee, and by 1928 the Pennsylvania Railroad had es-

tablished container service between New York, Baltimore, Philadelphia, Pittsburgh,

Cleveland, and Buffalo. However, in 1931, the Interstate Commerce Commission

(ICC), which licensed and controlled insterstate transportation, passed down the In

re Container Service decision that quashed truck and rail integration for nearly twenty

years in keeping with its emphasis on weakening railroads’ monopolistic powers by

sustaining modal competition. In their findings, the ICC concluded that because

the shippers greatly benefited from TOFC services, this sought-after service did “not

need the encouragement” of lower rates, which it deemed “unreasonable, unjustly

discriminatory, and unduly prejudicial.” The higher rates subsequently attached to

container services than breakbulk shipping by the Commission made it impossible for

the railroads to compete with other modes of transportation, and they thus ceased

such services, reestablishing the modal separation between trucking and rail (Palmer

43

and DeGiulio 1989, 290–292).

Meanwhile, in the UK during the late 1920s, British Railways also introduced rail-

way containers to gain all the benefits attributed to boxing cargo (reduced breakage,

reduced theft, ability to prepackage, and intermodal convenience). These containers

were small by current standards (about 2–4 tons), some were refrigerated, and some

were open. There were also cargo-specific containers. The extent of intermodalism’s

importance was such that some containers were “demountable truck bodies which can

be freely moved by both rail and road vehicles.” (Hammond 1957, 208)

3.2.3 Mid-century

Military role

The military’s role in developing and disseminating technological advances in logis-

tics has been consistent and significant over the twentieth century, if of secondary

importance. At mid-century, the military contributed in two ways. First, logistics

took on renewed importance after early failures at the beginning of World War II

and led to the improvement of pallet design and use. Maloney (1996, 2–3) argues

that between the wars, commercial shippers had realized cost savings by cutting back

on packaging, but that this weaker packaging was not up to the rigors of military

deployment. After spectacular losses as cartons of supplies disintegrated under wet

conditions in Guadalcanal and in the North African and Pacific Theaters, the military

reinvigorated its logistics departments, which issued new standards for packaging and

developed techniques for palletizing cargo.

The military’s second contribution during World War II and the Korean War

was to expand the use of pallets and mechanization to less developed ports, like

Tacoma (Magden and Martinson 1982) as well as overseas ports (Cargo Handling

44

1956a). This facilitated the expansion of mechanized loading and unloading in the

commercial sector as similar technologies could be employed at more points in the

system.

Efforts toward unitization

The military was not alone, of course, in recognizing the advantages of palletiza-

tion and seeking more efficient and secure packaging methods. Experiments with

palletization and unitization dating from the late 1920s led to rapid adoption, deep

penetration, and international standardization by the late 1950s. This evolution of-

fers three insights into the development of containerization. First, it makes it clear

that unitization was a technique for reducing labor costs. Second, it suggests that

the history of the container is more a product of gradual systemic evolution than a

disruptive innovation. Third, it shows that efforts toward intermodal standardization

precede the container-proper.

As discussed previously (Section 3.2.2), efforts to containerize cargo stretched back

to the early twentieth century. They took on renewed vigor after dockworkers had

asserted their dominance over production on the docks (see Chapter 8 for a fuller

account). Though counteracting labor gains was the major motivation for increasing

palletization, palletization possessed other positive attributes that attracted employ-

ers and were later used to praise containers. They reduced breakage, as goods were

prepackaged and therefore handled less and better insulated against shocks. They re-

duced theft, as it was not easy to discreetly abscond with an entire pallet or to open

one up. And they increased the flow rate of goods through the port, as packaging

and sorting was handled elsewhere, relieving dockside congestion.

Tooth (1956, 225) calls palletization “an integral part of the revolution in goods

handling methods which is still taking place in industry generally and which is fun-

45

damentally affecting cargo handling.” The extent of this revolution from scanty use

prior to World War II to the end of the 1950s is conveyed in Table 3.1, which shows

the proportion of palletized cargo at selected East and Gulf Coast ports. While the

range of values reflect the nature of the cargo handled by the different ports, together

they demonstrate that palletization was widespread by 1957 and that it dominated

freight movements in some ports, e.g., Manhattan, where 94 percent of cargo was

palletized.

Port LT (%) MT (%)

New Orleans Commercial 74 65Baltimore Commercial 90 91Hoboken Commercial 36 38Manhattan Commercial 94 80Brooklyn Commercial 17 34New Orleans Army Terminal 66 40Hampton Roads Army Terminal 16 20Brooklyn Army Terminal 41 50

Source: Maritime Cargo Transportation Conference (1957)

Table 3.1: Percent of palletized cargo in long tons (LT = 2,240 lbs) and measurementtons (MT = 40 cubic feet) in selected East and Gulf Coast ports in 1957

Refitting the freight system for palletization laid down some vital technological

components that contributed toward containerization. Hartman and Fish (1957) note

Rotterdam’s paving of its docks while repairing World War II damage specifically to

facilitate the use of forklift trucks. This growing trend contributed to trucks’ abil-

ity to access piers, which is a necessity for containerizaiton. Tooth (1956) reports

that flatbed vehicles for stacking pallets were introduced after the war, a precur-

sor to the chassis used to mount containers today. He also states that warehouses

were being specially built for pallets, e.g., measured in multiples of pallet dimensions

that were subsequently marked out on the floors. This, too, reflected a steady ad-

46

vancement toward the standardization of space and warehouse construction exhibited

today (Bowen 2008). The use of forklifts inside ships also spurred ship redesign (Cargo

Handling 1956b), which became vital for the transition from tankers and freighters

to full-fledged containerships.

Finally, the impetus toward international standardization of pallet sizes was first

recognized and acted upon in relation to pallets, setting an example for contain-

ers. Pallet standardization began in Sweden in 1946, quickly spread throughout

Scandanavia, and was made international by 1952 (Tooth 1956). The importance

of standardizing pallets (and other technological products) lay in overall efficiency

considerations. In the warehouse example above, for instance, standard sizes ensured

that no space was wasted as pallets were brought in and out of the warehouse; there

would always be a space of exactly the right size. It was also vital for forklift manu-

facturing. Standard sizes ensured that a particular forklift would be able to handle a

given pallet and that the forklift itself could be sold to a wider market. Standardiza-

tion was also reflected in truck and ship design. In essence, standardization turned the

pallet into a “module” that could be efficiently plugged into any variety of transport

and storage modes, which often simplified and expanded markets (Sturgeon 2002), as

the container will later (Heins 2009).

Intermodalism

By mid-century signs of containerization’s imminent birth were plentiful. Efforts

toward intermodal transfers between sea and land were stepping up. And pressures

to increase the size of containers in use led to notable intermodal experiments in the

1940s, including the development of the first truly modern containers.

Following the railcar barge model operating in New York Harbor, Seatrain Lines,

Inc. began to transport railcars on specially designed ships across the Caribbean

47

from Belle Chasse, Louisiana to Havana, Cuba. Several years later they expanded

this service to Hoboken, New Jersey. The railroads viewed this service as a direct

competitor, as evidenced by a litany of unsuccessful litigation to stop the service

(Levinson 2006, 53; Palmer and DeGiulio 1989, 287–288). This litigation was initiated

primarily during the early 1940s, when the In re Container Service decision was in

effect. This suggests since the railroads were effectively priced out of the intermodal

market, they were employing legal means of protecting their market by entrenching

modal separation.

This did not stop them from moving back into intermodal operations as soon as

the opportunity arose, however. In 1953, the New Haven Railroad Company filed for

clarification of the ICC’s position regarding TOFC as embodied in the In re Con-

tainer Service decision. The decision handed down established that railroads served

in these joint transportation services as “connecting carriers” rather than shippers,

which maintained a separation of modes acceptable to the ICC (Columbia Law Re-

view 1966). This clarification led to an immediate and rapid expansion of TOFC

service, nearly doubling the number of Class I railroads offering such services in

the eight-month period preceding January 1955. Over the next eight years, TOFC

loadings quadrupled from 168,150 annually to 797,500 (Palmer and DeGiulio 1989,

292–293).

Illustrating perhaps a somewhat myopic view of intermodalism or support for

their own expansion into intermodalism, the railroads do not appear to have opposed

initial efforts by Trailerships, Inc. in 1947 to transport truck trailers on converted

military ships from New York City to Albany. Nor did they oppose similar service

offered by TMT Ferry, Inc. from Florida to Puerto Rico (Palmer and DeGiulio 1989,

288). Together this and the previous example demonstrate that intermodalism was

gaining momentum in the wake of World War II, facilitating shippers’ and shipping

48

companies’ efforts to circumvent growing labor power and costs in the ports.

In addition to the feasibility of transferring large containers, i.e., truck trailers

and box cars, between sea and land, there was a concurrent upward pressure on the

size of containers and pallets already in use. As Tooth (1956, 228) asserts, “It has

already been stated that the employment of pallets is affecting cargo-handling in most

countries and also that there is very little unit load traffic passing through the major

ports of the world. To explain this apparent contradiction, it must be said that in

many countries large pallets, capable of carrying economic loads for modern quay

cranes and ships’ purchases, are being used as dock tools.” The import here is that

ships and docks prefer pallets larger than those in use in factories and by rail (less

than 40x48 inches with 1–2 tons capacity). This, in turn, implies pressure from the

transportation industry toward scaling up cargo unit size by combining a number of

pallets into one unit, a unit that will eventually emerge as the container.

According to Palmer and DeGiulio (1989), Leathem D. Smith conceived of perhaps

the first truly modern container in the early 1940s. Though delayed by World War

II, he drew up blueprints in 1943, found an initial customer in Agwilines, a principal

New York shipper in 1944, and began production of containers in 1945 that were

purchased and then leased out by Safeway Container Corporation. In the case of

Smith v. Dravo Corp. in 1953 for unlawful appropriation of trade secrets, it was

reported that:

He envisioned construction of ships especially designed to carry their cargo

in uniformly sized steel freight containers. These devices (which, it ap-

pears, were the crux of his idea) were: equipped with high doors at one

end; large enough for a man to enter easily; weather and pilfer proof; and

bore collapsible legs, which (1) served to lock them (a) to the deck of

49

the ship by fitting into recesses in the deck, or (b) to each other, when

stacked, by reason of receiving sockets located in the upper four corners

of each container, and (2) allowed sufficient clearance between deck and

container or container and container for the facile insertion of a fork of

a lift tractor, and (3) were equipped with lifting eyelets, which, together

with a specially designed hoist, made possible placement of the containers

upon or removal from a ship, railroad car or truck, while filled with cargo.

By the spring of 1946, Brodin Lines, Grace Lines, Delta Lines, and Stockard joined

Agwilines in leasing his containers, and by March 1948 roughly 600 containers of his

design had been sold, though manufactured by Dravo Corporation after Smith’s death

in a 1946 boating accident.

Levinson (2006) prefers instead to attribute the design of the first container to

Keith Tantlinger, who worked for Brown Industries in Spokane, Washington. “In

1949, he had designed what was probably the first modern shipping container, a

30-foot aluminum box that could be stacked two high on barges operating between

Seattle and Alaska or placed on a chassis and pulled by a truck” (Levinson 2006, 49).

Construction materials appear to be the primary difference, but regardless of which

features one chooses to designate as those essential to the modern container and thus

as a starting point, the essential point here is that by the late 1940s the idea of large,

enclosed containers that could be stacked and transferred among ships, trains, and

trucks was clearly in circulation and gaining momentum.

3.2.4 Early Containerization

Pride of place for combining inland trucking with ocean shipping through container-

ization is generally granted to Malcolm McLean, former president of McLean Trucking

50

Company, most notably in Levinson’s (2006) recent The Box. The narrative so far

questions the originality of his ideas, though not the significance of his active inno-

vation. In fact, McLean may well have been the first individual to conceive of and

operationalize containerization as a system of integrated, multimodal transportation.

While Smith and Tantlinger’s early designs were intended to enable simple intermodal

transfers, they were offered simply as a shipping technology. McLean had sought to

bind the heterogeneous components of trucking, shipping, and loading into a single

system for door-to-door freight delivery, thereby developing a new second-order LTS.

According to Levinson (2006), McLean worked his way up from truck driver in

1934 to owner of one of the largest trucking companies in the U.S. twenty years

later. In 1953, concerned that increasing highway congestion was cutting into his

profits and that war-surplus cargo ships would be purchased by domestic shipping

lines at costs that would allow them undercut his prices, McLean had a brainstorm:

rather than drive on congested roads, load the trailers onto ships and sail around

the congestion. McLean’s technological style reflected his impetuous, strong-minded,

risk-taking character. Having decided to buy his own ships without recourse to careful

financial calculations, he brought Tantlinger from Brown Industries to his company,

Pan-Atlantic, and simply told him to design a system similar to his Seattle designs

that would work with dedicated boats.

McLean’s container service began in 1956, a decade after Smith’s “Safeway” con-

tainers. The system as a whole follows a familiar pattern: cranes lift the containers off

of truck chassis, load them onto ships by depositing them in specially designed frames

in the holds of converted military ships that would carry them to distant ports (orig-

inally East and Gulf Coast ports) to be offloaded by cranes back onto trucks for final

delivery. This basic model, as I have shown, had been developed and experimented

with over the preceding decades. However, this narrative does not intend to question

51

his adventurous entrepreneurial innovation. In the face of institutional obstacles to

owning both a trucking firm and a shipping firm, McLean sold his trucking firm, the

third largest in the US at the time, and gambled everything on his venture. So while

the concepts McLean employed represent a reintegration of existing concepts and is

thus not particularly noteworthy, his role as an innovator in sparking the wholesale

adoption of containerization cannot be disputed.

3.3 Stage Two: Transfer

3.3.1 Domestic

While McLean was shaking up the East Coast, Matson Navigation Company had

noted the potential of McLean’s operation and was moving steadily toward employ-

ing containerization for its routes from the West Coast to the Hawaiian Islands. But

Matson embraced a different technological style. For Matson, established in 1882,

shipping served as an adjunct to its main operations of sugar and oil production, and

it had no drive to move rapidly. Citing the lack of reliable information on the cost

effectiveness of containerized shipping, Matson adopted a more scientific approach

(Levinson 2006, 59–61). The company hired Foster Weldon, a Johns Hopkins Uni-

versity geophysicist and pioneer in operations research, who determined that almost

half of Matson’s existing door-to-door shipping costs was due to labor. “[T]his cost

has increased steadily in the past and will continue to do so indefinitely as long as

the operation remains a manual one. There is certainly no indication of a change in

the current trend of spiraling longshore wages with no corresponding increase in labor

productivity” (Weldon 1958, 652–653). The only solution deemed viable was automa-

tion. After more than two years of painstaking research, cautious experimentation,

52

and financial calculation, Matson finally intiated its container service on 31 August

1958 from San Francisco. By this time the company had begun developing fully ded-

icated ships for the trade, putting them into service by 1960. With Matson’s more

conservative endorsement of a full embrace of containerized intermodal transport on

the West Coast and McLean’s flashy intervention into the East Coast market, the

industry’s shift toward containerization was solidly established.

Opinions about the future success of containerization remained diverse for some

time, however. First, the wide variety of cargo, ships, and loading processes present in

most ports led some observers to abandon any hopes of standardizing dock procedures.

The types of cargoes handled in a port require operations on the whole

gamut of techniques, from the most primitive form of labor to automa-

tion. . . As ships differ considerably with each other and cargoes may differ

also even within the hold of the same ship, the level of mechanization or

automation required for the loading and unloading of a ship may also vary

considerably. . . Moreover, for a given ship and a given cargo the equipment

used may vary accordingly as the cargo is stowed in the bows or a layer

is about to be closed. . . The great diversity due to the ships used, the

area of the ship where cargo is stowed, the type of cargo handled and the

equipment involved make it extremely difficult if not well nigh impossible

to standardize operations. (Picard 1967, 9–11)

Second, some advocates of palletization believed that containers were too restrictive.

“I have never believed in containers. It is my opinion that it should be the exception,

not the rule, that transport concerns should supply containers. Goods should be

packed on or as near as possible to the production lines, and they should be so packed

and reinforced that they can withstand the treatment to which they are subjected by

53

reasonable and modern transport” (Markussen 1962, 9).

Others, however, were completely sold. “In my view on suitable distance services,

the containering of freight is the method of the future. Like most new ideas it is

difficult to start and in its early days is somewhat of a tender plant. In ten years

time I am sure that we shall wonder why we did not start this long ago.”(Sidey 1962,

19) In Cargo Handling, an industry journal oriented toward logistics providers, Crake

(1962, 21, emphasis in original) lists the advantages of containerization:

To the customer, a door to door service.

A decrease in transport costs.

Reduction in loss, damage and pilferage.

Reduction in packaging expenses.

Reduction in warehouse space and inventories.

Simplification of tariffs on the joint road, rail, water rate and under a

single Bill of Lading.

To the shipowner a quicker turnround in port.

An increase in the annual transport capacity per unit.

Maintenance of a given service with less ships.

Decrease in cargo-handling costs.

Decrease in claims for damage and pilferage.

These claims reflect those made in earlier stages of unitization and provide the un-

derlying logic for shippers and shipowners to adopt the new system. The importance

of countering labor costs (“cargo-handling costs”) will be addressed in more detail

in Chapter 8. Financial analyses of bulk, pallet, and container shipping costs in the

same journal added weight to the economic case for moving to containers, e.g., (David

1957).

54

Support from the military also helped tip the scales in containerization’s favor.

The military feared a third large conflict and recognized the need for rapid transport

and turnaround to maintain its effectiveness (cf. Virilio 2006) and the desirability of

achieving this by reducing handling. According to Crake (1962, 20), “The essential

point is to have a domestic system which dovetails readily in terms of standardised

material techniques and know-how.” The military settled on roughly seven-foot steel

cubes called “Conex boxes,” and military orders for these containers provided impetus

for domestic suppliers, which led to their promotion of private sector use by manu-

facturers. Suppliers “have aimed their sales policy at the shipper, the forwarder, the

domestic carrier and the shipping industry” (Crake 1962, 20).

3.3.2 International

Interest in containerization was not limited to the United States. Smaller containers

tailored to European rail gauges and truck chassis were coming into widespread use in

Europe, though they were steadily replaced by larger containers after standardization

agreements in 1966. The Japanese Shipping and Shipbuilding Rationalization Council

urged the Ministry of Transport in 1966 to endorse containerization. After conducting

studies of U.S. ports, new port legislation was approved in August 1967, and container

cranes began operating by the end of the year (Levinson 2006, 186–188). Meanwhile,

regular trade between the U.S. and Asia was given a boost by the military during the

Vietnam War.

According to Levinson (2006), early in the war, logisitics were hopelessly mangled

and supplies were failing to reach U.S. troops in a timely manner. This was due to a

lack of coordination among military divisions and to Vietnam’s shallow coastal waters,

which compelled goods to be offloaded in nets onto barges—sometimes as far as four

55

miles offshore—and ferried in to shore. Military logistics experts advocated increased

use of unitized cargo, which referred during World War II to carrying goods on pallets,

but to Conex boxes by this point. However, Robert McNamara, then Secretary of

Defense, was aware that shipping companies had developed new technologies and

brought a number of executives together for a meeting in late 1965.

The outcome was a series of minor experiments that produced no results other

than demonstrating that the private sector might possess efficiencies that the U.S.

military lacked. Finally, another logistics crisis in late 1966 and a need for increased

supplies compelled the military to contract out container freight transport. McLean’s

Sea-Land won the bidding and immediately transformed Cam Rahn Bay and Da Nang

into massive container ports. Soon after they went into service in the second half of

1967, the cargo backlog was cleared and supply chains smoothed.

However, the ships, laden with military cargo to Vietnam, returned empty on the

return journey. Japan’s rapidly growing export market provided the perfect oppor-

tunity to increase the load factor of each shipping company’s system by increasing

the total cargo transported by each ship on its return journey. Japan lay near the

shipping route, had television, radios, and other goods to ship to the U.S., and, as

mentioned above, was actively embracing containerization. Matson had entered the

trade early to ship Japanese products to California and hoped to supplement this by

supplying military bases in Japan and Korea. Sea-Land was in the opposite position.

With a military contract, it hoped to pick up goods on the way back to the U.S. Mat-

son partnered with Nippon Yusen Kaisha Line (N.Y.K.), and Sea-Land with Mitsui.

A number of Japanese carriers also entered into competition as well. Overcapacity on

the route laid the foundation for a phenomenal growth in transpacific trade (Levinson

2006).

56

3.4 Stage Three: Growth, competition, and

consolidation

3.4.1 Growth: The homogenization of space

While load factor concerns drove growth in the transpacific arena, the greatest ob-

stacle to the growth of containerization was a lack of standard container sizes and

mounting equipment. Each company was employing often uniquely sized containers

to suit their specific needs. Rail systems were incompatible on both sides of the At-

lantic and within the U.S. itself. Because of the inflexible mounting systems installed

in ships to ensure that containers would not shift during transit and threaten to cap-

size the ship, the ships were only capable of carrying containers designed specifically

for them. Thus, if this situation were to persist, each container shipping company

would have to maintain its own fleet of containers, supply them to all its customers,

and operate its own cranes and other equipment in every port of call, a sure show

stopper.

In 1958, the United States Maritime Administration (Marad) decided that stan-

dardization was required. It was motivated primarily by an interest in ensuring that

companies that received its subsidies to build ships did not go bankrupt, leaving

Marad with a ship that was useless to any other container operator. The Navy also

lent its support for this effort, since it feared logistics chaos if it commandeered ships

with unique container systems in the time of a conflict. Thus, standardization arose

in part out of the Navy’s need to be able to quickly assemble a second-order freight

transport system in a time of war. Levinson (2006, 127–149) provides an account

of the twelve year struggle that went into the International Standards Organization

(ISO) full specfication for containers that constituted a major step forward in allowing

57

a container to be packed in Masan, Korea and transported by truck, ship, and rail to

Chicago without being unpacked.

As Sturgeon (2002) suggests, modularity in a value chain can increase efficiency

and reduce costs. Using the example of the electronics industry, he argues that the

ability to codify specifications for transfer from a brand name firm to a contract

manufacturer through the use of such tools as electronic data interchange (EDI) and

computer-aided design (CAD) files based on ISO standards enables better economic

performance. Codification removes firm-specific data transfer standards, thereby at-

tenuating the close bond between design (and marketing) and production. This flex-

ibility is expressed in the lead firm’s ability to work with a number of contract man-

ufacturers and in the contract manufacturer’s ability to work for a number of lead

firms. (See also North [1981, 26] on the role of standardization in reducing transaction

costs.)

The standardization of containers achieved a similar end. When a producer uses

a container that can be transported by most trains, trucks, and ships, the producer is

able to contract out the movement of its goods to any number of logistics providers.

Symmetrically, a logistics provider can be assured that it can handle the products,

whatever they may be, of a wide variety of potential customers. In effect, shipping

companies ship homogenous units of space rather than specific goods. This arrange-

ment not only allows for price competition but also enables companies in rapidly

changing markets to shift their business or acquire business that more directly re-

flects its immediate needs.

58

3.4.2 Competition: Deregulation and overcapacity

The shipping industry has been plagued historically by boom and bust cycles. In times

of increasing trade, cargo space becomes scarce, transport prices rise, and shipping

companies start placing orders for new, expensive, and often larger ships. However,

the long time required to construct a cargo vessel leads to construction backlogs, and

many ships are not built before the global trade again turns downward. This results

in a glut of cargo space, plummeting transport prices, and inevitable bankruptcies

(Levinson 2006, 222, 227).

To combat the destructive elements of this cycle, shipping companies traditionally

formed “conferences” along established shipping routes that set minimum prices for

cargo. The goal was to eliminate race-to-the-bottom strategies of underbidding one’s

competitors and encourage competition on the basis of service instead. They were

also intended to guarantee a share of trade for each member and limit competition.

Sanctions for violations were directed more toward shippers than shipping companies.

Liner companies would refuse to carry the goods of any shipper who employed ser-

vices outside the conference. For larger shippers, this was a very effective deterrent

(Branch 1996; Clarke 1997; Kendall and Buckley 2001). The U.S. Shipping Act of

1916 recognized that this system held some benefits and sought to control rather than

abolish the system. It did so by requiring that conferences justify their rates in terms

of operating costs and be approved by the Federal Maritime Commission (Palmer

and DeGiulio 1989, 317). In effect, liner conferences functioned as legally sanctioned

cartels that dominated shipping over particular geographical areas (cf. Chapter 8).

In a system based on operating costs, the dramatic savings of containerized trans-

port over traditional break bulk transport undermined the conference system from

the beginning. Within a few years of McLean’s first containership, Sidey (1962, 18)

59

asks the rhetorical question, “Is it right that one or two small Lines in a Conference

can stop or delay the bigger and more forward thinking members going forward with

the new development? This has happened in a number of Conferences. The result is

that the development, if any, is stunted and mis-directed and also on many occasions

one is left wondering whether the Conference looks upon container operators as allies

or as people who should be stopped at every possible point.” Much of the struggle

played out in container liner companies’s efforts to shift rate setting away from the

railroad-style setting of rates for each commodity by weight and toward single prices

for single containers, regardless of weight (Levinson 2006, 224–227). In this sense,

the conference system acted as a reverse salient in the development of intermodal

transportation.

The system was substantially weakened during the era of deregulation in the late

1970s and early 1980s. Prevailing opinion in the U.S. was that prices and services

should be set in an open market. Following deregulation of the airlines, trucking,

and railroads under President Jimmy Carter, shipping took its turn with the Ocean

Shipping Act of 1984 (Lewis and Vellenga 2000, 28; Rose et al. 2006). The act

strengthened conferences’ antitrust immunity and shifted the burden of proof in rate

setting from the conferences to the shippers and the FMC. Conference members were

also permitted for the first time to offer through rates, making the 1984 Act the

major step forward toward true intermodalism. Despite these gains, however, the

legislation embodied several clauses that would further undermine the conference

system. First, through rates had to be negotiated by shipping companies individually

with landside transportation providers. Second, conference members were given the

right to independently file a tariff outside of the conference, which essentially gave

stronger players the freedom to opt out. The significance of this clause is evident in the

thousands of actions that were filed very soon after the legislation passed. And finally,

60

to counter the monopoly power of the conferences, shippers were legally authorized

to form associations for bargaining rates with conferences (Lewis and Vellenga 2000,

27–29).3 Deregulation of shipping was more or less completed with the amendment of

the Shipping Act of 1984 by the Ocean Shipping Reform Act of 1998 (OSRA). This

act further undermined the conference system by allowing shipping companies to

enter into confidential contracts with shippers while limiting conferences to toothless

guidelines for member contracting (Federal Maritime Commission 2001; Lewis and

Vellenga 2000). The provisions of this law were also actively taken up, as evidenced

by the 200 percent increase in service contacts and amendments in the space of roughly

a year (Federal Maritime Commission 2001).

Deregulation led to one other major development for intermodalism. Soon after

the passage of the Motor Carriers Act of 1980 and the Staggers Rail Act of 1980

that deregulated these industries, the courts abandoned the ICC’s historical efforts

to prevent cross-ownership of transportation modes (Shashikumar and Schatz 2000;

United States Court of Appeals 1984). As a result, rail companies could now own

trucking firms or shipping firms and vice versa, allowing for the first time truly inter-

modal firms. Firms that chose to exploit this possibility and cross modal boundaries

were able to assemble a second-order transportation system and greatly expand their

geographical reach.

With deregulation and the entrance of major new competitors in containerized

shipping, competition became fiercer than ever. Many of these new carriers flew

Asian flags, reflecting growing U.S. and European trade with Asian manufacturers

(Slack 2004, 26). Perversely, a primary means of competing in a market characterized

3Conferences and the emergence of shippers’ associations express the organizational componentof spatial struggles among factions of capital. This struggle reflects that discussed in chapter 8between labor and capital to establish geographical monopolies. It remains unexamined here, butwill surely be a component of future research.

61

by overcapacity is for the individual carrier to increase capacity. This is because the

per tonne or per TEU shipping costs decrease as ships increase in size (Cullinane and

Khanna 2000, 186). Thus a shipping company can offer lower rates if it uses larger

ships. As a consequence the size of containerships has grown rapidly, particularly

since 1995. Containerships had grown rapidly until the mid-1980s, when they levelled

out at 4,500 TEUs (the equivalent of an average Manhattan block stacked nine or ten

high), the maximum size that would fit through the Panama Canal. From 1995 larger,

so-called “post-Panamax ships” were introduced, which have quickly exceeded 8,000

TEUs and even reached 12,000 TEUs (the equivalent of a 20-story building covering

an average Manhattan block). And though 12,000 TEU ships will probably remain

the largest in operation for some time, 18,000 TEU vessels are now being considered.

These ships require deeper channels and berths than many ports can offer, putting

pressure on ports to undertake massive, expensive dredging projects to accommodate

them (Cullinane and Khanna 2000, 182–184).

3.4.3 Consolidation: Mergers, acquisitions, and alliances

Of course, for such large ships to make a profit at their lower rates, they must fill the

great majority of their container slots. That is, as ships scale up in size, maximizing

load factor takes on increasing importance for shipping companies’ financial viabil-

ity. Recognizing the tendency toward overcapacity and seeking to reduce the risk of

investing in such large ships, carriers faced three options, according to Clarke (1997,

21–23). First, they could enter into consortia agreements so long as total market

share was restrained to roughly one-third. Second, they could form global alliances

that would share resources across the globe. The primary difference between these

two options is that the latter goes beyond pooling and rationalizing ships to sharing

62

equipment, technical standards, and vessel ownership. Finally, the third option would

be to remove the anti-trust immunity of shipping conferences. To these can be added

a fourth, mergers and acquisitions (M&As). In practice, all four have been pursued,

but M&As and alliances have dominated the landscape.

As firms invested in large capacity that they could not sustain, they became targets

of takeovers. According to Slack (2004, 26–27), there were at least ten major M&As

between 1990 and 2000. One particularly striking example is that of CP Ships, a

smaller, Canadian niche carrier, which grew to be one of the top ten container carriers

by acquiring at least six lines after 1995: CAST, Lykes, ANZDL, Contship, TMM,

and Italia (Alix, Slack, and Comtois 1999; Slack 2004). This and other mergers and

acquisitions have concentrated ownership into fewer hands.

In a dynamic relation with these M&As, global alliances formed and reformed

at a rapid rate during the 1990s. First, an alliance between Sea-Land and Maersk

formed in 1991. In 1995, APL, OOCL, MOL, and Nedlloyd formed the Global Al-

liance. In response, NYK, P&O, Hapag Lloyd, and NOL formed the Grand Alliance

and Hanjin, DSR-Senator, and Cho Yang came together in the United Alliance. The

composition of these alliances shifted as M&As took place with companies from com-

peting alliances, notably the acquisition of the U.S. line APL by Singapore’s NOL

and the merger of Britain-based P&O with the Dutch line Nedlloyd. As a result

Nedlloyd left the Global Alliance to join the Grand Alliance, and NOL, APL, and

MOL reformed the Global Alliance as the New World Alliance with HMM (Slack

2004, 25–27). By 1995, John Snow, CEO and chairman of CSX, stated that just

three alliances controlled 70 percent of the market (quoted in Brooks 2000, 169).

63

Maersk and Sea-Land: Slow dance to merger

One alliance that led to a merger is of particular pertinence to this dissertation,

that of Maersk Line and Sea-Land Service Inc. Sea-Land’s origins under under Mal-

colm McLean have already been discussed (Section 3.2.4). Sold by McLean to R.J.

Reynolds in 1969, divested by R.J. Reynolds in 1984, and taken over by CSX, a large

conglomerate with strong holdings in rail, by the late 1980s Sea-Land moved rapidly

into cooperative arrangements at the global scale. According to Brooks’s (2000, 166–

171) authoritative account, these began with joint ventures with French and Italian

companies to serve the Middle East. Taking advantage of the provisions for coopera-

tive working agreements permitted by the Shipping Act of 1984, Sea-Land chartered

five of its new Econships, the world’s largest at the time, to P&O and Nedlloyd with

transatlantic vessel sharing agreements. These agreements and others with Norasia

were undertaken with“the express purpose of improving asset utilization and reducing

costs” (Brooks 2000, 168), i.e., increasing load factor.

Maersk Line, a susidiary of the A.P. Møller Group, is one of the oldest and largest

shipping lines in the world. Well established on the Europe-Far East and trasnpacific

routes, Maersk decided in 1987 to make an aggressive push to complete its global

service by entering the North Atlantic container trade. Though it managed to acquire

ten percent of the transatlantic market within three years, a lack of profitability

drove it to rationalize its transatlantic operations, including negotiations with CSX

for the purchase of Sea-Land. Though negotiations only resulted in a vessel sharing

agreement, “throughout the early 1990s, rumours of the Sea-Land Maersk relationship

being a slow dance to merger circulated in the industry” (Brooks 2000, 169). Even

after CSX CEO and chairman John Snow adamantly rejected these rumors in 1995,

Maersk and Sea-Land announced a global alliance in 1996 when P&OCL left its

64

alliance with Maersk to join Grand Alliance.

This alliance was considered one of the most complete at the time and served both

companies well. Geographically, their routes were generally complementary, with Sea-

Land strong in the Northeast and Maersk in the Southeast. CSX operated U.S. rail

lines that provided a valuable land-bridge across the country. In 1996, Maersk’s

profits rose 23 percent to almost US$350 million with slot utilization (a measure of

load factor) in the high 80s, well above the industry average of 62 percent. Meanwhile,

Sea-Land was expected to cut costs by US$100 million per year by 1998 and as much

as US$150 million by 2000.

By the late 1990s, perhaps in response to the alliance’s success, Snow softened his

stance on a possible sale, saying that as a conglomerate with a portfolio of unrelated

firms, it was willing to sell Sea-Land at the right price (Brooks 2000, 171; Tirschwell

1999). In a press conference in April 1999, Snow sought to quash speculation about

an imminent sale after it announced a three-way split of Sea-Land into separate oper-

ating companies for terminals, international container transportation, and domestic

shipping, a strategy that often presages a sale of one of those components (Watson

1999c). Despite his denials, CSX agreed to sell Sea-Land’s international service to

Maersk mere months later for $800 million, including 18 terminals (Brennan 1999d).

Impacts of M&As and alliances

M&As and alliances have produced three primary changes in the container shipping

industry, which reflect the primary motivations of Hughes’ third stage: rationalizing

the system, making it more efficient, and concentrating capital. Slack, Comtois,

and McCalla (2002) describe two of them. First, alliances have introduced more

uniformity in the industry as they have expanded service across the globe, offering

similar numbers and frequencies of services to a similar number of ports, though the

65

ports themselves often differ radically (Slack 2004, 36). Second, container shipping has

become more intensified as asset pooling has allowed the companies to incorporate a

greater range of ports into their regular services and to build up services, transporting

greater volumes on larger ships.

The third impact has been the exertion of new pressures on ports. These pressures

take two primary forms. First, as discussed above, there were new demands on physi-

cal infrastructure due to the larger ships employed, which continue today. Progress in

ship size is to some extent constrained by limitations in the physical infrastructure of

ports, and this has been identified as a new reverse salient in the growth of the system

overall. The wider, deeper hulls demand costly dredging, the widening of the Suez and

Panama Canals, and even, in one case, the raising of an entire bridge (United States

Army Corps of Engineers New York District 2009). Second, there were competitive

pressures as alliances sought to consolidate their port calls at load centers.

The concept of the load center was originally introduced by Hayuth (1981). Be-

cause of the high daily operating costs of a ship, which increase with size, operators

have two options for reducing time and therefore overall costs in port. The first is to

decrease the turnaround time in port, and the second is to reduce the number of ports

of call. These two approaches cut out port charges and allow the ships to transport

more goods over a given time period. Affiliated with the reduction of the number

of ports of call is the trend toward concentrating traffic along a smaller number of

primary routes. Channeling services on high volume routes contributes greatly to

maximizing load factor and thus profitability. The result is a hub-and-spoke network

in which feeder services transport goods between smaller ports and load centers, while

high volume “expressways” ferry goods back and forth between distant load centers.

Hayuth argued that this establishes a hierarchy among ports, with load centers

dominating subnetworks. By concentrating throughput in one port, this arrangement

66

promises large profits for actors in those ports that become load centers while other

ports lose traffic and income. Alliances and large companies that sought to establish

load centers in the 1990s therefore had a great number of supplicants willing to make

significant sacrifices to obtain the supposed benefits of becoming a load center. Thus,

interport competition increased significantly during this decade of consolidation. One

such example will be explored in Chapter 5. However, as Notteboom (2009, 61) notes,

the pressure toward load centers is stronger at the level of the individual shipping line

or alliance than it is at the port level, since not all lines will choose the same port to

serve as their load center.

The introduction of more uniformity suggests the rationalizing of the freight trans-

portation system, as mergers and alliances reorganize and expand it in their quest

for profits. Increased profits require increased efficiency, and sector consolidation has

been oriented toward increasing the load factor by reducing overcapacity. Load factor

maximization has suggested to a number of shipping companies the use of load centers

and ever larger, ever more expensive ships. Landside, these expenses are reflected in

the higher costs of erecting larger cranes and dredging deeper channels. Both intensify

the use of capital and entail ever larger risks, as the cost of failure mounts.

3.5 Conclusion

This chapter has described the evolution of containerization and intermodalism as

a large technical system that has steadily merged existing first-order systems into

a novel second-order system. The account is at variance with the stages identified

by Hughes (1983) and Joerges (1988) only in suggesting that intermodalism was

not developed de novo as a complete system, like the electicity networks explored

by Hughes, but rather existed conceptually for some time and acquired the various

67

components that comprise it today through gradual accretion.

The first stage of invention, development, and innovation was characterized by

a seamed system. U.S. legislation geared toward combatting the abuses of the rail

barons of the late-nineteenth century legally isolated each mode of transport and

granted the ICC the power to set rates and license routes. Traffic congestion, labor

struggles, and modal competition drove shipping companies and stevedoring compa-

nies to actively pursue mechanization, leading ultimately to McLean and Matson’s

innovative efforts to ship cargo in containers that could be rapidly switched from

ships to trucks, crossing modal boundaries and merging modes into a second-order

system. The benefits of this system were quickly recognized, and the container made

its appearance around the world.

Containerization has grown rapidly down to the present day, though it is unclear

how the recent economic crisis will impact its long term prospects. Rapid growth was

fueled not only by the increase in global trade facilitated by containerization but also

by a perverse pressure to compete through economies of scale. More expensive ships

with greater capacity was the preferred method of reducing shipping costs and com-

peting in a glutted market. This unsustainable contradiction between overcapacity

and continuing growth produced financial strains on firms that soon led to mergers,

acquisitions, and alliances that concentrated ownership, capital, and control in the

shipping industry. These mergers rationalized the transportation system by reducing

overcapacity. Concurrent deregulation allowed firms to cross organizational bound-

aries by acquiring or merging with firms that provided other modes of transport. The

result has been the emergence of a seamless, second-order system that operates on a

much broader geographical scale than previously. The spatial implications of these

changes will be explored in the Chapter 4.

68

Chapter 4

Spatial change

As a result of the rapid intermodal transfer enabled by containerization, the logistics

infrastructure has spatially reorganized itself over the last half century, redirecting

cargo flows across the national landscape. From a port perspective, in the late 1800s

the labor-intensive transfer of individually stowed goods from sea to land and vice

versa generated a major mechanical break in the flow of goods that encouraged ag-

glomeration (Glaeser and Kohlhase 2003). As discussed in the earlier chapter on

sociotechnical change (Chapter 3), containerization has removed the obligation to

break bulk on the docks. Instead, within minutes or hours, goods can continue virtu-

ally uninterrupted toward inland destinations where they can be unpacked, repacked,

and redistributed. These inland destinations are the primary points of impact in this

spatial restructuring.

The following analysis functions in two ways. First, it identifies those areas for

which the importance of logistics-related employment is increasing or decreasing,

which offers some information for industrial and infrastructural planning at both

the local and national level. Second, it establishes an essential claim for Chapter

7. That chapter explores whether or not various sectors of industry are locationally

69

coupled with break points in the transportation infrastructure. This chapter provides

evidence that the transportation infrastructure has spatially reorganized, particularly

in warehousing, setting up the test for whether industry has followed this geographical

change and thus remains coupled to the infrastructure network, or whether industry

has remained in place and thus has decoupled from the wider transportation system.

4.1 Ports

The physical requirements for ports limit the number of suitable locations and have

thus resulted in a relative constancy in port location. Ports require deep channels

and berths to allow the now massive ships to access and anchor alongside docks

that span the ships’ length. These docks are preferably in sheltered harbors with

ample dockside storage space and ready access inland (Frankel 1987, 43–45). Today,

suitable port sites are generally not located upriver, because this adds time and cost

to shipping. These demands have always limited the number of available sites and

encouraged the active development of an even smaller number. With today’s massive

ships, the requirements for dock length and channel and berth depth have increased

enormously, resulting in the winnowing out of some ports, including some formerly

major ports like Philadelphia and Boston. It has also led to the active expansion of

more accessible ports through the construction or redesign of terminals. But it has

not led to the construction of entirely new ports in the U.S.

The concentration of container shipping activity in a select number of ports has

also been influenced by the nature of global trade. The growth of trade between Asia

and North America has shifted the center of gravity from the populous East Coast

to ports along the Pacific Rim (Dicken 2003). This trend has been exacerbated as

ships have outgrown the Panama Canal. Thus, the largest transpacific trade ships

70

are currently unable to pass through the Panama Canal and reach the East Coast,

resulting in more cargo being discharged on the West Coast. Table 4.1 provides the

total container volume handled in each of the three port ranges in which US ports

are located: East Coast North America (ECNA), which includes Canadian ports in

the Maritime Provinces; United States Gulf Coast (USGC); and West Coast North

America (WCNA), which includes Canadian ports in British Columbia and Pacific

Coast ports in Mexico. The emerging dominance of West Coast traffic is evident,

surpassing East Coast volume in the mid-1980s and now exceeding it by roughly

two-thirds.

y1976 y1981 y1986 y1991 y1996 y2001 y2006

WCNA 2184 3746 6183 8819 11667 15967 26636ECNA 3318 4413 5909 6738 9669 11840 16361USGC 371 721 911 1118 1250 1598 2243

Source: Containerisation International Yearbook

Table 4.1: Total North American container traffic by port range (1,000 TEUs)

Figures 4.1, 4.2, and 4.3 further emphasize the dominance of the West Coast ports

in container activity. Each port range exhibits a clear focus of container throughput

in one region. On the East Coast, it is the Port of New York and New Jersey. On

the West Coast, it is the adjacent ports of Long Beach and Los Angeles. And on the

Gulf Coast it is Houston. In 1976 the total container volume handled by the adjacent

ports of Long Beach and Los Angeles combined was nearly as much as that handled

by the Port of New York, while Houston lagged far behind with roughly 200,000

TEUs.1 By 2007, Long Beach and Los Angeles each handled roughly eight million

TEUs, while the Port of New York and New Jersey handled just over five million

1A twenty-foot equivalent (TEU) is the basic unit of measurement for container volume, repre-

senting a 20′ × 8′ × 8 12′

space.

71

Con

tain

ervo

lum

e(m

illion

TE

Us)

1975 1985 1995 2005

01

23

45

67

89

Vancouver

Ensenada

Long Beach

Los Angeles

OaklandSavannahSeattleTacoma

Figure 4.1: Container volume for major West Coast ports by yearSource: Containerisation International Yearbook

TEUs, and Houston approached the volume New York and New Jersey had already

reached in 1976.

It is also worth noting that these four ports far outstrip their nearest competitors.2

New York and New Jersey handle more than twice the volume of Charleston, SC,

and Norfolk, VA; Los Angeles and Long Beach handle more than four times that of

Oakland, Tacoma, Seattle, and Vancouver; and Houston handles roughly eight times

the traffic in New Orleans and Gulfport. These ports are all located in the highest

and densest population centers within each port range, lending weight to the claim

that global liner companies choose to direct their traffic as close to their customer

base as possible.

There are at least three countervailing tendencies to this shift toward the West

2This may suggest that the port system follows a Zipf’s law distribution akin to that of primateand secondary city rank-sizes (Berry, Horton, and Abiodun 1970).

72

Con

tain

ervo

lum

e(m

illion

TE

Us)

1975 1985 1995 2005

01

23

45

6

HalifaxBaltimoreBoston

Charleston

Miami

Norfolk

Philadelphia

Pt Newark

Figure 4.2: Container volume for major East Coast ports by yearSource: Containerisation International Yearbook

Con

tain

ervo

lum

e(m

illion

TE

Us)

1975 1985 1995 2005

01

2

FreeportGalvestonGulfport

Houston

MobileNew Orleans

Figure 4.3: Container volume for Gulf Coast ports by yearSource: Containerisation International Yearbook

73

Coast, however. First, the Panama Canal Authority has granted a contract to widen

and lengthen its locks that will permit the passage of today’s largest ships. It is

expected to be completed in 2014 (Panama Canal Authority 2009). The second is

a distant possible future. If the global warming hypothesis is correct and the Arctic

polar cap melts, there is some possibility of a contemporary (summertime) “North-

west Passage” (Sevunts 2005). The third is shippers’ response to the ten-day 2002

West Coast port lockout. Failed labor negotiations between the International Long-

shoremen’s and Warehousemen’s Union (ILWU) and the Pacific Maritime Association

(PMA) in 2002 led to the shipping organization locking out the union for ten days

before President George W. Bush intervened, invoking the Taft-Hartley Act to force

employers to reopen the docks and workers to return. The impact of this crippling

event has induced a number of large producers, distributors, and retailers to increase

the flexibility of their supply chains by developing routes to both coasts (Hall 2004).

4.2 Rail

The immense sunk costs and difficulty in obtaining continuous, linear rights of way

has fixed rail in place since the early twentieth century. The most important re-

structuring efforts in freight3 have been organizational. The limited range of physical

transformation has focused on increasing clearances to allow for the double-stacking

of containers on railcars and on developing inland intermodal terminals.

3There has been a major movement away from passenger travel, however (Rodrigue 2006).

74

4.3 Warehousing

The most significant spatial change in the logistics infrastructure is the location of

warehousing. Since the advent of containerization, warehousing has increased as an

activity, has become geographically more evenly distributed, and has developed new

concentrations a few hundred kilometers in from the coast and across the middle

of the country. Warehousing here refers to general warehousing and storage (SIC

4225 and NAICS 493110 and 593130). The original SIC grouping was broken into

two components under the NAICS classification. The first of these two types is

that which is generally considered warehousing: “establishments primarily engaged

in operating merchandise warehousing and storage facilities. These establishments

generally handle goods in containers, such as boxes, barrels, and/or drums, using

equipment, such as forklifts, pallets, and racks. They are not specialized in handling

bulk products of any particular type, size, or quantity of goods or products” (U.S.

Dept. of Commerce, Bureau of the Census 2003). The second type refers to those

“establishments primarily engaged in renting or leasing space for self-storage. These

establishments provide secure space (i.e., rooms, compartments, lockers, containers,

or outdoor space) where clients can store and retrieve their goods” (U.S. Dept. of

Commerce, Bureau of the Census 2009).

To illustrate warehousing’s spatial shift, a series of maps and regression data fol-

lows. The maps are of two types. The first set of maps (4.4, 4.5, 4.6, and 4.7) shows

total employment in warehousing for each county. The second set of maps (4.8, 4.9,

4.10, and 4.11) illustrates the county-level concentration of warehousing employment.

They employ the location quotient for warehousing (as reported by U.S. Dept. of

Commerce, Bureau of the Census (1987, 1998, 2007, 2009)), which represents the rel-

ative concentration of warehousing employment in the county relative to the country

75

as a whole. Values greater than one represent greater than average concentration,

and values less than one less than average. A value of two would represent a concen-

tration of twice the national average, and a value of one-half would represent half the

concentration.

76

Fig

ure

4.4:

Map

ofto

tal

emplo

ym

ent

for

public

gener

alw

areh

ousi

ng

in19

74

77

Fig

ure

4.5:

Map

ofto

tal

emplo

ym

ent

for

public

gener

alw

areh

ousi

ng

in19

84

78

Fig

ure

4.6:

Map

ofto

tal

emplo

ym

ent

for

public

gener

alw

areh

ousi

ng

in19

94

79

Fig

ure

4.7:

Map

ofto

tal

emplo

ym

ent

for

public

gener

alw

areh

ousi

ng

in20

07

80

Fig

ure

4.8:

Map

oflo

cati

onquot

ient

for

public

gener

alw

areh

ousi

ng

in19

74

81

Fig

ure

4.9:

Map

oflo

cati

onquot

ient

for

public

gener

alw

areh

ousi

ng

in19

84

82

Fig

ure

4.10

:M

apof

loca

tion

quot

ient

for

public

gener

alw

areh

ousi

ng

in19

94

83

Fig

ure

4.11

:M

apof

loca

tion

quot

ient

for

public

gener

alw

areh

ousi

ng

in20

07

84

4.3.1 Overall increase in warehousing employment

The maps of total employment quite strikingly illustrate the massive overall increase

in warehousing employment. Simply, since darker colors represent higher numbers of

warehousing workers, as the maps grow darker across the decades, they are reflecting

this increase. Table 4.2 shows these figures more precisely. Warehousing employment

has increased exponentially, tripling from roughly 30,000 in 1970 to almost 100,000

in 1999 and then increasing by well over 400 percent to over half a million in the last

decade. Unlike other aspects of freight handling work, warehousing has also increased

as a proportion of total employment, particularly over the last decade. The primary

reason for this 400 percent increase lies in changes within supply chain management.

The final stages of production, including packaging (in boxes or displays) and some-

times assembly, has been shifted to locations as close to the customer base as possible,

since this reduces overall shipping costs and facilitates local customization, which is

increasingly demanded (Bowen 2008). In summary, these maps and the table show

that warehousing employment has increased both in absolute and relative terms.

4.3.2 Deconcentration

Even as total warehousing employment has grown, it has become more evenly dis-

tributed geographically. The growth of inland distribution centers (DCs), for instance,

is well documented (e.g., Notteboom and Rodrigue 2005) and often takes place in

conjunction with the development of inland terminal facilities. In the 1974 maps,

warehousing employment (represented by shaded counties) is scarce and highly scat-

tered, though the location quotient map shows a high concentration upstream along

the Lower Mississippi. By 1984 this concentration has evaporated and warehousing

employment is growing on the east side of the Rocky Mountains and the Sierra Madre

85

Mountains in Colorado and New Mexico and spreading into the Plains states. In the

following decade, warehousing expands geographically in a dramatic fashion. While

continuing to grow along the West Coast and the megaregion that spans the East

coast from Boston to Washington DC, rapid diffusion is evident in the Southeast,

the Midwest, and the Rocky Mountain states. By 2007, most non-mountainous ar-

eas of the US are coated in a veneer of warehousing employment. Clear gaps in the

Northeast mirror the Appalachians and in the West the Rockies. Additionally, there

is a general tapering off as one moves west into the agricultural regions of the Plains.

Over the last three decades warehousing has transformed from a scattered collection

of islands to a fairly even distribution across the US.

1970 1979 1984 1989 1994 1999 2004

-0.0

50.

000.

050.

100.

150.

20

port

1970 1979 1984 1989 1994 1999 2004

-0.0

50.

000.

050.

100.

150.

20

airport

1970 1979 1984 1989 1994 1999 2004

-0.0

50.

000.

050.

100.

150.

20

intermodal

Figure 4.12: General Warehousing and Storage (SIC 4225 and NAICS 493110 and531130): Regression coefficients for distance (in 100km) from closest port, airport,and intermodal terminal by year against the logarithm of employment. NAICS andSIC comparable.Point represents estimate. Wide line indicates the 50 percent confidence interval, and thin line represents the95 percent confidence interval for the estimate. Dotted grey lines indicate industrial classification transitions.Red line indicates trend in estimators.

The results provided by a visual inspection of the map series are reinforced by

regression analysis of the locational determinants of warehousing employment (Table

C.1 and Figure 4.12). The regression results (Table C.1) show three significant trends,

one of them spatial. The non-spatial trends are the increases in the coefficients for

population density and per capita income. Indeed, the size of these coefficients make

them the overwhelming drivers of warehousing employment location. Roughly, a one

percent increase in either population density or income per capita is reflected in a

86

one percent increase in warehousing employment. This finding significantly qualifies

Bowen’s (2008) findings that within Municipal Statistical Areas (MSAs), i.e., large

municipalities, warehousing establishments tend to locate in counties with greater

access to highways and airports.4 While there may be important localized trends

toward locating in suburban and even exurban locations, especially for the largest es-

tablishments, the broader, driving impetus is proximity to the consumer base: denser,

wealthier counties.

The third trend lies in the spatial relation of warehousing employment to ports.

In 1974, for every 100 kilometers a county was from a port, warehousing employment

in that county would increase by almost a tenth of a percent. That is, for every

1,000 kilometers (620 miles) from a port (approximately the distance from New York

City to Detroit), warehousing would increase by 0.7 percent, suggesting an increased

need for warehousing goods in the center of the United States. By 2007, this figure

had dropped effectively to zero (0.2 percent), implying that the geographical relation

between ports and warehousing employment once population density and per capita

income had been accounted for has weakened significantly. This means that relative

to population, warehousing has indeed become more evenly distributed.

4.3.3 Reconcentration

Though the overall trend in warehousing has been toward deconcentration, this move-

ment has been accompanied by an evenly distributed reconcentration, or clustering.

As noted above, the greater initial concentration of warehousing employment inland

in 1974 fades over the next decade. Two forms of concentration then emerge over the

following decades.

4Indeed, if there is a trend in the relation to airports, which is weakly suggested by the upwardtrend in the airport plot in Figure 4.12, then it is a move away from airports.

87

The first type of concentration is a band of warehousing employment paralleling

the coasts two to three hundred kilometers (roughly two hundred miles) inland. This

band is less evident on the West Coast, as the Cascade and Sierra Nevada Mountains

press population closer to the shore. On the East Coast, however, one can identify a

nascent band of concentration, particularly in Eastern Pennsylvania and the Caroli-

nas, in 1984. By 2007 a clear band is evident just east of the Appalachian Mountains.

This band appears in both sets of maps, indicating that warehousing is increasing

both in absolute numbers relative to other counties in the US and in importance to

those counties employment portfolios.

The second form of concentration is the evolution of evenly distributed points of

concentration in the middle of the US. The general vacuum of warehousing employ-

ment in the 1970s and 1980s develops a rather smooth coverage by the current day,

but this layer is punctuated by fairly regular concentrations in the Southeast and

Midwest (once one has accounted for mountainous terrain). This can be seen in the

2007 location quotient map (Figure 4.11).

4.3.4 Summary

Geographical change in the location of warehousing employment represents perhaps

the most important element of the spatial restructuring of the logistics network. It

has expanded not only in volume but also in space. Following a see-saw motion, ware-

housing employment expanded out from its initial concentrations in 1974 to more or

less evenly distribute itself across the US by 1994 and then to develop new concen-

trations by 2007. The most significant aspect of this contemporary reconcentration is

that it occurs along a band paralleling the coast, inland from the dense populations

that line the oceans’ shores and abutting the mountain ranges that lie within.

88

The primary driver of warehousing location is proximity to its customer base,

wealthier and more populated counties. This larger national level trend far outweighs—

and has outweighed for decades—the importance of transportation access. Of course,

this relationship is affected by the naturally high correlation between population den-

sity and infrastructure access, so some degree of dependence cannot be discounted.

Overall, though, it is likely that population densities and the warehousing they attract

lead to investment in infrastructure via a shortage of capacity rather than an excess.

That is, a lack of warehousing facilities in populated areas leads to investment in

warehousing in those areas rather than a glut of warehousing facilities lowering costs

to a level that attracts population (Hirschman 1958).

4.4 Freight trucking

Freight trucking has grown over the last forty years but not in relation to overall

employment. Freight trucking establishments (SIC 4210 and NAICS 484100 and

484200) are primarily engaged in furnishing trucking and transfer services for a wide

variety of commodities, generally palletized and transported in a container or van

trailer, but including specialized freight that because of its size, weight, or shape

requires specialized equipment (U.S. Dept. of Commerce, Bureau of the Census 2007).

Though employment in trucking has increased from under a million in 1970 to just

under 1.5 million in 2007, as a proportion of national employment, it has fallen from

1.65 to 1.26 percent (see Figure 4.2). Thus, while greater volumes of trade have indeed

increased jobs in trucking, the declining relative importance of freight trucking in the

national employment portfolio counters the idea that transportation necessarily grows

faster than other sectors (Cooley 1894).

As the plots in Figure 4.13 and regression in Table C.2 show, the location of freight

89

1970 1979 1984 1989 1994 1999 2004

-0.1

0.0

0.1

0.2

0.3

0.4

port

1970 1979 1984 1989 1994 1999 2004

-0.1

0.0

0.1

0.2

0.3

0.4

airport

1970 1979 1984 1989 1994 1999 2004

-0.1

0.0

0.1

0.2

0.3

0.4

intermodal

Figure 4.13: General freight trucking (SIC 4210 and NAICS 4841// and 4842//): Re-gression coefficients for distance (in 100km) from closest port, airport, and intermodalterminal by year against the logarithm of employment. NAICS and SIC comparable.Point represents estimate. Wide line indicates the 50 percent confidence interval, and thin line represents the95 percent confidence interval for the estimate. Dotted grey lines indicate industrial classification transitions.Red line indicates trend in estimators.

trucking establishments, which includes dispatching centers as well as headquarters, is

driven primarily by population density and per capita income. A one percent increase

in population density will increase employment in general freight trucking by about

0.85 percent, almost a one-to-one relationship. While still demonstrating a strong

relationship to per capita income, however, this has been dropping over time. From

a height of approximately a two percent increase in employment for every $1,000

increase in income, the correlation has dropped to roughly half a percent for every

$1,000. Thus, trucking establishments, like warehousing, locate proximate to their

customer base, but they may be moving toward the poorer counties in those areas,

presumably due to zoning and land costs. At the margin, trucking employment also

locates with a intertemporally consistent relation to ports and airports, though not

with intermodal terminals for transferring goods from rail to either ships or trucks.

For every thousand kilometers away the average county is distant from a port, trucking

employment increases by 0.3 percent. This is virtually constant across the last three

decades. Distance from customs airports demonstrates a similar steady rate over

time but is ten times stronger. Over a thousand kilometers, trucking employment

will increase three percent. However, because airports are distributed throughout

the interior, there are very few counties this far away. Trucking’s relationship to

90

intermodal terminals is indeterminate or nonexistent. Taken together, these results

suggest that trucking has not moved relative to transportation infrastructure during

the period of containerization, though it does tend to locate farther away from ports

and airports once population density is taken into account.

4.5 Conclusion

Since the advent of containerization, the logistics network has spatially reorganized,

altering the flow of cargo so that it “slows down and deposits its load” in territories

other than that hosting its port of entry. While capital intensive components of the

network, specifically ports and railroads, have remained spatially fixed, less capital

intensive aspects, specifically warehousing, overwhelmingly follow their customer base

rather than any particular infrastructural node.

As the section on ports illustrates, the center of gravity for container traffic has

shifted from the East Coast to the West Coast since 1976. This is attributable to two

primary factors: first, growth in trade with East and Southeast Asia has increased

traffic on transpacific routes as they are the most direct; and second, ships have out-

grown the Panama Canal, constricting transpacific trade’s access to the East Coast.

The dominance of a single port location within each port range has increased over

time. Such imbalanced growth has been dependent not only on port characteristics,

like channel depth, but also access to the largest population concentrations within

those ranges.

The series of maps, however, also suggest a development strategy for localities

a few hundred kilometers inland. This band is well situated to host warehousing

establishments—with the constraint that they provide air and highway access (Bowen

2008). To the extent that such establishments offer quality working conditions and

91

function as final, value-adding stages of the production process, they offer direct

economic development opportunities and could potentially attract preceding stages

of the production process as well (cf. Chapter 6).

92

1970

1974

1979

1984

1989

1994

1999

2004

2007

Warehousingtotal

30582

43482

48783

25991

59002

78397

96764

427365

527478

Warehousingpercent

0.06

0.07

0.07

0.03

0.07

0.08

0.09

0.38

0.46

Tru

ckingtotal

915628

1001687

1226113

1139225

1408099

1552343

1350038

1410742

1458068

Tru

ckingpercent

1.65

1.62

1.66

1.48

1.59

1.66

1.26

1.27

1.26

Marineca

rgohandlingtotal

94502

82244

93970

73048

50061

52093

47906

64427

67410

Marineca

rgohandlingpercent

0.17

0.13

0.13

0.09

0.06

0.06

0.04

0.06

0.06

Deepseafreighttotal

23919

17426

22694

22014

16455

11323

13390

11242

11217

Deepseafreightpercent

0.04

0.03

0.03

0.03

0.02

0.01

0.01

0.01

0.01

Freighttransp

ortationarrangem

enttotal

79442

58413

79454

89841

102509

111605

149229

174322

212148

Freighttransp

ortationarrangem

entpercent

0.14

0.09

0.11

0.12

0.12

0.12

0.14

0.16

0.18

Source:County

BusinessPatterns

Tab

le4.

2:U

.S.

emplo

ym

ent

by

indust

ryas

tota

lan

dp

erce

nt

ofU

.S.

tota

lem

plo

ym

ent

93

Chapter 5

Global capital and interport

competition

If it becomes a battle between New York and New Jersey, the

region as a whole will only suffer. We could blow this whole

thing out of the water if we fight over it tooth and nail. (Chris

O. Ward, director of port redevelopment for the Port

Authority of New York and New Jersey quoted in Kannapell

1997)

5.1 Introduction: Territory and terrain

Chapter 4 explored the spatial reorganization of the freight transportation network,

arguing that containerization has facilitated a shift in freight handling activities,

especially warehousing, inland away from seaports. This shift alters the terrain of

economic activity. This dissertation will define terrain as the momentum of capital

flows associated with a particular economic activity over a surface, in this case land

94

and sea. Momentum is considered the quantity of capital passing through a given

space over a given period of time. For our purposes, an increase in the volume or

turnover of capital in a given area corresponds with an increase in momentum. 1 For

example, keeping all else constant, if a business were to move its activities from one

city to another, the momentum would decrease in the first and increase in the second,

altering the flow of capital and hence the terrain.

Terrain is to be distinguished from territory. As described by Jessop et al. (2008),

territory is a form of sociospatial structuration associated with efforts to bind, parcel,

and enclose. The object of such efforts is appropriately left open, but for the purpose

at hand, we will focus simply on flows of capital. Specifically, the administrative

boundaries of local, municipal, state, and federal governments function as negotiated

boundaries with attendant mechanisms for controlling flows of capital. For example,

municipal boundaries define those properties from which a given governing body can

extract taxes for the provision of services. Customs laws may block the flow of

endangered species parts past national boundaries into a given country. And most

importantly for our case here, efforts to attract a business to a given location are

attempts to tie the flow of profits to that location through business and sales taxes

as well as jobs and indirect spending.

Territory—at least administrative territory—is much less flexible than terrain.

Long historical processes and organizational inertia tend to keep administrative terri-

tory fairly fixed, while the mobility of capital and people facilitates much more rapid

shifts in terrain. Because administrative boundaries have remained relatively con-

stant over the past fifty years, technological change has allowed logistics providers to

alter the economic terrain underlying governmental territories. As shown in Chapters

1This approach is similar to the economic concept of the velocity of money or circulation exceptthat it explicitly introduces a spatial component that remains implicit in the economic definition.

95

3 and 4, logistics providers are increasingly operating at the continental scale, which

implies that no single administrative territory can guarantee capture of the flow of

capital they generate. As a result, smaller territories are compelled to compete for

those flows. The results of these struggles, by fixing investment in infrastructure

and plant, to varying degrees and for varying time periods spatially fix the terrain of

economic activity (Harvey 1982).

This chapter explores how territorial struggles over terrain played out in one in-

stance of interport competition. Because of the high costs of port infrastructure, such

episodes of restructuring typically arise at intervals of decades and thus structure

economic activity for the 25–30 years contracts usually last. When long term con-

tracts between Sea-Land and the Port Authority of New York and New Jersey were

approaching their end, Sea-Land and it alliance partner Maersk used the newfound

flexibility of the logistics network and their stated intention of developing an East

Coast load center to initiate a cycle of bidding that has resulted in a transfer of funds

from the region’s residents and visitors to the shipping companies.

The case study demonstrates four points. First, the broader territorial range

of the shipping companies gives them greater control over the terrain of economic

activity than port authorities and labor unions. Second, restructuring the freight

transportation system requires extensive changes to the physical infrastructure across

a complex array of interrelated modes. Third, interport competition burdens port

authorities with as much risk as possible, while specific fractions of capital or labor

benefit. And fourth, the associative approach to port governance embodies inherent

weaknesses vis-a-vis hierarchical approaches.

96

5.2 Port authorities

5.2.1 Characteristics

The concept of a “port authority” has expanded from its narrow origins in the Pro-

gressive era as a public authority insulated from outside political influence to include

any public seaport agency. The term does not refer to ports that are privately owned,

even if these privately owned facilities offer the same range of services to the shipping

public as public facilities do. It does however refer to government entities that man-

age more than one port. Beyond this, the range of institutional forms, administrative

forms, powers, and financing are myriad.

Institutional forms

One institutional form is an authority proper: a quasi-independent, self-sustaining

public agency with private characteristics (Fainstein 2005), which is represented in

this analysis by the Port Authority of New York and New Jersey. This form derives

from the Progressive era “suspicion of concerted political power and a renewed sense

of upper-class business virtue” (Walsh 1978, 26), which prescribed agencies insulated

from the vagaries of shifting political power so that they might apply rigorous pro-

fessional techniques to the solution of public problems (Brown 2008, 4). They are

expected to harness the best features of private enterprise–prudence, efficiency, and

economy–for the public interest (Caro 1975, 16). As numerous accounts have attested,

however, complete separation of politics from authorities’ operations is not possible,

though some authorities have been more successful than others in minimizing political

influence (See, for example: Caro [1975], Doig [2001], and Walsh [1978]). Fainstein

(2005, 90) elaborates on the practical relation between authorities and politicians:

97

The institutions act largely according to the same norms that drive the

private sector, even if their staffing is more politicized and they are po-

tentially more susceptible to popular pressure. Although they are under

the control of elected officials, their goals largely coincide with the goals of

those politicians, who thus usually do not feel the need to exercise much

supervision. The motives of politicians and public authorities may differ

in that the primary concern of politicians is reelection rather than the

creation of revenue producers, especially when the revenue stream goes

elsewhere. But the construction of visible projects serves their common

interests.

However, to the degree that authorities do enjoy some isolation from the passing

whims of politicians and their constituents, their independence most often comes at

the cost of having to finance their projects by raising capital on bond markets and

other investments. There are two broad categories of municipal bonds that author-

ities issue (Walsh 1978, 55–83). The first are general obligation bonds, which are

backed by the taxing power of their host municipalities and are thus cheaper for the

authority because of the perceived guarantee of government backing. The second are

revenue bonds, which are backed only by the revenue streams of the authority in

question. Most authorities have access only to the latter type. As a consequence,

such authorities must operate as profitable enterprises if they are to secure future

funds through bond issuances. As Doig (2001, 71) points out, “Port Authority bonds

could be sold only if investors believed the agency would be able to operate the new

enterpriseson a break-even basis with a slight surplus.” This reliance leads politicians

to appoint authority board members predominantly from the business community,

particularly the finance community. And, as Walsh (1978, 204) suggests, “Dominance

98

by business is dominance by business politics.” The necessity of showing a positive

revenue stream thus binds authority interests to those of the financial community and

constrains authorities’ ability to take on loss-generating projects, even if they are in

the public interest.2

Other port authorities are not true authorities. Rather, they are “integral admin-

istrative divisions of state, county or municipal government” (Sherman 2002, 3). In

this analysis, this form is represented by the Maryland Port Administration. These

port authorities function as more or less independent branches of formal government.

Therefore, they are more subject to political pressure. If Fainstein is correct that the

interests of port authorities generally align with politicians’ interests, then this may

not hinder such agencies’ decision-making. It does generally give these authorities

access to much deeper funding sources. Not only are their budgets part of a larger

entity’s, which allows for greater levels of direct funding, but also as government agen-

cies they can issue general obligation bonds, accessing cheaper credit. Other forms

of port authority include special purpose subdivisions and shells to provide bonding

authority for a port facility financing.

Adminstrative forms

Port authorities cover a range of administrative forms: from landlord ports that

contract out all operations to operating ports that maintain ownership of all port

facilities and directly operate them (World Bank 2007). In either sense, port authori-

ties function as the formal governance mechanism for coordinating port functions and

establishing and enforcing port regulations.

Based on a small sample of port authorities (60), Van der Lugt and De Langen

2One prime example of this is the Port Authority of New York and New Jersey’s long fight tokeep PATH, New Jersey’s loss-making commuter rail, out of its portfolio (Doig 2001, 383).

99

(2007) identify three types of authority goals: one oriented to the broadly defined

success of the port cluster, one oriented to profit maximization, and one a mixture

of the two. These essentially embody public, private, and public-private goals and

reflect sources of financing. While all port authorities collect users fees, their level

of dependence on these for operations and future developments varies. More pub-

licly focused authorities have greater access to government funds and more private

authorities depend on market mechanisms for provision.

Powers

While port authorities may be responsible to government officials at the local, county,

district, state, multi-state, or (in Canada) federal levels, they are generally given

expansive, though pared down, government powers. According to Sherman (2002, 3):

Port authorities are typically empowered to exercise the powers of eminent

domain, to conduct studies and develop plans, levy facilities charges, issue

bonds, to sue and be sued, to apply for federal grants, to act as the local

public assure for federal navigation projects, to enter into contracts and

agreements, and , frequently, as the Massachusetts statute states,“to do all

acts and things necessary or convenient to carry out the powers expressly

granted in this act.” (emphasis in original)

While some authorities are narrowly restricted to shipping concerns alone, others are

permitted to operate airports, commuter rail systems, bridges, tunnels, free trade

zones, and a variety of other pursuits. If one pairs the governmental, territorial

monopoly over shipping facilities (and other facilities) granted to port authorities

with the regional viewpoint these large infrastructures naturally engender, it is easy

to understand why Doig (2001) calls the Port Authority of New York and New Jer-

100

sey (PANYNJ) the closest thing the New York metropolitan region has to regional

government.

Port strategy

Given the rapid changes in logistics described in Chapters 3 and 4, there has been

a great deal of debate over effective port strategy and an emphasis on port reform.

In these discussions, many writers conflate the wide variety of port stakeholders with

the port authorities, failing to distinguish the potential variance in goals (Fleming

and Baird 1999; Van der Lugt and De Langen 2007). While the orientation toward

profit maximization does tend to align the interests of port authorities with some port

stakeholders, notably firms, the frequent inclusion of public goals indicates that port

authorities may have additional or other goals. Thus, it is essential to distinguish

port authorities from other stakeholders as a unit of analysis. That said, the main

emphasis of port reform has been a movement away from goals other than commercial

success. Proposals for port authority reform generally range from outright privatiza-

tion through corporatization to commercialization. Privatization involves selling the

port facilities to a private entity. Corporatization is essentially privatization with the

government retaining a large share of the resulting company. And commercialization

indicates an increasing profit orientation.

This trend toward commercialization has tended to shunt port authorities out

of direct port operations and into the status of landlord ports (Ircha 2001, 132–

133). This has led to an investigation into the repurposing of port authorities, since

their profit-orientation demands that they define their value-adding activities. While

loss of cargo has led some port authorities to enter entirely new economic activities

(Brown 2008), there is an increasing consensus that the coordination role of freight-

oriented port authorities fits this bill. Authors like Chlomoudis et al. (2003) have

101

argued that port authorities emphasize their government-like roles in setting targets

for cooperation among stakeholders, directing port development by defining the op-

erational framework for regional port production, and forecasting. These reflect some

of the economic arguments for planning in a market society delineated by Klosterman

(2003), respectively: resolution of prisoner’s dilemma questions, provision of public

goods, and dissemination of information necessary for informed market choices.

Since this coordinating role cannot be limited to the immediate port cluster, we

must adopt the thinner conception of port clusters as incorporating related but ge-

ographically dispersed actors—both public and private (cf. Porter 1998; Whitford

and Potter 2007; Zeitlin 2005). The trend toward port regionalization implies such

a geographical transformation. Port authorities, Notteboom and Rodrigue (2005,

307) claim, must increasingly look to form associative relations with entities outside

their territorial jurisdiction to enhance their competitiveness. They list several major

areas of possible cooperation between port authorities and inland distribution cen-

ters, including traffic management, site issuing, hinterland connections and services,

environmental protection, marketing, and research and development. For example

the PANYNJ has spearheaded efforts to share market information like that in the

introduction, has begun to collaborate with ports and governments throughout the

Northeast Corridor, and has even signed a memorandum of understanding with the

Panama Canal Authority. This example indicates two things: first, that port authori-

ties are also beginning to head up efforts to develop translocal networks to coordinate

the provision of translocal public goods; and second, that the region itself may still

be too small a geographical unit for translocal port planning.

102

5.2.2 U.S. port authorities

Sherman (2002, 1–4) claims that roughly 115 of the 183 deepdraft ports in the United

States are currently governed by port authorities. Unlike many countries, the U.S.

has no national port authority. While some port activities are subject to federal

law and jurisdiction, like security and interstate commerce, the federal government

has no power over state or locally legislated port authorities. This is because the

U.S. Consititution grants the federal government power over the country’s navigable

waters, including channels and harbors, but not over the land adjacent to those waters.

That land is governed by local municipalities and states. Thus, the U.S. port system

is not really a system at all but an amalgam of independent, competing port systems.

5.2.3 Canadian port authorities

The Canadian port system underwent a significant transformation during the height

of the competition discussed below. With the National Harbours Act of 1936, the

Canadian government assumed direct administrative control of fifteen major, finan-

cially troubled ports. The Canada ports act of 1983 relaxed the central government’s

control over local boards without completely devolving management. The central

government also managed two other groups of ports. There were nine harbour com-

mission ports established through other parliamentary acts. And there were more

than 500 other small ports administered directly by Transport Canada (Sherman

2002, 6).

Under the belief that central control of the ports hindered the system’s adapta-

tion to rapidly changing commercial circumstances and political appointments were

leading to poor decisions, the government began moving toward liberalization of the

marine sector in the early 1990s. The Canada Marine Act passed in June 1998 shifted

103

major ports toward commercialization by dissolving the Canada Ports Corporation

and designating the major ports as Canada Port Authorities (CPAs). CPAs are

non-profit corporations directed by a board of user representatives selected by the

Transport Minister, one municipal representative, one provincial representative, and

one federal representative. In addition to creating the CPAs, the Canada Marine

Act began the process of transferring remaining ports to municipal, provincial, and

private interests (Ircha (2001, 133–134); Sherman (2002, 6–8)).

A CPA is“an agent of the Crown for port activities related to shipping, navigation,

the transportation of passengers and goods and the storage of goods to the extent that

these are specified in the letters patent,” according to the law. They are forbidden

from engaging in activities not directly related to port operations. They are required

to be self-sufficient, and like public authorities in the U.S., they must turn to the

private sector for borrowing and are without recourse to government insurance for

repayment (Sherman 2002, 6–8).

5.3 Three ports

The following sections describe the basic characteristics of the three port finalists in

the competition for a unified Maersk and Sea-Land terminal on the East Coast that

will be described below: Baltimore, Halifax, and New York and New Jersey. Each

section provides background material relevant to the broader narrative and analysis.

This information includes details on cargo handled and port facilities. It also identifies

the pertinent outlines of each port’s governance structure in line with the discussion

above in Section 5.2.

104

5.3.1 Baltimore

The contemporary history of the Port of Baltimore begins with the foundation of

the Maryland Ports Authority (MPA) in 1956. Prior to this, the port had been

a “railroad port,” serving the interests of the railroads rather than the carriers or

shippers. Port operations discriminated against the rapidly growing trucking interests

with which the railroads that had long served the port were competing. Storage and

handling rates were higher. Railroads would keep berth assignments secret until the

last minute. Trucks had difficulty physically reaching the piers due to the tracks.

Worse, the piers were falling into serious disrepair (Keith 2005, 13–15). As was

the case in the formation of many port authorities (Sherman 2002, 2), Baltimore-

based commercial interests convinced the Maryland General Assembly to create the

Maryland Port Authority to counter the railroads’ power with the explicit goal of

serving local shippers in the local port cluster (Hall 2003, 354), the owners of cargo

rather than the carriers of cargo. The MPA moved quickly to transform the port to

handle truck-borne cargo. In 1959, it purchased Harbor Field, an airport six miles

east of the city, converting it to Dundalk Marine Terminal. In 1964, the MPA took

a forty-year lease on B&O’s Locust Point piers just north and west of Dundalk (Hall

2003; Keith 2005).

In 1971, the MPA was brought under the control of Maryland’s Department

of Transportation (DOT) and renamed the Maryland Port Administration (MPA)

(Keith 2005, 15). Since this time it has been one of five intermodal agencies integrated

through the DOT. Its budget and political support for large capital expenditures

now depend on the Secretary of Transportation and the Maryland State Legislature,

which have generally been sympathetic (Hall 2003, 355). Though still reflecting a

hierarchical approach to transportation infrastructure management, Sherman (2002,

105

3) describes the MPA and the Virginia Port Authority as “quasi-independent cor-

porate enclaves within their respective state departments of Transportation.” As a

consequence of this institutional change, the MPA’s bonds have changed from rev-

enue bonds that are paid back on the financial performance of the authority to general

obligation bonds that are backed by the state’s ability to raise funds through gasoline

taxes, tolls, and other transportation-based sources of revenue (Hall 2003, 355). Thus

funds are now more constrained by political interests but have potentially greater

depth.

As a shippers’ port, the Port of Baltimore has operated as a common use port.

While many port authorities increasingly act as landlord ports that rent terminal

facilities to single users for extended periods of time, often 25 years or more, the Port

of Baltimore has stuck with the model of common use in which ships are assigned to

berths on a first come, first served basis. According to Hall (2003, 355), this is because

the port already contains a number of private terminal facilities in addition to the

public facilities and opened its own container terminal at Dundalk in response to Sea-

Land’s construction of a private terminal. As a common use port administrator, the

MPA leases space through long and short term contracts to operators who actually

handle the cargo, but it also is involved on a day-to-day basis in allocating berths

and facilitating meetings between shipping companies when there are conflicts.

The Port of Baltimore is located eight and a half hours steaming time up the

Chesapeake Bay from the Atlantic Ocean. As the closest port to the Midwest, Bal-

timore long enjoyed a mileage advantage for rail service. The Shipping Act of 1984

altered this advantage by allowing shipping companies to offer joint tariffs for sea

and land transport. This advantage was also weakened as operating costs increased

with ship size, making the two day round trip less cost effective than rail or truck

transport from ports like Norfolk and Charleston that are right on the ocean (Hall

106

(2003, 351); Keith (2005, 15–16)). Still, the port boasts ready access to I-95, other

interstate highways, and the fourth largest consumer market in the U.S.

Despite dredging its main channels in 1991 from 42 to 50 feet, a depth sufficient

for the largest container ships, the Port of Baltimore has failed to recover its peak

container flow in 1984, when volume reached nearly three-quarters of a million TEUs

(Hall 2003, 351) (See also Figure 5.1). At the time of the Maersk deal, volume had

levelled off at just under a half million TEUs annually. However, the Port of Baltimore

has generally pursued a strategy announced in its 1996 Strategic Plan that focuses on

non-containerized cargo. The mainstays of this strategy are automobiles and Ro-Ro

(heavy equipment that can roll on and roll off the vessel under its own power, such

as tractors and bulldozers). In 1999, the port was the fifth-largest automobile port

in the U.S., and the largest export port (Hall 2003, 352). By 2003, it ranked second

only to New York and New Jersey for automobile imports and remained the largest

export port (Keith 2005, 17). As of 2003, the Port of Baltimore handled 42 percent

of all Ro-Ro equipment. It is also the leads the nation in forest products (Keith 2005,

17).

The port’s public facilities include: Dundalk Marine Terminal, Fairfield Automo-

bile Terminals Intermodal Container Transfer Facility, North Locust Point Marine

Terminal, Seagirt Marine Terminal, South Locust Point Marine Terminal, and Cruise

Maryland Terminal. The two main container terminals are Dundalk and Seagirt.

The latter opened in 1990 and covers 284 acres (112 ha) with a capacity of 450,000

containers per year. Dundalk is twice the size at 570 acres (230.8 ha) and handles

the Port of Baltimore’s full complement of cargo. Dundalk was the proposed site

for the Maersk-Sea-Land terminal and has been hosted a private terminal owned by

APM Terminals, Inc., a subsidiary of the AP Moller-Maersk conglomerate, since 1993.

Though it is now served by both CSX and Norfolk-Southern rail services, at the time

107

of the Maersk deal, CSX provided the only direct rail access (Brennan 1999e).

Baltimore

1975 1985 1995 2005

01

Halifax

1975 1985 1995 20050

1

New York and New Jersey

1975 1985 1995 2005

01

23

45

6

Figure 5.1: Annual container volume in millions of TEUs for three finalistsSource: Containerisation International Yearbook

5.3.2 Halifax

The Port of Halifax is located on the east coast of Nova Scotia and acts as Canada’s

“Atlantic Gateway.” The harbor enjoys a natural endowment of channel and berth

depths of 60 feet (16+ meters) at low tide, the deepest on the east coast of the

continent and sufficient for the largest of todays ships. It is ice-free year-round.

And it is the closest east coast Atlantic port to Europe by roughly one day’s sailing

compared to New York and New Jersey (Halifax Port Authority 2010; World Port

Source 2010).

The port is served by Canada National Railway (CN), which offers on-dock double

stack service. In another sign of the intermodal turmoil in North America during the

1990s, CN was privatized in 1995 with the provisions that no shareholder could own

more than 15 percent of the company and that its headquarters remain in Montreal,

ensuring that the company remained Canadian. Aggressive rationalization pared

down CN’s network to a primary east-west line stretching from the Port of Halifax

to the ports of Vancouver and Prince Rupert. The company also worked to establish

108

a north-south network by purchasing Illinois Central Railroad in 1998, which gave

it critical access to the major rail hubs of Chicago, St. Louis, and Kansas City and

south to New Orleans (CN 2010; Murray 2004).

The Port of Halifax serves as a major port for bulk cargo, primarily forest products,

rubber, and steel, handling almost 9m metric tonnes in 2004. In addition, it offers a

range of other terminals, including one of North America’s largest vehicle processing

and trasshipment facilities, ro-ro, and two dedicated container terminals that handle

roughly half a million TEUs annually. Fairview Cove Container Terminal, which sits

on 70 acres and offers over 2,000 linear feet of dock, is currently leased to Cerescorp,

which is the Canadian incarnation of Ceres Terminals, a stand-alone company in the

Harbour Division of the NYK Group. The second container terminal is the South

End Container Terminal, which is roughly the same size as Fairview. It is operated

by Halterm Container Terminal Limited, which is an independent operator offering

common-user facilities (Ceres 2010; Halifax Port Authority 2010; Halterm Limited

2010).

The Port of Halifax was one of the initial set of ports to implement the adminis-

trative changes demanded by the Canada Marine Act of 1998 (Halifax Port Authority

2010), receiving its letters patent on 1 March 1999 (Sherman 2002). As a Canada

Ports Authority (CPA), Halifax is considered vital to domestic and international trade

by the federal government, and it operates as a non-profit Agent of the Crown that is

responsible for managing the port without financial recourse to the public sector. The

Halifax Port Authority, like all CPAs, is forbidden from investing in non-port activi-

ties (Sherman 2002, 6–8). The Port of Halifax’s mandate is “to develop, market and

manage its assets in order to foster and promote trade and transportation and serve

as a catalyst for the local, regional and national economies” (Halifax Port Authority

2010), and it is under the hierarchical control ultimately of Trasnport Canada. As

109

the transition from a CPC port with operations clearly oriented toward the public

interest into a CPA with a strict self-sustenance constraint took place over the period

of the competition described below, the Port of Halifax can be considered a mix of

public and private interest.

5.3.3 New York and New Jersey

The Port Authority of New York and New Jersey operates the largest port on the

East Coast. Located at the heart of the New York City metropolitan region, the port

is “the gateway to the most concentrated and affluent consumer market in the world”

(Port Authority of New York and New Jersey 2010). It handles all forms of cargo,

but it is particularly noted for containerized cargo. Indeed, as Chapter 3 pointed out,

the PANYNJ was home to the first container ships and terminals.

With the introduction of containerization, shipping activity moved rapidly away

from the Manhattan waterfront to New Jersey, Brooklyn, and Staten Island. The

Brooklyn terminals handle primarily breakbulk cargo, though the South Brooklyn

Marine Terminal handles some ro-ro and the Red Hook Container Terminal han-

dles ro-ro and containers (Port Authority of New York and New Jersey 2010). The

Howland Hook Terminal on Staten Island, reopened in 1996 after almost a decade

of dormancy, has a container handling capacity of a million TEUs (Rodrigue 2004).

The main container terminals, however, are located in New Jersey, where there is

direct access to interstate highways and rail networks (CSX, Norfolk Southern, and

Canadian Pacific) that stretch across the continent. Three large terminals in Newark

and Elizabeth handle about four million TEUs per year. The Port Newark Container

Terminal covers 180 acres (71ha) and offers 4,400 feet (1,165m) of ship berth at a

depth of 40 to 50 feet (12.2–15.2m). Maher Terminal, the largest terminal operator

110

in the port, covers 445 acres (180ha) and offers over 10,000 feet (3,000m) of dock at

a depth of 45–50 feet (13.7–15.2m). Finally APM Terminal, a subidiary of the AP

Møller-Maersk Group, covers 350 acres (142ha) and offers about 6,000 feet (1,829m)

of ship berth at a depth of 45–50 feet (13.7–15.2m) (Port Authority of New York

and New Jersey 2010). The current configuration is a result of major restructuring

following the competition discussed in this chapter.

Since the 1990s, channel and berth depths have been a challenge to the growth of

the port. While Panamax container ships could sail through a 35-foot (10m) channel,

the post-Panamax ships introduced in the mid-1990s required depths of 42–52 feet

(13–16m). Prior to the deal discussed here, the maximum clearance of 40 feet (12m)

in the PANYNJ was more than adequate for Panamax ships, but woefully inadequate

for the newest generation. When the Regina Maersk, the first post-Panamax ship,

called in New York in July 1998, it could not have reached the dock if it had not

discharged a significant portion of its cargo in Halifax prior to arriving (Rodrigue

2004, 73–74). As part of the PANYNJ’s contract with Maersk and Sea-Land and as a

growth strategy, in 1999 the Army Corps of Engineers (ACE) began dredging the Kill

van Kull channel, which transits across the north shore of Staten Island and leads to

Newark Bay. The dredging cost $700 million, 35 percent of which was borne by the

PANYNJ. Even before this work was completed in 2005, the PANYNJ was envisioning

the key channels being dredged to 50 feet (15.2m). It agreed with the ACE to pay

50 percent of the estimated $1.8 billion cost and expects this work to be completed

by 2014 to coincide with the anticipated completion of the Panama Canal expansion

(Port Authority of New York and New Jersey (2010); Rodrigue (2004, 74–75)).

The Port Authority of New York and New Jersey was created by the state legis-

latures of New Jersey and New York in 1921 to rationalize regional transportation in

the face of the railroads’ inability to coordinate services and boost the regional econ-

111

omy (Doig 2001; Sherman 2002). It thus represents an associative approach to port

governance in which the two states freely entered into a contract for the management

of resources perceived to be common. The purpose of the Port Authority is presented

obliquely in the introduction to the Port Compact of 1921. Arguing that the growth

of commercial activity around the harbor has integrated the “territory in and around

the port” into “one center or district,” the Compact states its signatories’ belief that

“better coordination of the terminal, transportation, and other facilities of commerce”

would generate “great economies” that would benefit the two states and the nation.

It then states that cooperation between New York and New Jersey is required to en-

courage investment on the necessary scale and to formulate and execute the necessary

physical plans. The essence of this argument is that commercial activity had scaled

up geographically to the regional level and required a new mechanism to coordinate

its infrastructure. In the language introduced at the beginning of this chapter, the

PANYNJ established a new territory in the face of a changing economic terrain.

Interestingly, the emphasis on national growth has been lost and the states them-

selves have disappeared in the PANYNJ’s orientation as it has grown. Its current

mission statement is “To identify and meet the critical transportation infrastructure

needs of the bistate region’s businesses, residents and visitors: providing the highest

quality, most efficient transportation and port commerce facilities and services that

move people and goods within the region, provide access to the rest of the nation and

to the world, and strengthen the economic competitiveness of the New York/New Jer-

sey metropolitan region” (Port Authority of New York & New Jersey 2006). Though

the original emphasis on state and national benefits may have responded to the ne-

cessity of state legislative and Congressional approval (Port of New York Authority,

Development and Operations Department 1938), in the new mission statement the

nation has become an adjunct to the region and the state’s role merely organizational.

112

This may reflect the need of Cohen and other early actors to persuade state and na-

tional policians to support this somewhat novel form of governance. It may also

reflect the emergence of a belief in the regionalization of the global economy (Knox

1995; Scott 2001; Storper 1997). In either case, the current position runs counter to

the position supported in the current analysis, which is that ports serve national or

continental interests rather than regional interests.

The PANYNJ covers a region of 1500 square miles (3880 km2) in roughly a 25-

mile radius circle around the Statue of Liberty, which stands in the middle of New

York Harbor. It includes New York City as well as Newark, Jersey City, and over 300

smaller cities and towns. In addition to the port itself, the PANYNJ manages a broad

portfolio of facilities, including all bridge and tunnel crossings between New York City

and New Jersey, three major airports, the PATH heavy rail commuter line, the Port

Authority Bus Terminal, the World Trade Center site, and a number of development

initiatives (Doig (2001, 69); Rodrigue (2004, 67–69)).

The governors of New Jersey and New York each appoint six of the twelve members

of the PANYNJ’s Board of Commissioners, subject to state senate approval. They, in

turn, appoint an executive director to take responsibility for managing the PANYNJ

in keeping with the policies established by the Board. The governors retain the right

to veto actions taken by Commissioners of their respective states (Port Authority of

New York and New Jersey 2010). In practice, however, such vetoes rarely occur. And

more generally, the board itself, comprised of individuals for whom the position is

usually a side interest, rarely opposes the suggestions of the Port Authority staff and

executive director (Walsh 1978, 180–181).

113

PANYNJ Financials

The PANYNJ must raise all its funds through fees and the private market, primarly

revenue bonds. To this end, according to Walsh (1978, 96), “the management of the

Port Authority has contantly invoked investors as its most important constituency.”

To give the PANYNJ more secure financial footing during its financially shaky early

days, the state legislatures granted it the power to build and operate bridges and

tunnels between the two states. The steady stream of toll revenue from these facilities

and its refusal to break down its operating expenses under a policy of “consolidated

financing” allowed the PANYNJ to disguise loss-making activities and present overall

positive returns, easing its way into the pockets of potential investors (Walsh 1978).

Table 5.1 illustrates the effects of “consolidated financing.” The policy itself has

concealed the details of individual operations through much of the PANYNJ’s his-

tory. Exceptions were made when the New York State comptroller’s office conducted

detailed audits of the Port Authority’s finances in 1970 and 1974 (Walsh 1978, 91).

The PANYNJ began publishing more detailed financial reports in 1991, and these

are represented here. There are four observations to make at this juncture. First,

total operating revenue has remained positive since at least 1969. Second, there are

two major generators of positive cash flow: the airports and the bridges and tun-

nels. Third, there are two major sources of negative cash flow: PATH and the port

terminals. Finally, the only terminals that have done slightly better than breaking

even are the Auto Marine Terminal and the Elizabeth Marine Terminal. The latter

is particularly significant since, as mentioned above, it is home to the APM terminal,

which is the subject of this chapter’s study. Because it shares the port facilities with

Maher’s terminal, it is impossible to determine whether or not one is subsidizing the

other, nor is there any indication of how channel dredging costs are distributed among

114

1969 1973 1991 1994 1997 2000 2003 2006

Air Terminals *JFK 20 15 52 84 106 131 67 116LaGuardia 8 5 11 25 25 37 12 34Newark 4 (2) 52 62 99 132 110 163Total 31 18 114 169 228 304 200 220

Port Commerce *Elizabeth Marine Terminal 4 4 11 9 8 14 3 9Brooklyn 1 (2) (8) (12) (11) (12) (37) (40)Red Hook 1 (-1) (1) (5) (8) (5) (6) (25)Howland Hook (5) (8) (5) (3) (10) (19)Auto Marine 2 2 0 2 2 2Total 10 (2) (25) (34) (29) (6) (57) (84)

Tunnels, Bridges and Bus Terminals29 38 120 122 123 119 181 163

PATH (14) (28) (164) (199) (177) (180) (218) (187)

World Trade Center 50 (1) 5 33 113 276

Regional development (8) 0 (15) (11) (90) 11

PFC Program � 100 121 118 90 121

Other � (26) (2) 28 (3) 666 (2)

Combined total 57 26 60 153 282 372 883 519

Source: 1969 and 1973 from Walsh (1978, 92–93). Note that 1969 does not include depreciation. 1991 to 2006

from Schedule E of PANYNJ annual reports.

* Air terminal and port commerce totals also include several smaller facilties.

� The Passenger Facility Charge (PFC) is a small tax levied on airline passengers.

� This category primarily covers income and loss on damage to the World Trade Center in 1993 and 2001. It also

includes non-freight and non-passenger port activities.

Table 5.1: Port Authority facility income (loss) in millions

the terminals. However, the historically positive revenue stream from the Elizabeth

Marine Terminal suggests that while the two largest container handling operations

may not generate significant revenue, they are not a drag on the PANYNJ either.

115

5.4 Competition

As Chapter 3’s narrative suggested, the late 1990s were years of structural transfor-

mation for the freight industry. In particular, 1998 and 1999 were pivotal for Maersk,

Sea-Land, and the ports of the North Atlantic. As Maersk and Sea-Land were ap-

proaching the end of their slow dance toward merger behind the scenes, publicly they

were announcing an intention to further rationalize their operations by merging their

terminals. Sea-Land’s 25-year contract with the Port Authority of New York and

New Jersey was expiring in February 1999, and Maersk’s subsidiary Universal Marine

Terminals, Inc. had a contract with the PANYNJ that would expire in two stages in

October 1999 and November 2000 (Tirschwell 1998). The companies announced in

December 1997 that they would pursue a hub-and-spoke load center approach that

would concentrate the bulk of their container traffic in a single North Atlantic port.

By February 1998 it was clear that Maersk and Sea-Land would consider ports other

New York and New Jersey (Tirschwell 1998). And in early May the companies asked

six ports for proposals (Brennan 1999j).

5.4.1 Background: Easy marks

The shipping industry entered a downcycle in 1998. Liner companies seeking to cut

costs were finding it difficult to identify areas where costs could be cut. In a Journal

of Commerce article on the state of the shipping industry based on interviews with

industry analysts and a Moody’s report that had recently been released, Mongelluzzo

(1999b) claimed that carriers had already streamlined their own operations by cen-

tralizing documentation and customer service functions. They had introduced larger,

more efficient vessels. And they had negotiated advantageous intermodal contracts

with railroads in the early 1990s. An industry consultant and former chief officer of

116

“K” Line America, Theodore Prince, said, “The next largest expense left is the termi-

nals. If they’re going to get another $100 per container in savings, the only place they

can go is the terminal. There’s no other fruit out there” (Quoted in Mongelluzzo

1999b).

To this end, liner companies and alliances were consolidating their vessel calls at

fewer hub ports, or load centers. The large volume of cargo controlled by carriers

became a bargaining tool in negotiations with ports. The report goes on to quote Leo

Donovan, vice president of Booz Allen & Hamilton Inc., as suggesting that shipping

lines had historically believed that concessions could more easily be wrested from

port authorities than private sector vendors, since the latter are profit-driven. “Ports

are more driven by market share. The lines felt they were an easy mark” (Quoted in

Mongelluzzo 1999b).

As the review of port authorities above (Section 5.2) indicated, not all port au-

thorities are driven by motives other than profit. The PANYNJ was a self-sustaining

authority with a market-incentive to maintain profitability. In line with this financial

position, the Port Authority of New York and New Jersey adopted a much more hard-

line stance, despite the fact that Maersk and Sea-Land constituted roughly 25 percent

of the port’s total throughput. In the renegotiation of the Sea-Land contract, the PA-

NYNJ demanded market rate terms. Because Sea-Land’s expiring 25-year contract

from 1974 had been granted at exceptionally favorable terms with a clause prohibiting

any escalation of fees to encourage the growth of the port, market rate prices would

translate into a 300 percent increase (Brennan 1998c; Doig 2001; Mongelluzzo 1999b).

117

5.4.2 The drudgery of dredging

One major obstacle in the way of a successful bid facing the Port Authority was

channel dredging. New York Harbor’s shallow drafts did not permit the largest con-

tainerships to enter fully loaded. As a consequence, New York was losing up to ten

percent of its cargo to other ports, mainly Halifax, which has deep drafts and direct

rail access to the Midwest (Gersten 1998). Maintenance dredging in the harbor and

channel deepening to 45 feet for Sea-Land and Maher terminals had succeeded in

bringing some cargo back by winter of 1998(Gersten 1998), but it was still insufficient

for the largest ships. While 45 feet was seen as a minimum depth requirement, 50 feet

was seen as desirable. Uncertainties surrounding the PANYNJ’s ability to finance and

complete dredging to such depths was reportedly one of the major reasons Maersk

and Sea-Land opened bidding up for their joint terminal (Brennan 1998b,d).

There were three reasons for uncertainty: the toxins, the tunnel, and the tax.

First, environmental groups’ opposition to dredging’s disturbance of habitats and

chemically tainted sediment had been rising through the 1990s. In practice, this did

not block dredging; it simply slowed down the process and introduced adjustments

to dredging plans (cf. Hevesi [1994] and Revkin [1997]).

The second concern was the alternative plan favored by Congressman Jerry Nadler

and Senator Daniel Patrick Moynihan, New York City Mayor Rudolph Giuliani, and

others: the Cross Harbor Tunnel. For an estimated $4 billion, proponents argued that

a rail tunnel should be bored under New York Harbor to connect Brooklyn to the

mainland railways either directly in New Jersey or via Staten Island. Two arguments

were marshalled in favor of this plan. First, depths at the Brooklyn terminals were

already sufficient for the largest ships. Second, New York’s maritime activities had

withered with the introduction of containerization because of insufficient access to rail

118

(Bagli 1998; Bowen 1998). Regardless of its merits or lack thereof, this plan served

and continues to serve as ammunition in disputes over the distribution between the

two states of the flow of revenues from Port Authority activities.

The major ongoing uncertainty, however, became the Supreme Court’s late March

1998 ruling that the harbor maintenance tax on exports was unconstitutional. Created

in the 1980s, the harbor maintenance tax replaced general treasury funds for dredging

and similar harbor maintenance activities. Once this was struck down, there was no

clear source of funding for the enormous undertaking that dredging in New York would

entail (Journal of Commerce 1998a). While the Clinton administration proposed the

Harbor Services Fund Act, which ultimately failed, and promised initial funds toward

the $621 million project to dredge to 45 feet, long term funding remained in question

(Brennan 1998c).

5.4.3 The request for proposals

In the face of rising rates and uncertainty over the port’s future ability to handle

large containerships, in May 1998 Maersk and Sea-Land requested proposals for ter-

minals from seven ports: Norfolk, Va.; Baltimore; Philadelphia; New York-New Jer-

sey; Quonset Point, R.I.; Boston; and Halifax. Proposals had to offer a facility that

could handle 550,000 container movements annually with 6,000 contiguous feet of

berth space capable of handling up to four post-Panamax ships and sixteen cranes

at one time and on-dock or near-dock rail access, preferably to more than one rail

operator (Brennan 1999j). The companies also demanded operational control of the

facility (Getzfred 1998).

Proposals were submitted in September of 1998. Two of the seven terminals

were knocked out fairly quickly. Boston had declined to bid, citing a lack of space.

119

Philadelphia was rejected early on for its reluctance to hand over operations to Maersk

and Sea-Land. Two more ports were rejected in a 10 December 1998 announcement.

The companies rejected the Virginia Port Authority’s bid for a terminal at Norfolk

because of the cost of developing the site, inability to get operational control of the

terminal, and rail access only via Norfolk-Southern rather than dual access with CSX

Transportation, which owned Sea-Land. And Quonset Point was turned down over

concerns that it would not be fully operational early enough as the new port had yet

to handle any cargo, some dredging would be required, and the rail link had not yet

been completed (Brennan 1998c).

The three finalists were: Baltimore, Halifax, and New York-New Jersey. The

Port of Halifax proposed to build a new, customized terminal capable of handling

550,000 TEUs alongside CN’s railyard (Brennan 1998c; Tower 1998b). By the end

of September, the port had secured guarantees for a mixture of funds necessary to

build the terminal (about US$360 million) from banks and the federal and provincial

governments. Although passage of the Canada Marine Act would forbid Transport

Canada funding for port construction, the director of the Halifax Port Authority was

confident that other federal funding sources would be available. Further the provincial

premier publicly pledged that funding would not be a problem (Tower 1998b). Soon

after the submission deadline, CN and a private sector terminal operator, Halifax

International Terminals Ltd. (or Halterm), proposed an alternative common-user

facility (Tower 1998a). Port analysts, however, accurately believed there was small

chance that Halifax would be successful in winning the bid due to its small local

market. They thought, however, it was possible that Halifax could be part of a two-

port strategy in which cargo destined for the Midwest would be unloaded in Halifax

before ships proceeded south to New York-New Jersey or Baltimore (Brennan 1998a).

Lack of direct rail access to the New York market was also cited as a weakness (Tower

120

1998a). Others suggested that Halifax’s hopes faded a month after ports submitted

their bids when the Ocean Shipping Reform Act was passed (14 October 1998). By

allowing shipping firms to sign confidential contracts, the law eliminated Halifax’s

advantage of freedom from conferences and rail rates that allowed shipping companies

the strategic advantage of setting their own rates without communicating those rates

to the competition (Prince 1999).

Baltimore emerged as a dark horse candidate. Port officials proposed converting

570-acre Dundalk Marine Terminal into a container terminal. Dundalk was highly

underutilized at the time, basically “sitting there empty and can be up and running

almost overnight” (Brennan 1998a). The Port of Baltimore also offered significantly

lower lease rates than other ports (between $30,000 and $39,000 per acre) in addi-

tion to solid rail connections to the North Atlantic interior, existing 50-foot deep

channels, and access to the Washington D.C. market, the nations fourth largest. It

eventually included perks like a dozen expensive gantry cranes and downtown office

space. Challenges centered primarily on rail. While the port enjoyed full double-stack

access to the south, there was no fully cleared double-stack route north to Chicago

and other destinations. There was also only a single rail link to the terminal: CSX’s

rival Norfolk-Southern (Brennan 1999e; Watson 1999b).

The Port Authority of New York and New Jersey proposed a 350-acre terminal

with 700,000 TEU capacity and on-dock rail facilities via the PANYNJ’s ExpressRail

service (Brennan 1998d). The two main issues with the PANYNJ proposal from the

shipping companies’ perspective were the uncertainty that the channels would be

dredged and the cost. The PANYNJ’s proposal offered just under $500 million in

subsidies over the 27-year life of the proposed lease, rising from $18,000 per acre to

$106,000 (Brennan 1999e).

121

5.4.4 What was at risk?

What did the the PANYNJ stand to lose? Maersk and Sea-Land represented roughly

25 percent of the port’s annual 1.7 million TEUs. A Journal of Commerce editorial

painted a worst-case scenario if Maersk and Sea-Land were to leave the port (Journal

of Commerce 1998b). There was no guarantee that another business would replace

them: “The amount of global cargo is increasing. The number of shipping lines,

however, is not.” Plus, the investment level required to equip a major terminal, an

estimated $150 million or more, is not available to many shipping lines (Journal of

Commerce 1999a). Estimates of time taken to recover 1998 levels of throughput

ranged from five to fifteen years (Brennan 1999k; Journal of Commerce 1999c).

In the meantime, the loss of traffic would directly affect cargo assessment fees,

which could potentially have chased away other port operators. Cargo assessment

fees were levied to provide benefits to longshoremen, including retirement coverage

and more importantly a guaranteed income if there was no work. As the total amount

per longshoreman was fixed in labor contracts, a loss in cargo would mean that the

fees would have to be distributed across a smaller volume of cargo, thereby raising the

price of shipping that cargo through the port. Conversely, it also meant that more

work reduced the amount necessary. The New York Shipping Association (NYSA) had

been chipping away at the assessments since 1987 and made a major cut in December

1997, as the competition began. As the number of longshoremen covered by these

benefits decreased by natural attrition from 20,000 in the 1960s to 2,700 in 1997, the

rates had come down and were considered to be contributing positively to the growth

in port traffic. The PANYNJ’s manager for port marketing and labor argued that the

supply and demand for labor had reached equilibrium for the first time in thirty years

(Newman 1997). Were the port to suddenly lose 25 percent of its traffic, climbing

122

rates would have the cumulative effect of driving existing operators out of the port,

which would in turn lead to yet higher fees and drive more traffic creating a “carrier

exodus” in the words of ILA president John Bowers (Brennan 1999c). Ultimately this

would lead to the port’s catastrophic implosion. This would have severely affected the

estimated 170,000 port-related jobs located in the region (Brennan 1999k; Journal of

Commerce 1999c).

5.4.5 Boxed in

Though Maersk and Sea-Land were to initially announce their decision in December

of 1998, instead they narrowed the contestants down to the three discussed above and

repeatedly delayed making a final decision for five months as they negotiated with

various actors. By February 1999, it became clear that Halifax had been crossed off

the short list, leaving Baltimore and New York-New Jersey. While the latter held

firm to its initial proposal at first, as the Port of Baltimore offered more and more

incentives to Maersk and Sea-Land, cracks began to appear.

Labor fragmentation

First, Maersk and Sea-Land took advantage of union fragmentation to negotiate con-

cessions from the International Longshoremen’s Association (ILA) locals. The hard-

won unity the ILA had built in the 1970s broke down ten years later as the industry

began to change and employers more actively utilized non-union labor in the right-

to-work states of the South and the Gulf (see Chapter 8 for this background). While

the ILA leadership pushed continuously for higher wages and greater benefits, South

Atlantic dockers began to worry that wages and benefits negotiated with regard to

conditions in New York and New Jersey were undermining their own competitiveness.

123

Work was being shifted to non-union employers and workers (cf. Erem and Durren-

berger [2008] for a compelling account one episode in this struggle). And there was

fear that cargo was going to shift to other ports. These concerns manifested them-

selves in 1986 when West Gulf and South Atlantic locals made concessions on wages

and benefits in defiance of the union’s leaders. Over the next decade, the ILA frag-

mented as it abandoned its policy of “one port down, all ports down” and relinquished

negotiations to locals.

Thus, outside of a few items negotiated in the Master Contract between the ILA

and employers’ association, U.S. Maritime Alliance Ltd. (USMX), the union’s locals

negotiated only over terms for their own port. This allowed Maersk and Sea-Land

to go to Baltimore and negotiate a letter of intent to keep truck gates open longer,

to keep the railroads open every day, and to expand starting times, which would cut

down on overtime pay (Brennan 1999b). In early March, New Jersey locals, partly at

the insistence of New York Governor Pataki, made concessions that matched those

of the Baltimore ILA locals in overall value (Brennan 1999h). Fragmented organi-

zation thus weakened labor vis-a-vis employers and resulted in diminished working

conditions. By comparison, a competition between Long Beach and Los Angeles for a

Maersk terminal during the same period did not involve labor issues. This is almost

assuredly due to the fact that the International Longshore and Warehouse Union

(ILWU) represents all longshoremen on the West Coast, eliminating any possibility

of playing one local off against another (Mongelluzzo 1999a).

Port Authority fragmentation

Though relations among PANYNJ board members were traditionally in keeping with

the “cordial co-operation” called for in the Port Compact of 1921, Walsh (1978, 182)

reports that after 1965 relations became more strained. In the 1990s, these strains

124

became more evident. New York had come to believe that Port Authority revenues

and benefits were inequitably distributed in New Jersey’s favor. As was shown in

Table 5.1, the two airports in New York City generate a significant stream of revenue

that is redirected to subsidizing PATH and the ports, which are overwhelmingly in

New Jersey. While New York representatives complain that their state sees little

benefit from Port Authority activities (Bagli 1999a; Journal of Commerce 1999b),

New Jersey representatives counter that the state collects income taxes from New

Jersey residents who work in New York City and that roughly 40 percent of port-

related jobs, particularly freight forwarding jobs, are located in New York, providing

ample compensation (Armbruster 2000a).

This conflict slowly rose to the surface during the competition. First, it was the

Cross Harbor Tunnel. In the early 1990s, Jerry Nadler introduced the Cross Harbor

Tunnel discussed previously (Section 5.4.2) as a potential means of redirecting port

job creation to Brooklyn. It has served as a political whip wielded by New York

to capture benefits from the Port Authority. Giuliani picked up this call in 1996

(Bagli 1998). And as the competition began, New York State Senators Moynihan

and D’Amato publicly endorsed the tunnel side-by-side with Giuliani and announced

a study of its feasibility. Presumably, this was to indicate that New York wanted a

greater share of the benefits that port growth could generate. As a spatial strategy, the

Cross Harbor Tunnel would shift the economic terrain within the territory governed

by the PANYNJ by permitting the efficient flow of goods beyond the eastern coast of

the mainland to Long Island.

Then, soon after a New York University Taub Center report dismissed the Cross

Harbor Tunnel as“Tunnel Vision”(Bagli 1998), Giuliani proposed something radically

different: withdrawing from the Port Authority (Giuliani 1999). In the late 1940s,

New York City had run into financial difficulty supporting its airports and worked

125

with the Port Authority for the latter to lease the land upon which they were located

and operate the airports (Doig 2001, 288–314). Renewed in the 1980s, the contracts

were set to expire again in 2015. Just as Maersk and Sea-Land were taking advantage

of their contract renewals to extract benefits from the PANYNJ by leaving, Mayor

Giuliani sought to wring revenues from the authority by threatening its livelihood.

We have got to get out of the Port Authority. This makes no sense to be

partners with our friends in New Jersey because they extract more out

of it than we do. I think that’s been proven over and over again. So

what we’re going to do this year is we’re going to announce that we’re

not going to renew the lease when it comes up in 2015, and we’re going

to immediately put out Requests for Expressions of Interest by private

companies - we know of two that are very interested - in taking over the

two airports, and to begin now to make concrete plans for how they would

change those airports. Start it now. That may be a little bit soon—but

actually, given the way things work, it isn’t so soon. And hopefully by

doing that, we will increase the pressure on them maybe to move aside.

Because they’re not going to be continued. We’re not going to continue

to allow them to lease that space from the City of New York any longer.

They have proven that they can’t do a good job of running the airports

in the interests of the City of New York. It isn’t personal. It’s purely

business.

This approach would have reasserted the primacy of state boundaries in circumscrib-

ing the flow of capital in the region. This territorial strategy would have retained the

revenues from renegotiated airport contracts within the territorial boundaries of the

city and state while leaving the costly regional infrastructure to New Jersey.

126

From the perspective of jobs, this position was a fairly easy one for New York

to take. While New Jersey’s jobs were directly related to the concrete movement of

freight, New York’s jobs were less tangibly connected. It is clear from much literature

that the financial activities related to manufacturing and transportation are driven

more to colocate than to relocate when manufacturing does (See, for example: Cohen

1981; Frobel, Heinrichs, and Kreye 1980; Sassen 2002). Additionally, for the other

job category most cited, freight transportation arrangement, population density far

outweighs infrastructure facilities as a determinant of location, as shown in Table C.3.

So, while New York certainly is home to many freight transportation arrangement

firms, as Figure 6.5 shows, it did not have to fear losing a great number of them

should Maersk and Sea-Land leave. As the facilities serving the maritime industry

are primarily located in New Jersey, support for these jobs became critical to the

governor. As Figure 6.1 shows, marine cargo handling jobs were virtually nonexistent

in New York at the time, while New Jersey recorded approximately 4,000. New Jersey

Governor Christie Whitman’s interest in keeping these jobs became clear at this time

as well. Just prior to Giuliani’s threat, she met directly with Maersk and Sea-Land

to convince them to build in New Jersey (Brennan 1999b). And from this point on,

she was very visibly and vocally on the side of labor, appearing at their rallies and

publicly deriding Pataki for threatening their jobs (Brennan 1999f). There was thus a

territorial divide in the nature of jobs and risk from a Maersk and Sea-Land departure

that followed the state lines down the Hudson River.

This territorial fragmentation within the PANYNJ may not have been entirely

counterproductive, however. Though analysts and actors were mixed on whether or

not Maersk and Sea-Land’s request for proposals was serious or just a ploy, opinion

appears to have fallen more heavily on the side of a ploy. Pataki apparently agreed

with the analysts and never believed that Maersk and Sea-Land would leave the port.

127

According to Brennan (1999i), a Pataki supporter familiar with the talks said, “New

York felt that the two lines never wanted to leave the region because it would have

meant a huge loss of market share for them in a market that has a huge number

of consumers.” In Levinson’s (1967) terms, Maersk and Sea-Land would not want to

leave the New York-New Jersey product market area for freight transportation. Pataki

also felt strongly that the port’s maritime division should be self-supporting since that

would reduce the state’s transfer of revenue from the airports (Brennan 1999e). This

gave him an intransigence that produced an inflexibility in the PANYNJ’s negotiating

stance. If he was correct, as many analysts believed, his position may have prevented

the PANYNJ from making even greater concessions (Brennan 1999b).

5.4.6 Shunted away

Maersk and Sea-Land delayed their announcement by a month in February to also

pursue negotiations with regard to rail in Baltimore. Port of Baltimore officials met

regularly with CSX, Norfolk-Southern, Maersk, and Sea-Land representatives to work

out ways in which CSX could have access to Dundalk terminal. Changes would have

required investments to alter clearances through Maryland and Delaware (Brennan

1999h; Watson 1999b). Meanwhile, in response to Maersk and Sea-Land’s consistent

concerns over dredging funds, Governor Whitman began unilaterally offering to come

up with over $120 million from state transportation funds to indirectly offset the costs

of dredging and make a final offer that she was convinced Maersk and Sea-Land would

accept (Brennan 1999f).

With Whitman sweetening the PANYNJ proposal, it appears that the shipping

companies were reaching a final decision. After the rail discussions, CSX presented

a list of improvements it felt would be necessary to make the Port of Baltimore a

128

competitive load-center. The $550 million price tag included a third main track in

Maryland, double-stack clearances, a new intermodal yard in New Jersey, Baltimore

port enhancements, and other rail improvements. CSX denied that this wishlist was

directly related to the competition, despite the fact that it was owned by the same

parent company as Sea-Land. Norfolk-Southern’s response reportedly outlined needed

improvements that totalled less than $40 million (Watson 1999d,e).

On 12 April 1999 both ports submitted their “last, best offers.” The Port of

Baltimore’s bid, though not public, reportedly included some enhancements to rail

that focused on dual access to the terminal rather than the broader list CSX pro-

posed, which John Pocari, Maryland’s secretary of transportation, described as “very

ambitious” (Watson 1999d,e). The PANYNJ bid included Whitman’s $120 million

sweetener and improved rates. The deal would also entail $1.6 to $1.7 billion in re-

lated infrastructure projects over the life of the lease. Though Pataki refused to sign

the proposal, Maersk and Sea-Land accepted the Whitman-approved proposal as fi-

nal. When the proposal was submitted, Pataki made the price of his approval clear:

restructuring the agency along stricter state lines (Brennan 1999a).

Four weeks later on May 7, the shipping companies made their final announcement.

They would build their new terminal in New Jersey. Lack of rail improvements were

cited as the major reason for not selecting Baltimore (Brennan 1999i). This suggests

that perhaps CSX’s list was the final price tag, or that the exceedingly pricey list was a

convenient way for Maersk and Sea-Land to finally break off the deal with Baltimore

now that it had the $121 million sweetener from Whitman, which the companies

cited as critical to establishing an economically competitive bid. James White, the

executive director of the Maryland Port Administration, supported this view when

he said that, “CSX didn’t do anything to help us,” an allegation Maersk and Sea-

Land described as only a minor concern in their decision (Watson 1999a). At any

129

rate, Maersk and Sea-Land wound up where they started with significant subsidies

compared to market rate leases, promises of dredging, and improved rail access.

According to Charles Gargano, deputy chairman of the Port Authority Board

of Commissioners (Quoted in Armbruster 2000b), the proposed lease would subsi-

dize Maersk and Sea-Land with $587 million over 30 years, not including dredging,

and would provide Maersk and Sea-Land with $30.4 million for infrastructure im-

provements. In exchange, following generally recognized best practice (Luberoff and

Walder 2000), Maersk and Sea-Land would guarantee unspecified minimum volumes

if dredging to 45 feet was completed by 2004 and 50 feet by 2009. For their own pur-

poses, Maersk and Sea-Land would likely invest $150–200 million in the new terminal,

though this was not specified in the contract.

5.5 Territorial squabbles

The contract, however, would not be finalized for more than a year. In a letter

dated April 20, just after the final proposal and before the decision announcement,

Pataki demanded a restructuring of the Port Authority. He specifically demanded

that New Jersey guarantee the entire subsidy for the Maersk and Sea-Land lease,

which would total over $600 million. While Whitman stated her willingness to discuss

restructuring prior to the final proposal, she and other New Jersey lawmakers argued

that the discussion should not hold up approval of the lease (Brennan 1999g). Pataki

opted to use the situtation to New York’s advantage, directing his representative

on the Port Authority board to refuse to vote on the lease (Armbruster 2000a). In

retaliation, New Jersey representatives refused to vote on projects New York badly

wanted (Smothers 2000). After more than a year of abortive board meetings, a deal

was finally reached in May 2000 and a memorandum signed and publicized on 1 June

130

2000.

In addition to the approval of the Maersk and Sea-Land lease and the Maher

lease, which had been held up by the delay, the memorandum of items to be acted on

at a special meeting of the Port Authority Board of Commissioners on 2 June 2000

included a number of terms that directed PANYNJ resources to New York (Journal of

Commerce 2000b). The main item was the creation of a Regional Bank for New York

totalling $250 million to be paid in equal increments over three years and to be spent

on “projects consistent with the purposes of the Port Authority.” The agreement

also stipulated immediate and rapid action on releasing and processing of a request

for proposals for the World Trade Center net lease, which had been agreed to in

1998 and was completed by the end of April 2001 (Bagli 2001). This project and

the authorization for the lease of air rights above the Port Authority Bus Terminal

were two projects that New York had wanted badly and were held up. Funding for

a JFK Airport link from Jamaica Station in Queens was approved. A number of

items were oriented toward improving conditions at New York’s marine terminals.

The deal included rent adjustments for Staten Island’s Howland Hook Terminal. A

rail connection between Staten Island Railroad and the Chemical Coast line and the

continuation of the Red Hook Barge Program were approved. Cranes were to be

purchased for Howland Hook and Red Hook, and a project to improve eleven acres

of Howland Hook was approved. New Jersey was released from its obligation to

pay the $120 million subsidy. The reconfiguration and redevelopment of New Jersey

marine terminals was approved, though this was necessary for the Maersk and Sea-

Land lease to go forward anyway. Also a feasibility study of extending PATH to the

Newark Airport Monorail was approved.

All in all, by raising the spectre of reasserting its territorial monopoly over the

flow of revenue from its airports, New York was able to redirect the flow of some of

131

the PANYNJ’s revenues back into its territory. These gains were reinforced several

years later when New York City Mayor Bloomberg negotiated extensions of the Port

Authority’s leases for La Guardia and John F. Kennedy Airports until 2050 for a $700

million upfront payment and significantly higher annual rents (Cooper 2003).

5.6 Costs of competition

It seems fairly clear that Maersk and Sea-Land benefitted from the competition they

initiated, successfully reducing the PANYNJ’s initial lease to one that subsidized the

companies at over half a billion dollars over thirty years. It also seems that New

York turned the situation to its advantage as well. Though losing the possibility of

investing the subsidies in New York-oriented projects, Pataki managed to guarantee

Port Authority investment in New York. But what were the costs?

5.6.1 Port costs

The first question to address is whether or not Maersk and Sea-Land would really

have left New Jersey. As discussed above (Section 5.4.5), most analysts did not

believe that this would happen. The regional consumer base, the product market

area for freight transportation services, was considered too large for the companies to

abandon. Most likely the worst case scenario would have been a scaled down terminal

with discretionary cargo, i.e., cargo with destinations other than the New York-New

Jersey region, being shipped through Halifax or Baltimore. This is in part due to the

high rail costs of moving freight from Baltimore (Brennan 1999i). But it is also due

to fragmentation within the imminent Maersk and Sea-Land merger. Brooks (2000,

90–91,170) documents conflicting network strategies between Maersk and Sea-Land.

While Maersk had a predilection for hub-and-spoke networks with load centers, Sea-

132

Land had a preference for increased direct calls at ports, which became one of the

alliance’s central goals. It is thus highly unlikely that the two companies would have

completely abandoned the port.

But suppose they had. Would New York and New Jersey have suffered from the

loss? Evidence seems to indicate that the loss would have been minimal. Even as

the deal was being negotiated, Hanjin, P&O Ports, and a cosortium led by Ameri-

can Stevedoring, Inc. were bidding for the 160-acre Maersk Universal terminal that

Maersk would abandon (Armbruster 2000c). Hanjin’s bid was already higher than

Maersk and Sea-Land’s (Bagli 1999b), but reports suggest that P&O Ports, which

eventually won the lease, was offering much more, perhaps two or three times the

$19,000 per acre that Maersk and Sea-Land were to pay (Journal of Commerce 2000a;

Storey and Armbruster 2000). One presumes that the time required to negotiate and

redevelop the terminals in accordance with the needs of these new tenants is respons-

bile for analysts claim that the loss “would have undercut New York’s traditional role

as the Northeast’s transport and distribution hub for at least five years” (Brennan

1999k). Arguably, this would have been a small loss, as the largest subsidies for

Maersk and Sea-Land were in the first years of its new lease.

It is also highly unlikely that the port would have lost container volume, leading

to the collapse of the port as some doomsayers suggested (Brennan 1999c). Figure 3.2

shows that traffic has continued to increase steadily throughout the U.S., implying

that what was at stake was more market share than absolute loss of throughput, as

many analysts have suggested (Mongelluzzo 1999b). More specifically, as Figure 5.1

shows, container traffic grew in Baltimore and leveled off in Halifax, while it skyrock-

eted in New York and New Jersey. Baltimore, thus, seems to have gained from the

exposure of the competition, and Halifax remained unaffected. It is probably safe

to assume, then, that New York and New Jersey’s traffic would also have increased.

133

Indeed, the gain of roughly 2.5 million TEUs since the deal was established far out-

weighs the doubling of container throughput to 700,000 TEUs made possible by the

new terminal. Surely, a momentary drop in port rankings is probably the worst that

the PANYNJ would have experienced.

However, the concessions made by the PANYNJ have not been contained to the

Maersk and Sea-Land deal. Eight years after the deal was signed, Maher filed a law-

suit before the Federal Maritime Commission against the Port Authority for violations

of the Ocean Shipping Reform Act of 1984. Specifically, the terminal operator is seek-

ing a “cease and desist order and reparations for injuries caused to it by PANYNJ’s

violations. . . , because PANYNJ (a) gave and continues to give an undue or unreason-

able prejudice or disadvantage with respect to Maher, [and] (b) gave and continues

to give an undue or unreasonable preference or advantage to” APM Terminals North

America, Inc. (Maher Terminals, LLC 2008). Dennis Lombardi, the Deputy Director

of Port Commerce at the PANYNJ, was under the impression that the suit was likely

to lead to an out-of-court settlement of some unspecified value, adding to the Port

Authority’s loss on the deal (Interview with Dennis Lombardi, March 25, 2008).

5.6.2 Labor costs

It is also highly likely that labor has suffered as a result of its fragmentation due

to the competition among ports. First, New York and New Jersey longshoremen

were unlikely to take a significant loss if Maersk and Sea-Land left the port. Second,

concessions made by locals have diminished their working conditions as subsequent

contracts have begun from the terms agreed to attract Maersk and Sea-Land. Third,

broader, long term trends in the industry suggest an ongoing loss of port-related

employment.

134

As discussed above, the interest of other terminal operators to locate in New

York Harbor and more general growth of container traffic would have sustained labor

demand, even if Maersk and Sea-Land chose to leave for Baltimore or Halifax. The

growing demand for longshore labor led, in fact, to the first opening in the New York

and New Jersey ILA rolls since they were closed in 1966 just one month after the deal

was accepted in principle by Maersk and Sea-Land (Brennan 1999l). This implies that

longshoremen were close to full employment and existing union members would not

have been affected by a slight decline in jobs going forward. Further confirmation was

provided by Sea-Land’s regional manager at the time, Jim Devine, who was quoted as

saying, “We’re pressed for workers at our piers right now and will be even more pressed

for those who have adequate training as our volumes increase.” Meanwhile, Baltimore

labor was “eager for the work” at its “struggling” terminal (Brennan 1999b). It thus

seems reasonable to agree with Theodore Prince’s editorial claim that “the gain of the

ILA members in one was a victory to be won at the expense of their union brothers

in the other” (Prince 1999).

More significantly, the concessions offered by labor to attract Maersk and Sea-Land

became the new starting point for negotiations with other shipping and stevedoring

firms. For instance, in subsequent negotiations in the Port of Baltimore for stevedoring

services, the local ILA was forced to grant the same concessions it had offered Maersk

and Sea-Land.

Finally, as Chapter 6 will demonstrate, trends in port activity employment are

generally downward and are likely to continue in that direction over the long term.

Containerization reduced the number of workers required to operate terminals over

time. This is particularly so on a per TEU level. As terminal automation continues

to advance, even the small proportion of remaining jobs will be pared down. Thus, it

is unlikely that even short term gains will persist for long.

135

5.7 Conclusion

This chapter has traced central aspects of one episode in interport competition that

took place at a transitional moment in U.S. maritime evolution. Legal changes, merg-

ers and acquisitions, alliances, and lease expirations all conspired in 1998 to offer an

example of how global shipping companies have become able to take advantage of

their intermodal networks’ flexibility to initiate incentive competition among ports.

The account provided here reinforces Markusen and Nesse (2007) institutional ap-

proach to interurban competition, which looks beyond simple, microeconomic models

of profit maximization toward the particular institution settings of localities in shap-

ing the nature and outcom eof interlocal competition, including political cycles and

local interests. This case highlights the complexity of local, regional, national, and

even international political and economic factors in shaping the dynamics of compe-

tition. However, it augments the two causes of the rise of such competition since the

1980s with a third. Markusen and Nesse suggest that the emergence of site consul-

tants and the devolution of responsibility for economic development to subnational

units has substantially aggravated interurban competition. The case examined here

suggests a complement to devolution that I will call de facto devolution. As net-

works of production and particularly transportation have scaled upward, entities like

port authorities and even states that once enjoyed territorial monopolies over regional

product market areas have lost them and thus experienced the same loss of control.

The case study also illustrates four points relevant to the spatial strategies em-

ployed in contemporary intermodal systems. First, intermodal technologies and orga-

nization have expanded the effective area of production for each product market area,

shifting power toward shipping companies and away from ports and labor. That is,

while ports at one time monopolized transportation access to their hinterlands, in-

136

termodalism has eased shipping companies’ capacity for serving a given region from

multiple ports, weakening ports’ bargaining power. In the case examined here, Maersk

and Sea-Land were exploring ways to serve the New York region from Halifax and

Baltimore, forcing the PANYNJ ultimately to make sizable concessions.

Second, the complex reorganization and adjustments to the rail system that

Maersk and Sea-Land demanded from Maryland highlights the interdependency of

the contemporary intermodal freight network. It also offers an example of Hughes’

reverse salients. Progress in advancing the competitiveness of Chesapeake Bay ports

was being held back by a lack of development elsewhere in the system. The interstate

coordination that would have been necessary in this case suggests that a larger scale

mechanism for coordinating the development of the national freight system might

facilitate system development.

Third, following Harvey (1989), it is possible to suggest that interport competition

foists the greatest proportion of risk on ports. They must invest (with the federal

government) enormous sums to dredge for ever larger ships that may not come. In

compelling port authorities to offer large subsidies that reduce their direct revenue be-

low break-even levels in hopes of generating port-dependent jobs, shipping companies

effectively minimize their operating risk and ensure profits.

Finally, the wrangling between the governors of New York and New Jersey dur-

ing the deal underlines the potential fragility of associative forms of collaborative

governance. By insisting that the PANYNJ’s spending be more evenly distributed

between the two states, Pataki undermined the Port Authority’s regional focus. By

threatening to withdraw from the Port Authority compact, Pataki and Giuliani were

reasserting their territorial dominance over the flow of revenues generated within their

territory, regardless of the regional basis for those revenues. Though the fractious res-

olution of the dispute suggests some stability in the associative form of governance

137

and the possibility of extending it to include other ports, the squabbling also prefig-

ures the types of conflicts that might be anticipated if this coordination mechanism

for governing the national freight system were adopted.

138

Chapter 6

Direct Employment

6.1 Introduction

In this chapter and the following, I turn to the question of whether fostering port

freight volume through incentives continues to provide the regional benefits it tradi-

tionally has. The political justifications offered to the public for incentives fall into

three general categories: direct employment in logisitics services, indirect and induced

employment, and support for economic activities whose cost competitiveness is de-

pendent on an efficient, low-cost, local transportation infrastructure. By highlighting

the role of technological change in spatio-organizational restructuring, these chap-

ters demonstrate that the impact of ports as nodes in global production networks

on local and regional economies is declining, calling into question local motivations

for subsidizing infrastructure improvements that will benefit the private sector and

not the residents themselves. This chapter proceeds by reviewing the probably insur-

mountable weaknesses of port impact evaluations to date. It then addresses direct

employment trends in port labor, trucking, and warehousing, demonstrating a decline

in the first and suggesting that the latter have grown through urbanization economies

139

rather than localization economies, i.e., through population growth rather than port

activity. Recent research on the quality and character of these jobs is then reviewed,

suggesting that the growing number of jobs in these sectors may not be the safe,

stable, well-paying jobs that governments should seek for their residents.

6.2 Economic impact analyses

As early as Waters (1977), academics have bemoaned the difficulties of calculating the

economic impacts of port activities on a region. The benefits are generally defined as

the income a community receives per ton of cargo moving through its port in the form

of jobs, sales, income, and taxes (Davis 1983; Waters 1977), and they are generally

broken down into four categories (Davis 1983). Direct impacts include sales, jobs,

and incomes derived from owning logistics firms or being employed by one. Indirect

impacts include sales, jobs, and income generated by the purchase of goods and

services by firms directly involved in providing logistics services. Induced effects refer

to sales, jobs, and incomes created by the personal expenditures of employees and

owners of directly related firms. Finally, Port-dependent impacts, which are the most

difficult to measure, are the sales, jobs, and income derived by firms as a result of the

lower transportation costs achieved by locating proximate to a port.

Measuring these values in a manner that enables port planning has proven nearly

impossible. Commonly recognized obstacles include defining port activities, utilizing

aggregate multipliers, accounting for modal substitution, extrapolating from existing

services rather than calculating the effects of incremental changes, and sidestepping

technological change and its spatial reorganization of the logisitics system. To this

list, we can add an evaluation of the quality of the jobs created.

The first problem is that of identifying port-related activities. This obstacle is

140

not one of identifying the character of these activities but rather one of locating

those activities within available data. Though there is a generally agreed upon set of

activities directly involved in port functions (see Table B.1), there is no single Stan-

dard Industrial Classification (SIC) code or North American Industrial Classification

System (NAICS) code that encompasses all these activities (Davis 1983; de Langen

2007). The result is that economic impact studies tend to share an overlapping core

of activities but lack the consistency necessary for effective comparison. This problem

is aggravated by censorship in the County Business Patterns data. Not only does the

data set suppress data that discloses too much information about individual firms

but it sometimes also suppresses data on entire sectors. For instance, rail-related

employment, which is often touted as a major beneficiary of port operations, is not

reported at all.

The second problem is the use of single aggregate multipliers for disparate indus-

trial sectors. As Davis (1983) argues, three of the four methods generally used to

determine economic impacts (economic base analysis, income-expenditure analysis,

and the application of previously established multipliers) yield single aggregate mul-

tipliers. But as Waters (1977, 15–17) claims, single multipliers can greatly distort the

nature of geography- and sector-specific economic impacts. For instance, his analysis

shows that the multiplier for forest products in the state of Washington is 27 and that

for agricultural products is 60, the latter being vastly different from the typical range

of 20 to 37. This suggests that a single multiplier will fail to capture the true economic

impact of change in port capacity or local economic activity, since it fails to reflect the

particular composition of the local economy and the differential impact of transporta-

tion on its various components. To amend this problem, most contemporary analysis

employs input-output modeling, e.g., Lahr and A. Strauss-Weider Inc. (2004) and

Yochum and Agarwal (1987). This approach requires expensive, customized tables

141

for each region or a more generic set of assumptions about the relations between in-

dustrial sectors in addition to identifying port-specific sectors. Another approach is

to create weighting measures for the value-added by given commodities (Haezendonck

2001). This approach, however, also faces the problem of aggregation, as the weights

are generally assumed to be constant across geographically distant ports, even though

these ports support different combinations of economic activity that may not reflect

the weighting appropriate to the benchmark port (de Langen 2007, 189). Additionally,

the spatial fragmentation that has facilitated the decentralization of production has

also permitted the centralization of command functions (cf. Fainstein and Fainstein

1989), which implies that many of the finance and insurance jobs associated with port

activities need not colocate with ports’ freight-movement functions to be successful.

The third problem is ignoring the potential for substitution in shipping. Though

idenitified by Waters (1977) as early as 1977, this problem was ignored until Hall

(2004) examined the situation closely after the 2002 West Coast port lockout. Devel-

oping ideas latent in Anderson’s (2002) flash analysis of the situation, Hall demon-

strates that firms dealt with the two-week shut-down by substituting other modes

of transport for shipping (often air freight) and by shipping increased quantities of

goods prior to the lock-out in a form of intertemporal substitution. The long run

impact of this incident has been that shippers have developed redundant capacity in

East Coast and Gulf Coast ports to ensure that supply chains will not break down

should something similar occur in the future. The basic point is that simply because

a port’s activities “support” certain jobs, that port does not guarantee the ongoing

existence of these jobs.

The fourth problem is that studies base projections on simple extrapolations of ex-

isting relationships rather than on incremental changes in port services (Davis 1983).

Rather than assuming a fixed relation between port capacity and regional economic

142

activity, more rigorous economic analysis would suggest that the sector might en-

counter diminishing marginal returns as services increase. The development of new

facilities may not attract the same amount of activity that earlier facilities did. One

might cite Baltimore’s ambitous construction of a container terminal at great taxpayer

expense in the late 1980s that has remained underutilized (Luberoff and Walder 2000).

“Build it and they will come,” is not a motto on which the port industry can rely.

There are two additional problems with port economic impact analyses that con-

stitute the primary content of this chapter and the next. The first of these is that

the static nature of the coefficients generated by any of the methods, including input-

output analysis and value-added analysis, cannot account for the dynamism of tech-

nological change (Davis 1983). As technology shifts, the relation between firms also

shifts, as does the relation between capital and labor, resulting in a constantly chang-

ing set of coefficients to describe those relationships. While coefficients generated at

one time may fairly accurately describe the relation between a port and local eco-

nomic activity, they fail to indicate the direction in which those coefficients have

moved or are likely to move in the future, greatly reducing the ability to employ them

in successful planning.

In the logistics industry, the spatial impact of containerization adds an additional

layer of complexity. By reducing the time, labor, and costs of moving freight through

ports and the remainder of the logistics network, containerization and other tech-

nological improvements have expanded port hinterlands (Rodrigue and Notteboom

2009) and facilitated the spatial fragmentation of production (Whitford and Potter

2007), redefining the region impacted by a port. As a consequence, the localized im-

pact analyses typically conducted are no longer sufficient (de Langen 2007). Worse,

they are becoming irrelevant. As port hinterlands overlap and merge (Notteboom

and Rodrigue 2005, 2007), ports are coming to serve the same economic base. This

143

has two implications. First, it boosts substitutability within shipping. For instance,

as Hall (2004) and others have noted, a growing number of firms in the US have

developed bicoastal shipping strategies to increase flexibility, particularly since the

West Coast lock-out. Second, this substitutability makes it practically impossible to

distinguish which firms and sectors are reliant upon services of a particular port.

Finally, while economic impact analyses focus on the number of jobs created and

to some extent the income associated with those jobs, they do not question whether

or not these are good jobs. As the analysis below shows (see Figures 6.3 and 6.4),

most job growth has occurred in warehousing and trucking. However, recent research

on the impacts of deregulation in these sectors (Bensman 2009) indicates that they

are precarious, dangerous, poorly paid jobs with no benefits for non-union workers.

The following sections examine employment trends in directly related industries.

Indirect and induced employment is not evaluated, as these would require vastly dif-

ferent methodologies and are nonessential to the chapter’s argument. The dependence

of other industries on proximity to ports is addressed in the next chapter.

6.3 Port operations and marine cargo handling

There are a great variety of professions directly involved in port operations, including

tug boat operators, ship repairmen, crane operators, and, of course, longshoremen.

The original SIC classification (4463) was broken up into two parts under the NAICS

classification (488310 and 488320). Port and harbor operations (NAICS 488310)

comprises “establishments primarily engaged in operating ports, harbors (including

docking and pier facilities), or canals” (U.S. Dept. of Commerce, Bureau of the

Census 2007); and marine cargo handling (NAICS 488320) comprises “establishments

primarily engaged in providing stevedoring and other marine cargo handling services

144

(except warehousing)” (U.S. Dept. of Commerce, Bureau of the Census 2007). The

latter is what we traditionally refer to as longshoring, and constitutes roughly 87

percent of the earlier SIC classification that combines these two. Thus, figures for

employment in this category refer primarily to longshoremen.

1970 1980 1990 2000

05000

15000

Totalem

ploymen

t

MarylandNew JerseyNew YorkPennsylvaniaVirginiaFloridaGeorgiaSouth Carolina

1970 1980 1990 20000

5000

15000

CaliforniaOregonWashington

Figure 6.1: Total employment in marine cargo handling transportion (SIC 4463 andNAICS 488310 and 488320) by state (East Coast left, West Coast right)Source: Country Business Patterns

Employment in these areas, particularly that of freight handling, has demonstrated

a remarkable decline since 1970. Total employment declined by fifty percent from just

under 100,000 in 1970 to less than 50,000 in 1999 before climbing again to approach

70,000 today. As Figure 6.1 indicates, though there is generalized marginal growth in

several other states as well, almost the entirety of the recent upturn in employment

has occurred in California, where containerization growth has been highest. 1 Still,

as a fraction of total employment (see Table 4.2), the importance of longshoring has

declined even more steeply, from 0.17 percent of total employment to 0.06 percent, a

decline of two-thirds.

Table 6.1 offers another perspective on this relationship. It shows the number of

1Note that this growth represents new jobs, which are not offered guaranteed annual incomesunder the 1960s Mechanization and Modernization Agreement.

145

1979 1984 1989 1994 1999 2004

California 15.13 3.31 1.74 1.10 0.87 1.27Oregon 5.47 19.35 7.86 4.41 5.13 7.98Washington 0.58 2.06 1.39 1.18 1.02 1.16WCNA Average 8.17 3.42 1.80 1.23 0.98 1.30

Florida 28.62 16.05 4.12 3.26 1.89 2.50Georgia 12.64 3.75 4.55 4.85 2.20 1.23Maryland 16.33 7.78 5.82 5.55 4.12 3.40New Jersey Inf Inf 60.13 Inf Inf 111.64New York Inf Inf 13.00 Inf Inf 30.04Pennsylvania 18.25 10.39 23.98 13.50 9.04 7.72South Carolina 6.50 4.97 2.91 3.15 1.44 1.11Virginia 9.15 10.08 3.93 2.92 2.10 1.03PA and NJ 57.12 38.23 43.17 50.92 28.69 26.59PA via NJ Inf Inf 21.21 Inf Inf 34.77ECNA Average 24.56 12.45 5.76 5.03 2.89 2.49

Alabama 262.20 19.76 33.73 59.11 31.52 14.04Louisiana 42.46 21.55 19.80 15.60 18.14 14.31Texas 42.69 22.97 9.40 9.34 3.49 2.85USGC Average 49.73 22.20 13.57 12.23 6.58 4.57

PA and NJ combines volumes and employment for the two states.New York and PA via NJ use cargo volume from NJ only.Averages are based on totals for the states listed.Source: Country Business Patterns

Table 6.1: Employment in port operations and marine cargo handling by state per1,000 TEUs moved through ports in that state

jobs for every thousand TEUs2 handled by ports in the listed states.3 There has been

a marked decline from anywhere between 3.62 (New Jersey) and 14.95 (California)

for major container ports down to roughly one and a half longshoreman for every

1,000 TEUs and as low as one job in New Jersey. Note that the higher figures for

Oregon, Georgia, Maryland, Pennsylvania, and the Gulf Coast states reflect the much

larger role of bulk commodities in these states’ ports.4 The Port of South Louisiana,

2TEUs, or twenty-foot equivalent units, are basically half of the most common container. Thus,it is the equivalent of a twenty foot by eight foot by eight and a half foot box.

3While the West Coast ports are generally located well within state boundaries, allowing us toassume that all longshoring employment in these states are in their ports, there is potential overlapon the East Coast, particularly in New Jersey. Philadelphia’s location on the border of Pennsylvaniaand New Jersey may lead to a minor overestimate for New Jersey figures, as establishments thatwork in Philadelphia may be located in New Jersey. However, the fact that the ports of Newark andElizabeth handle more than twenty times that of Philadelphia makes any discrepancies minimal.Similar effects may be in play in the Gulf Coast ports as well, though the role of bulk goods, likeagricultural products and petroleum are most likely responsible.

4Note that even the ratios for major container port states include those marine cargo handlerswho handle bulk commodities and autos, which is a not insignificant portion of these ports’ activities.

146

for instance, is the largest volume shipping port in the Western Hemisphere, but

since it deals primarily in petroleum, farm products, steel, and chemicals rather than

containers, the per TEU level of employment is greatly inflated. And automation is

poised to decrease this ratio even further.

Still, this represents a sizable level of employment. For example, at this rate, the

new Maher Terminals terminal in Prince Rupert, British Columbia, is expected to

handle 500,000 TEUs and would employ 750 longshoremen at this rate. And the

Maersk-SeaLand terminal in Port Newark, which was predicted to handle 1.2 million

TEUs by 2028, would employ almost two thousand workers directly at the East Coast

average and 1,200 at the New Jersey rate.

6.4 Deep sea freight transportation

Deep sea freight transportation (SIC 4410 and NAICS 483111) refers to establishments

“primarily engaged in providing deep sea transportation of cargo to or from foreign

ports” (U.S. Dept. of Commerce, Bureau of the Census 2007). It thus refers to those

shipping companies that organize and provide the transport of goods between US and

foreign ports. These companies generally own ships and should be distinguished from

freight transportation arrangement establishments (SIC 4710 and NAICS 488510),

which will be discussed later.

Table 4.2 shows that deep sea freight handling has dropped to less than half

(11,217) of its 1970 level (23,919) nationwide. It has also dropped in proportion to

national employment, making up 0.04 percent of all employment in 1970 and less

than 0.01 percent today. This trend can be seen in more detail in Figure 6.2, which

Handling such cargo is more labor intensive than container handling. For instance, each auto has tobe driven by an individual onto or off the ship and earlier techniques for handling agricultural goodsinvolved smaller units requiring more labor power for unit of volume. Thus, these figures should betaken as upper end estimates.

147

shows the general decline in employment in deep sea freight transportation for five of

the major Northeast port states. These statistics offer an important indicator of the

decreasing benefits that deep sea freight transporters offer their host states.

1970 1980 1990 2000

04000

8000

12000

Totalem

ploymen

t


1970 1980 1990 2000

04000

8000

12000 California

OregonWashington

Figure 6.2: Total employment in deep sea freight transportion (SIC 4410 and NAICS483111) by state (East Coast left, West Coast right)Source: Country Business Patterns

There are three basic trends in the data. The most obvious is the rapid evaporation

of employment in New York State. This loss is a result of the port’s center of activity

shifting to the New Jersey waterfront, which accounts for that state’s steep rise in

the 1980s. Note, however, that New Jersey’s employment declines again rapidly after

1990 and converges toward the other ports, suggesting a minimal number of employees

required for contemporary freight providers. The second trend is the relatively even

keel of employment in Pennsylvania and Maryland, each of which (like New York and

New Jersey) have one main port. This supports the steady minimum necessary to

operate a freight operation. Finally, the third trend is the growth of employment in

Virginia, which has aggressively developed several new terminals since 1990.

If we consider these trends with respect to the total volume of cargo moving

through ports, the decline becomes even steeper. As is evident from figures 4.2, 4.1,

and 4.3, the total volume of containerized cargo is increasing in all but a few isolated

ports. Thus, for deep sea freight handling, where overall employment has decreased

148

1979 1984 1989 1994 1999 2004



Alabama 40.51 6.89 0.12 0.26 0.78Louisiana 14.42 8.19 4.90 4.99 3.07 2.21Texas 9.13 4.20 1.59 1.97 1.48 0.50USGC Average 12.52 5.99 2.66 2.87 1.75 0.73


Table 6.2: Employment in deep sea freight transportation by state per 1,000 TEUsmoved through ports in that state

precipitously, the per unit level of employment has dropped even further. Table 6.2

shows that average employment in deep sea freight transportation on the West Coast

has dropped almost 97 percent from 2.63 jobs per thousand TEUs to a mere 0.09;

that it has fallen 85 percent from 2.81 to 0.42 on the East Coast and half that in New

Jersey; and that it has sunk 94 percent from 12.52 to 0.73 on the Gulf Coast. Notable

exceptions include Florida, which posted major gains in the mid- to late-1990s and

Virginia, which has exhibited a slow, unsteady climb since 1990.

Overall, employment in firms that provide deep sea freight transportation has

plummeted since the 1970s. This can be attributed to three factors. First, technolog-

ical improvements in communications and data processing have reduced the number

of employees required to operate these firms. Second, increasing ship size has reduced

the relative size of administrative overhead. And third, consolidation among shipping

149

companies has further reduced administrative overhead, allowing the firms to trim

down over time.

6.5 Warehousing and storage

Warehousing and storage have been discussed in Section 4.3. That analysis showed

that warehousing employment has grown exponentially over the last four decades

and that it has grown in importance in the national economy as a whole. It also

demonstrated that warehousing has experienced a spatial deconcentration and recon-

centration, directing much activity to a band of warehousing a few hundred kilome-

ters inland from the ocean coasts. Closer inspection of employment in the five states

evaluated above reinforces our understanding of this pattern. Figure 6.3 shows that

state-level employment in general warehousing and storage has climbed steadily in all

five states from 1970 to the late 1990s and then steeply to 2007. The greatest gains

have been in Pennsylvania, which has exceptional interstate access and lies within the

inland band. Pennsylvania is now home to the largest number of warehousing and

storage workers on the East Coast.

Warehousing employment in terms of throughput has also been rising for reasons

discussed previously. On the West Coast, employment fell from an average of 2.5

employees per thousand TEUs in 1979 to just above one employee in the late 1990s

before tripling to roughly three employees per thousand TEUs. The East Coast follows

a similar pattern, declining in the late 1990s before tripling from roughly three to nine

employees per thousand TEUs. These per unit increases imply strong job creation

possibilities from port expansion in warehousing, though not necessarily near the port

itself.

Growth in the absolute number of warehousing jobs in the immediate vicinity of

150

1970 1980 1990 2000

020000

40000

60000

Totalem

ploymen

t


1970 1980 1990 2000

020000

40000

60000 California

OregonWashington

Figure 6.3: Total employment in general warehousing and storage (SIC 4225 and 4226and NAICS 493110) by state (East Coast left, West Coast right)Source: County Business Patterns

ports is evident in the total employment series of maps (Figures 4.4, 4.5, 4.6, 4.7,

4.8, 4.9, 4.10, and 4.11). For example, LA County shows the highest number of jobs

(over 1,000) in 1974, with the surrounding counties having less than 500. By 2007,

LA County has roughly 18,000 warehousing jobs and the surrounding counties have

over 1,000. Elizabeth and Newark, New Jersey, meanwhile, have fewer than 500 in

1974 and over 5,000 in 2007.

However, the relative importance of warehousing employment as a percentage of

total employment in the economies of these counties has remained steady and even

low relative to the national average. With minor and temporary exceptions, the loca-

tion quotient of counties proximate to ports remains consistently below 0.50 relative

to total employment in this sector for the nation as a whole. This suggests that ware-

housing employment plays only a minor role in these counties’ overall employment,

though it does not indicate whether or not the role it plays is critical to the success

of other economic activities in those counties.

The latter concern is countered by two factors. First, as mentioned previously

(Section 4.3.2), warehousing employment is determined primarily by proximity to

customer bases. The fact that these bases are geographically coincident with ports is

151

1979 1984 1989 1994 1999 2004





Table 6.3: Employment in general warehousing and storage by state per 1,000 TEUsmoved through ports in that state

a matter of historical development rather than contemporary need. Second, the in-

creasing deconcentration of warehousing employment would suggest that warehousing

is important to economic activity in general rather than port activity specifically.

Though warehousing plays a small role in a county’s economy relative to other

activities, it does still offer significant potential gains in both counties and states

that host ports in terms of the number of employees. Though policy makers in the

late 1990s could not have foreseen such astounding growth in this sector, it certainly

stands as one of the major benefits of port expansion during this period. The ques-

tion going forward is whether this trend will continue or whether the shift has been

completed. Trend lines show conflicting directions. While some states continued to

increase rapidly prior to the economic crisis of 2008, others had already begun to turn

downward.

152

6.6 Freight trucking

Freight trucking has also been discussed earlier (see Section 4.4). It was shown that

though trucking has experienced mild but steady growth, employment has dropped

relative to total employment at the national level, particularly from the late 1990s.

Analysis also showed that trucking has not shifted relative to infrastructure nodes

but, rather, is tied to population density and per capita income.

Figure 4.13 shows that this is the case for the five states being examined here.

Trucking employment has grown steadily but weakly, except for a decline in the late

1990s that seems to have levelled out. Again, for ports this does not seem to be a major

gain. The highest number of truckers work in the warehouse state of Pennsylvania,

establishing themselves inland, perhaps near the end of the day’s hauling and lower

cost housing. And trucking exhibits a pattern similar to that of general warehousing

and storage, in that over the time period examined here employment increases as one

moves further away from ports (see Figure 4.13).

Though much freight trucking is related to urban concentration (cf. Table C.2),

its ratio to container throughput is significantly higher than the other sectors explored

so far. On average, employment has dropped by three quarters to 23 employees per

thousand TEUs on the East Coast, by over ninety percent to eight employees per

thousand TEUs on the West Coast, and by more than half to eighty employees per

thousand TEUs on the Gulf Coast. A closer look at the East Coast reinforces the

trend observed in the previous paragraph. Trucking employment per unit is lowest in

New York and New Jersey, where the major port is located, and nearly fifty percent

higher in Pennsylvania, which hosts the greatest number of general freight truckers

on the East Coast. Additionally, though there were significant gains in New York and

New Jersey in the late 1980s and early 1990s, these states have experienced a much

153

steeper decline than Pennsylvania.

1979 1984 1989 1994 1999 2004





Table 6.4: Employment in general freight trucking by state per 1,000 TEUs movedthrough ports in that state

All in all, one is compelled to conclude that trucking as well does not contribute

significantly to the economic wellbeing of areas adjacent to ports. While providing

a significant number of jobs overall, the total number of jobs is declining, especially

near ports, and is shifting inland. Further, it is not clear that jobs in this sector are

as safe, steady, and well-paid as they could be (see Section 6.8 below).

6.7 Freight transportation arrangement

The U.S. Dept. of Commerce, Bureau of the Census (2007) defines freight transporta-

tion arrangement as “arranging transportation of freight between shippers and carri-

ers.” Personnel in this sector are often known “as freight forwarders, marine shipping

154

1970 1980 1990 2000

050000

100000

Totalem

ploymen

t


1970 1980 1990 2000

050000

100000

CaliforniaOregonWashington

Figure 6.4: Total employment in general freight trucking (SIC 4210 and NAICS484100 and 484200) by state (East Coast left, West Coast right)Source: County Business Patterns

agents, or customs brokers and offer a combination of services spanning transporta-

tion modes.” Some of the decline in deep sea freight transportation employment is

surely attributable to the expansion of the freight transportation arrangement sector,

which generally goes by the term “third party logistics” (3PL). Due to regulatory

restrictions on direct transportation providers, firms in this sector have until recently

been the only source for coordinating door-to-door shipments of goods. They were

the only firms (other than shippers themselves) that could legally coordinate ship-

ment across multiple modes of transportation. They thus developed expertise in this

kind of coordination, and their importance has grown with containerization, reflect-

ing the increasingly transactional nature of trade, transportation, and supply chain

management.

Employment in 3PL has increased by roughly 250 percent since 1974 from just

under 60,000 nationwide to over 210,000. As a proportion of national employment,

the sector has doubled from 0.09 percent to 0.18 percent of all employment. It is not

surprising that this sector grew faster than the national average, since globalization

requires additional work to coordinate transportation overseas as well as domestically.

Growth has been uneven, however. On the West Coast, almost all expansion has been

155

1970 1980 1990 2000

010000

20000

30000

Totalem

ploymen

t


1970 1980 1990 2000

010000

20000

30000 California

OregonWashington

Figure 6.5: Total employment in freight transportation arrangement (SIC 4710, 4723and NAICS 488510) by state (East Coast left, West Coast right)Source: County Business Patterns

1979 1984 1989 1994 1999 2004





Table 6.5: Employment in freight transportation arrangement by state per 1,000TEUs moved through ports in that state

in California, which dominates container shipping. On the East Coast, there has been

a fairly consistent increase across all states, but New York is only now beginning to

recover from a significant decline during the 1980s and early 1990s that coincided

156

1970 1979 1984 1989 1994 1999 2004

-0.1

5-0

.10

-0.0

50.

000.

050.

100.

15

port

1970 1979 1984 1989 1994 1999 2004

-0.1

5-0

.10

-0.0

50.

000.

050.

100.

15

airport

1970 1979 1984 1989 1994 1999 2004

-0.1

5-0

.10

-0.0

50.

000.

050.

100.

15

intermodal

Figure 6.6: Freight transportation arrangement (SIC 4710 and 4723 and NAICS488510): Regression coefficients for distance (in 100km) from closest port, airport,and intermodal terminal by year against the logarithm of employment. NAICS andSIC comparable.Point represents estimate. Wide line indicates the 50 percent confidence interval, and thin line represents the95 percent confidence interval for the estimate. Dotted grey lines indicate industrial classification transitions.Red line indicates trend in estimators.Source: County Business Patterns

with the purchase of the last American shipping companies by European and Asian

interests. Still, New York, the traditional seat of trade on the East Coast, employs

the greatest number in the sector in the United States.

This sector has, however, declined relative to throughput. On the West Coast,

employment has declined over eighty percent from 8.4 to 1.6 employees per thousand

TEUs. On the East Coast, employment has shrank from about ten to about four

employees per thousand TEUs. And on the Gulf Coast, employment has declined

from over eighteen to under twelve employees per thousand TEUs. Looking more

closely at the East Coast, one observes a general rise in the ratio in the late 1980s

and early 1990s in New Jersey, Maryland, and Pennsylvania followed by a forty to fifty

percent decline. Employment relative to throughput in New York, by comparison,

deteriorates steadily from over eleven jobs per thousand TEUs passing through the

Port of New York and New Jersey to under three today, and employment in Virginia

exhibits a slow slide from 3.5 in 1984 to 1.7 in 2004. Additionally, in the south, the

sector has grown immensely, particularly in Florida, which serves as a gateway to

Central and South America (Friedmann and Wolff 1982).

Together, these trends suggest that even as employment in this sector has been

157

rising, its locus has been shifting toward states that host other types of transporta-

tion nodes, like airports. Figure 6.6 implies that this shift has been toward airports

and perhaps away from ports, though again population density is the main and an

increasingly important driver of location (cf. Table C.3). Thus, the geographical shift

in employment is not interregional but intraregional.

6.8 Good jobs or bad?

When politicians and analysts speak of job creation, they seldom address the char-

acteristics of those jobs beyond some vague assurances of respectable wages. In the

context of port labor, which has traditionally been blue-collar work, and deregulation

in the transportation industry (see Chapter 3), this issue takes on additional im-

portance. A recent report (Bensman 2009) argues that since the federal government

deregulated the trucking industry in 1980 through the Federal Motor Carriers Act, the

quality and safety of trucking employment has deteriorated significantly. Bensman

(2009) argues that an exploitative form of independent contracting between trucking

service providers and individual truckers that resembles contingent hiring more than

it does independent contracting has led to inefficiencies, environmental damage, more

dangerous highways, enormous public costs, and a degrading of port trucking jobs.

Truckers are now expected to own or lease and maintain their own trucks in which

they work on average for ten to twelve hours a day five days a week and to cover their

own medical expenses on an average annual income of $28,000. In the line of work,

they are compelled to wait long hours among lung-damaging diesel fumes to be sure

they will have work for the day, to often haul trailers on chassis they know are unsafe,

and to discover sometimes that their loads are improperly insured or overweight when

accidents do happen, often leading to personal bankruptcy.

158

In warehousing, the situation is no better. Most workers are contingent work-

ers who start at $9 per hour with no benefits and can hope at best to step up to

a permanent position at $15 per hour and no benefits. Additionally, the work is

very insecure. There are seasonal fluctuations with hiring peaks in advance of the

Christmas season and slumps after the holidays. There are also frequent closures,

as unprofitable operations are shut down and new facilities are constructed alongside

newly built infrastructure (Bensman 2008).

Benefits are also elusive in the one time bastion of labor power, longshoring.

Bensman (2008) reports that the Waterfront Commission of New York Harbor has

stated that one third of longshoremen hired since 2003 have not worked enough hours

to qualify for benefits. This is a result of declining strength and standards, especially

“on the East Coast where the ILA was forced to accept a nine-tier contract that

has reduced pay levels and reintroduced casual work for thousands of recent hires”

(Bensman 2008, 7).

The overall trend appears fairly clear cut: job quality, pay rates, benefits, and

security are all decreasing in logistics-related employment. In the sectors growing

fastest, port trucking and warehousing, conditions are deteriorating the most rapidly.

Though there are surely exceptions within each sector, jobs in the logistics industry

do not appear to be “good” jobs.

6.9 Conclusion

The preceding analyses call into question ports as effective generators of local, directly

related employment. Longshoring work has dropped toward half its level forty years

ago, and the trend toward increasing automation will continue to drive employment

still lower. Deep sea freight transportation has declined by fifty percent and almost

159

disappeared relative to total freight volume. While warehousing jobs have increased

signficantly, they have become less concentrated and play a smaller and smaller role in

those local economies that host ports. Finally, trucking employment has experienced

some mild growth, but the primary gains are in areas distant from the ports themselves

and the currently degraded quality of these jobs would suggest that they are not jobs

that local governments should be aggressively seeking to attract through subsidies

to terminal operators and shipping companies. Together, the declining levels and

quality of employment in these four areas of direct employment make port authority

and politicians’ claims that port subsidization will create jobs overly sanguine.

One bright note in the sector may be that of freight transportation arrangement.

This sector has been growing faster than national employment as a whole, though it,

too, is declining relative to total freight volume.

Sector WCNA ECNA USGCRatio Inc/Dec Ratio Inc/Dec Ratio Inc/Dec

Port operations and marine cargo handling 1.3 ⇑ 1.7 ⇓ 4.6 ⇓Deep sea freight transportation 0.1 ⇓ 0.4 ⇓ 0.7 ⇓General warehousing and storage 3.1 ⇑ 9.2 ⇑ 18.8 ⇑General freight trucking 7.9 ⇓ 22.7 ⇓ 81.1 ⇓Freight transportation arrangement 1.6 ⇓ 4.0 ⇓ 11.7 ⇓

Total 14.0 39.0 116.9

Arrows indicate whether the trend is increasing or decreasing.

Table 6.6: Estimated ratio of jobs per 1,000 TEUs of container traffic by sector

Table 6.6 summarizes the current ratio of employment by sector for every thousand

TEUs passing through ports in the three US port ranges and indicates the current

trend. At present, all sectors are declining relative to total freight throughput except

warehousing, which has been increasing as the final stages of production are shifted

into the logistics chain. The totals in this table provide a loose indication of how

many directly related jobs can be attributed to each thousand TEUs moved through

a given port range. It suggests that five million TEUs (roughly the total volume

160

in the Port of New York and New Jersey) coincide with 70,000 jobs on the West

Coast, 195,000 jobs on the East Coast, and 600,000 jobs on the Gulf Coast. One

half to two thirds of these jobs are in trucking. Of course, these figures should be

considered maximal employment generation figures, as bulk commodities, which are

not included in these calculations, are responsible for a significant proportion of jobs

in these sectors, especially on the Gulf Coast.

These are clearly not insignificant numbers. Three additional caveats are in or-

der, however. First, these jobs are distributed (and redistributed) throughout a given

port’s hinterland, and many of them are moving inland to areas where land and labor

are less expensive. This reduces the economic impact on areas proximate to ports.

Second, as Table 6.6 indicates, the job creation rate is moving downward. Thus, ex-

cept for warehousing, job generation estimates made on the basis of static coefficients

are likely to be wildly optimistic. Going forward, cost-benefit analyses must account

for these probable shifts and more actively develop projection ranges, as Flyvbjerg,

Bruzelius, and Rothengatter (2003) suggest. Third, deregulation, decreasing union-

ization, and loss of union power5 have degraded the quality of these jobs. They are

increasingly lower paid, more dangerous, and less secure.

We can conclude, then, that direct employment at the port is no longer the signif-

icant regional economic driver it once was. While quantitative expansion of capacity

can create new jobs, increases appear to be producing diminishing marginal returns

and diminishing job quality.

5This often occurs through processes of interport competition in the context of increasing supplychain flexibility, such as the case illustrated in Chapter 5.

161

Chapter 7

Decoupling

7.1 Port-related employment through localization

economies

The third argument for subsidizing ports is that they sustain and foster local economic

activity. That is, that port activities shape the economic terrain by attracting other

economic activites. The argument, borne out by the nineteenth century city, is that

the presence of port facilities serves as an anchor for localization economies, i.e., the

benefits of locating in the proximity of a specific facility (Henderson 1983; Rodrigue,

Comtois, and Slack 2006). As a result of reduced costs generated by ready access to

transshipment points, firms are expected to be more efficient, more competitive, and

thence more successful, providing employment and revenue for the local community.

In this chapter, the argument in favor of localization economies drawn from

Charles Horton Cooley’s work is tested against the neo-Marshallian argument that

economic activity has decoupled from transportation infrastructure nodes. Having

show in Chapter 4 that the U.S.’s transportation infrastructure has shifted geograph-

162

ically, I now introduce spatial regressions predicting the location of a representative

selection of the country’s top exporting industries. The results suggest a decoupling

of economic activity from transportation nodes, leading to the conclusion that ports

no longer serve as anchors for non-port employment.

7.1.1 Industry types and transportation

Glaeser and Kohlhase (2003) have recently called this relationship into question, how-

ever. They argue that transportation costs for cargo have dropped so significantly

that economists no longer need to treat them as determinants in the location of

economic activity. Citing the long-term trend of declining transportation costs in

the U.S. over the course of the twentieth century and the nation’s shift away from

manufacturing and natural resource production toward services, they argue that re-

cent “new economic geography” models like Krugman (1991) and Fujita, Krugman,

and Venables (1999) are not relevant for the current socioeconomic situation because

they perpetuate the emphasis on immovable natural resources made by early loca-

tion theorists (e.g., Losch [1954] and Weber and Friedrich [1965]). Instead, following

Marshall’s (1890) suggestion that industrial agglomeration is a result of information

spillovers and concentrated labor pools, Glaeser and Kohlhase argue that only those

transportation costs that enhance or inhibit personal transportation, like subsidies

for public transportation or gasoline taxes, are relevant as locational determinants of

economic activity.

McCann and Shefer (2004) introduce concepts from economic sociology and eco-

nomic geography proper to argue for a more subtle approach to transportation costs

than Glaeser and Kohlhase’s “neo-Marshallian approach.” They develop three ideal

types of industrial agglomerations based on the distinction made initially by Powell

163

(1990) that firms can be organized into markets, hierarchies, or networks. Powell

breaks down economic organization forms primarily through an analysis of transac-

tion costs (cf. Williamson [1975]). Economic activity can be organized as markets of

atomistic actors who communicate and interact through prices. Alternatively, market

activities can be organized into a hierarchy that persists over time to coordinate pro-

duction and exchange. Between these two extremes, Powell places networks consisting

of interdependent economic actors that coordinate their activities with multiple col-

laborators with whom they have developed trustworthy relations through repeated

interaction. McCann and Shefer (2004) mirror these three types of organization in

categorizing spatial clustering. In the market case, which they call “pure agglomera-

tion,” the critical transportation costs are those related to urban commuting as small

atomistic firms are dependent on workers’ ability to get to the workplace. Spatially,

pure agglomerations will be coupled to local urban infrastructure rather than inter-

regional infrastructure. In the case of hierarchical organization of economic activity

into large firms that locate in industrial complexes, traditional transportation costs

for goods are most relevant due to the high volume of inputs and outputs to the

production process. In the case of “social networks,” like the industrial districts of

Emiglia-Romagna (For example, Brusco 1982), transportation costs have no relevance

since the model depends on social infrastructure rather than physical infrastructure,

though spatial proximity is assumed to foster the trust necessary for efficient func-

tioning.

Following these authors’ lines of thought, we can distinguish at least six sectors

whose spatial agglomeration tendencies should be evaluated separately: natural re-

source extraction, heavy industry, networked light and medium industry, information-

oriented work, market-oriented light and medium industry, and retail. McCann and

Shefer’s (2004) industrial complexes can be divided into two subcategories. The first

164

is natural resource extraction, in which a few concentrated firms develop industrial

complexes to extract natural resources, like mining. The second is heavy industry,

which processes raw materials in large quantities at capital-intensive industrial com-

plexes. Similarly, social networks can be broken up into at least two subcategories.

The first is networked light and medium manufacturing, which is more likely to display

characteristics of industrial districts and consists of a larger number of smaller firms

than heavy industry producing more customized goods. The second is information

industries, which provide services rather than goods Glaeser and Kohlhase’s (2003)

and also tends to consist of a larger number of smaller firms. The pure agglomeration

category can also be divided into two subcategories. The first is market-oriented light

and medium industry, which is distinguished from networked industries by its focus

on simpler, more standardized products that are more conducive to market relations

since the products and the prices are more readily comparable (Gereffi, Humphrey,

and Sturgeon 2005; Sturgeon 2002). The second is retail, which is dependent primarily

upon urban agglomeration.

7.1.2 Dialectics of transportation

Due to their reliance on mathematical abstraction, these approaches unfortunately

tend to overlook the sociospatial organization of transportation itself as an active

force in their discussions. Rather, for these economic approaches, transportation

is represented simply by transportation costs over distance and time (Rietveld and

Vickerman 2004). Vested interests are overlooked and spatially concrete hindrances

like intermodal transfers of cargo are considered simply in terms of costs. For all

their spatiality, these approaches conceptualize mathematical fields of agglomeration

rather than the concrete mechanics of those agglomerations. Hence, transport theory

165

tends to treat transportation as a derived demand that responds to and facilitates

other agglomerative forces. Except perhaps for the case of road congestion, the role

of transportation in shaping the economic terrain by inducing demand is generally

disregarded (Rietveld and Vickerman 2004).

It thus seems useful to return to the early theorizations of one of the founders

of the Chicago School of Sociology, Charles Horton Cooley. Cooley emphasized the

dialectical relation between transportation as a mechanical force and the social forces

that developed it. “Precisely because transportation underlies social development it is

in turn determined by that development. It is a tool of the economic, the political, the

military organizations, and the character of the tool varies with their needs” (Cooley

1894, 41). For Cooley, as this quote implies, though social forces (in addition to

natural ones) shape transportation systems, the transportation system underlies and

integrates society. He has in mind one simple principle: “Population and wealth tend

to collect wherever there is a break in transportation” (Cooley 1894, 91, emph. in

original).

By ‘break’ Cooley (1894, 91) means any physical interruption in the movement of

goods that is sufficient to result in the transfer of goods and their temporary storage.

He then divides breaks into two categories: mechanical and commercial. A mechan-

ical break is a purely physical interruption in the flow of goods, while a commercial

break combines the mechanical break with a change in ownership of the goods. Be-

cause commercial breaks involve many other auxiliary services to facilitate economic

exchange, they tend to be the sites of greatest agglomeration. This conceptualization

accounts for the growth of great port cities like New York, Venice, and Amsterdam,

which combined the mechanical break between sea and land with innovative finance

and banking to grow into world economic superpowers (Braudel 1992). Note that this

view is not incompatible with a transportation costs approach. By locating in prox-

166

imity to mechanical breaks in transportation, firms are often able to reduce their costs

by minimizing the amount of handling involved in transferring cargo from one mode

to another, often the largest proportion of transportation costs. Cooley’s approach,

however, emphasizes how transportation acts as an agglomerative force, rather than

focusing on cost reduction.

As Chapter 4 demonstrated, transportation has changed dramatically since Coo-

ley’s time. From a port perspective, in the late 1800s the labor-intensive transfer of

individually stowed goods from sea to land and vice versa generated a major mechan-

ical break that encouraged agglomeration (Glaeser and Kohlhase 2003). As discussed

in the earlier chapter on technological change (Chapter 3), containerization has re-

moved the obligation to break bulk on the docks. Instead, within minutes or hours,

goods can continue virtually uninterrupted toward inland destinations where they can

be unpacked, repacked, and redistributed. The growth of inland distribution centers

(DCs) and warehousing concentrations is well documented (e.g., Notteboom and

Rodrigue 2005) and Section 4.3 illustrates this shift. Over the past three decades,

warehousing has spread out more evenly across U.S. counties and formed a concen-

trated band a few hundred kilometers inland and parallel to the coast. Thus, the

break point for intermodal cargo transfers has moved from the nation’s ports to in-

land warehousing districts.

So, if Cooley is correct, the tight coupling between breaking bulk and other eco-

nomic activity implies that agglomeration should be occuring around the new concen-

trations of warehousing activity and shifting away from ports. If the neo-Marshallians

are correct and economic activity has decoupled from its transportation base, eco-

nomic activity should not follow the restructuring of the transportation network. In

the following pages, I explore these two opposing hypotheses in relation to the six in-

dustrial sectors (i.e., natural resource extraction, heavy industry, networked light and

167

medium industry, information industries, market-oriented light and medium manu-

facturing, and retail1).

Neo-Marshallians Cooley

Natural resources Decoupled DecoupledHeavy industry Decoupled De/coupled

Networked manufacturing Decoupled CoupledMarket manufacturing Decoupled CoupledInformation industry Decoupled Decoupled

Retail Decoupled Decoupled

Table 7.1: Relationship of industrial sectors to transportation following Marshall andCooley

Table 7.1 summarizes how these two approaches would consider the relation of

broad industrial classifications to break bulk points in the logistics infrastructure.

They thus generate a series of testable hypotheses that will be investigated in the

following sections. Natural resources, as a product of the land or sea, are inherently

immobile and should thus be decoupled from the transportation infrastructure for

both the neo-Marshallians and Cooley. Heavy industry should also be decoupled

from the transportation system according to the neo-Marshallian perspective, since

transportation costs are no longer important in location decisions. From the Coolean

perspective, however, there is an aspect of uncertainty. The enormous sunk costs

required of these industries lends a degree of historical inertia to contemporary lo-

cations not unlike that of natural resources. This inertia makes it highly unlikely

that they would move to new locations to take advantage of changing transportation

infrastructure. However, the location of new facilities could still be coupled to break

bulk points for those industries that do not significantly reduce the bulk of the raw

1Though Cooley does not consider this, retail also serves a break-bulk point. It serves not onlyas a mechanical break where truckloads of goods are broken up and distributed to customers butalso as a commercial break where ownership is transferred from the retailer to the consumer.

168

materials. Generally, manufacturing will locate near the source of raw materials if

the bulk or weight is greatly reduced in processing, but it is otherwise free to locate

elsewhere. Cooley would suggest that “elsewhere” should be mechanical break points.

Networked manufacturing also should be decoupled in the neo-Marshallian view, but

for Cooley there should be a distinct coupling, as manufacturing spatially repositions

itself to take advantage of cost savings near new mechanical breaks by reducing the

number of modal transfers. Market-oriented manufacturing should follow the same

logic as networked manufacturing, though it may be expected to shift more quickly as

it is not held in place by organizational interdependencies. Fifth, the information in-

dustry, as it does not depend on the movement of goods in any significant way should

be decoupled for both neo-Marshallians and Cooley. Finally, retail should remain

decoupled from the interregional transportation network from the neo-Marshallian

position. Rather, retail should be coupled only to the intraregional transportation

network, but this is not the scale addressed in this paper. From Cooley’s position,

though he does not identify it himself, retail functions as a mechanical and commer-

cial break in the transportation system as goods are transferred from trucks that

deliver them to the store and the customers who take ownership of them. Thus, retail

should be tightly coupled with concentrations of population but not to the changing

transportation network. The primary area of disagreement between the two schools

of thought, then, is with regard to manufacturing.

7.2 Sector analyses

In the following sections, sample industries are drawn from each of the six categories

described above. They are selected on the basis of four criteria. First, they are

believed to be fairly representative of their respective fields with respect to indus-

169

trial structure. Second, they have consistent definitions under both SIC and NAICS

coding across the data employed, making the numbers intertemporally comparable.

Third, those sectors that offer tradeable goods (natural resource extraction, heavy

industry, and manufacturing) as opposed to generally non-tradeable services (retail

and information industries) are amost entirely listed by the International Trade As-

sociation (2010) as exporting industries. Industries selected represent a sampling

from the export rankings by value, as shown in Table 7.2. At least three of them

(semiconductors, pharmaceuticals, and petroleum products) are from the top twenty

exporting industries. As exporting industries are more likely to require the services

offered by container ships and airlines, they should have initially located proximate

to these facilities. Finally, when possible (one case), industries were selected from

Table B.1, which represent industries traditionally understood to be directly related

to port clusters.

7.2.1 Natural resource extraction

Natural resources behave as one would expect: they are completely indifferent to the

location of the wider transportation network. Because they are tied to the physical

composition of the planet rather than its social composition and remain immobile,

changes in the transportation infrastructure should have no impact on the location

of natural resource employment.

The industry selected here is iron ore mining (SIC 1010 and NAICS 212210).

Iron ore mining comprises “establishments primarily engaged in (1) developing mine

sites, mining, and/or beneficiating (i.e., preparing) iron ores and manganiferous ores

valued chiefly for their iron content and/or (2) producing sinter iron ore (except iron

ore produced in iron and steel mills) and other iron ore agglomerates” (U.S. Dept. of

170

Industry NAICS code Value Rank(USD millions) (out of 452)

Natural resourcesIron ore mining 212210 246 295

Heavy industryFlat container manufacturing 327211 816 168Glass container manufacturing 327213 182 330Cement, hydraulic 327310 68 391Ready-mix concrete manufacturing 327320 0 452Petroleum refineries 324110 8894 16

RetailBook, periodical, and music stores 451211–2, 451220 NA NA

Networked manufacturingPharmaceutical preparations 325412 9425 14Electroplating, plating, polishing, anodizing,and coloring

332813 NA NA

Machine tool (metal forming types) manufac-turing

333513 1248 126

Automatic environmental control manufac-turing for residential, commercial, and appli-ance use

334512 340 262

Semiconductors and related device manufac-turing

334413 59223 1

Market-oriented manufacturingAdhesive manufacturing 325520 804 172Gum and wood chemicals 325191 163 345Turned product and screw, nut, and boltmanufacturing

332720 NA NA

Information industriesPeriodical publishers 511120 NA NAData processing services 514210 NA NAMotion picture and video production 512110 NA NA

Source: Table 41. U.S. Total Exports, 1998–2003 from International Trade Association (2010).

Table 7.2: Selected industries ranked by value of exports in 2000

Commerce, Bureau of the Census 2007).

This industry has indeed remained decoupled from transportation nodes. The ut-

ter lack of relation to non-physical factors is evident in the lack of signficant estimators

in the regression equation below (Table C.4), and the low adjusted R-square values

imply that not only is mining not directly related to income or race but also not linked

to nodes in the transportation network. The only variable of significance is the slight

relation to airports. Iron mining employment is marginally more likely to be located

near customs landing airports than away from them; employment drops by a quarter

171

1970 1979 1984 1989 1994 1999 2004

-0.0

6-0

.04

-0.0

20.

000.

020.

04

port

1970 1979 1984 1989 1994 1999 2004

-0.0

6-0

.04

-0.0

20.

000.

020.

04

airport

1970 1979 1984 1989 1994 1999 2004

-0.0

6-0

.04

-0.0

20.

000.

020.

04

intermodal

Figure 7.1: Mining, iron ore (SIC 1010 and NAICS 212210): Regression coefficientsfor distance (in 100km) from closest port, airport, and intermodal terminal by yearagainst the logarithm of employment.Point represents estimate. Wide line indicates the 50 percent confidence interval, and thin line represents the95 percent confidence interval for the estimate. Dotted grey lines indicate industrial classification transitions.Red line indicates trend in estimators.

of a percent for every thousand kilometers one gets from such an airport. However,

this relation remains unaltered over the time frame considered here and probably has

more to do with historical coincidence than deliberate location decisions. No relation

to ports is demonstrated.

We can thus conclude that the identical neo-Marshallian and Coolean perspectives

are supported here. Natural resources have remained decoupled from the break bulk

points of the transportation network.

7.2.2 Heavy industry

Under the broad category of heavy industry, it is necessary to distinguish between

two types of products: bulk and general. General commodities can be containerized

and shipped. Bulk products are generally not containerized, though there are recent

increases in the containerization of bulk cargo (Rodrigue et al. 2006). This is im-

portant as the two categories demonstrate different changing relations to ports. Two

industries are selected for this section.

The first of these industries is petroleum refining, which de Langen (2007) cites as

a core port industry and which is one of the U.S.’s top export industries. A relatively

172

concentrated industry with 242 establishments and over 65,000 employees in 1997,

petroleum refining is defined as those establishments “engaged in producing gasoline,

kerosene, distillate fuel oils, residual fuel oils, and lubricants, through fractionation

or straight distillation of crude oil, redistillation of unfinished petroleum derivatives,

cracking or other processes. Establishments of this industry also produce aliphatic and

aromatic chemicals as byproducts” (U.S. Dept. of Commerce, Bureau of the Census

2007). This is a bulk shipping industry that does not lend itself to containerization.

Its products are more likely to be transported via pipelines or tanker trucks.

It will be seen from the regression (Table C.5) that despite the fact that refineries

are typically considered LULUs (locally undesirable land uses), refineries and urban

populations are coming into closer contact. Social variables other than population

density have little bearing on the location of refining, suggesting that the deciding

factor may be more geographically based. Lending support to this finding is that the

estimators for airports show that refinery employment is slightly more likely to be

located closer to airports and that it is moving closer to ports. The decline in refinery

employment as one moves away from ports has been cut in half from its 1979 level of

0.05 percent for every 100 kilometers. Together this suggests that de Langen’s claim

has some veracity to it and that bulk commodities are not following the shift toward

warehousing centers.

The second industry selected is glass container manufacturing, a concentrated

exporting industry with 61 establishments and over 20,000 employees in 1997 that

produces “glass containers for commercial packing and bottling, and for home can-

ning” (U.S. Dept. of Commerce, Bureau of the Census 2007) and as a non-bulk

good is more likely to use containerized shipping. This concentrated industry is only

marginally related to population density and has no statistically significant relation to

any of the infrastructural nodes studied. Representative of other container-oriented

173

1970 1979 1984 1989 1994 1999 2004

-0.2

-0.1

0.0

0.1

port

1970 1979 1984 1989 1994 1999 2004

-0.2

-0.1

0.0

0.1

airport

1970 1979 1984 1989 1994 1999 2004

-0.2

-0.1

0.0

0.1

intermodal

Figure 7.2: Petroleum refineries (SIC 2911 and NAICS 324110): Regression coeffi-cients for distance (in 100km) from closest port, airport, and intermodal terminal byyear against the logarithm of employment.Point represents estimate. Wide line indicates the 50 percent confidence interval, and thin line represents the95 percent confidence interval for the estimate. Dotted grey lines indicate industrial classification transitions.Red line indicates trend in estimators.

1970 1979 1984 1989 1994 1999 2004

-0.0

50.

000.

05

port

1970 1979 1984 1989 1994 1999 2004

-0.0

50.

000.

05airport

1970 1979 1984 1989 1994 1999 2004

-0.0

50.

000.

05

intermodal

Figure 7.3: Glass container manufacturing (SIC 3221 and NAICS 327213): Regressioncoefficients for distance (in 100km) from closest port, airport, and intermodal terminalby year against the logarithm of employment.Point represents estimate. Wide line indicates the 50 percent confidence interval, and thin line represents the95 percent confidence interval for the estimate. Dotted grey lines indicate industrial classification transitions.Red line indicates trend in estimators.

industries examined, these findings suggest that the sunk investments involved cause

these industries to behave like natural resources and remain decoupled from the wider

transportation infrastructure.

7.2.3 Manufacturing

Manufacturing covers a wide variety of industrial sectors, each with its own physical

requirements and historical trajectory. Thus, generalizing across sectors must only

be done with tentative footsteps. Nevertheless, the examples that follow do present

a relatively clear pattern of locational behavior.

174

Networked manufacturing

Networked manufacturing refers to those sectors that are likely to collaborate with

providers of related goods or services to collectively produce a final product. Though

their small size might simplify relocation, networks fostered through spatial proximity

may slow this process. To capture the breadth of industry, an handful of sectors were

selected and analyzed. First, under the broad classification of chemical manufactur-

ing falls pharmaceutical preparations (SIC 2834 and NAICS 325412), which refers

to “establishments primarily engaged in manufacturing in-vivo diagnostic substances

and pharmaceutical preparations (except biological) intended for internal and external

consumption in dose forms, such as ampoules, tablets, capsules, vials, ointments, pow-

ders, solutions, and suspensions,” constitutes a large sector with a moderate number

of establishments (838) and well over 100,000 employments in 1997. These estab-

lishments produce a mixture of standard and non-standard, property right protected

goods and represent one of the largest exporting industries in the U.S. The next two in-

dustries are in either primary metal manufacturing (NAICS 332) or fabricated metal

product manufacturing (NAICS 333). The first of these is electroplating, plating,

polishing, anodizing, and coloring (SIC 3471 and NAICS 332813), which comprises

“establishments primarily engaged in electroplating, plating, anodizing, coloring, buff-

ing, polishing, cleaning, and sandblasting metals and metal products for the trade.”

This sector had a large number of establishments (over 3,400) and almost 75,000 em-

ployees in 1997. Though this sector provides a non-tradeable service to other metal

working firms and thus does not export, it may still choose to locate near break

points to more effectively receive and send goods for electroplating. Machine tool

(metal forming types) manufacturing (SIC 3542 and NAICS 333513), on the other

hand, which refers to establishments“primarily engaged in manufacturing metal form-

175

ing machine tools (except handtools), such as punching, sheering, bending, forming,

pressing, forging and die-casting machines,” constitutes a fairly small but deconcen-

trated industry, with 225 establishments and over 14,000 employees in 1997. It is

also a major exporter producing goods that are highly customized and thus likely to

incorporate inputs for interdependent businesses. The final duo examined come from

computer and electronic product manufacturing (NAICS 334). The first is automatic

environmental control manufacturing for residential, commercial, and appliance use

(SIC 3822 and NAICS 334512), which comprises “establishments primarily engaged

in manufacturing automatic controls and regulators for applications, such as heating,

air-conditioning, refrigeration and appliances” (thermostats, basically). This sector is

a significant exporter and is a fairly small but deconcentrated industry, with just over

300 establishments and over 20,000 employees in 1997. The second sector is semi-

conductors and related device manufacturing (SIC 3674 and NAICS 334413), which

comprises “establishments primarily engaged in manufacturing semiconductors and

related solid state devices. Examples of products made by these establishments are

integrated circuits, memory chips, microprocessors, diodes, transistors, solar cells and

other optoelectronic devices,” which is a large and moderately concentrated industry

with roughly 1,100 establishments and just under 200,000 employees in 1997. Saxe-

nian (1996) and Sturgeon (2003) have shown that this sector can rely very much on

networked production, especially in California.

Market-oriented manufacturing

Market-oriented manufacturing tends to produce standardized products that compete

in purer markets on the basis of price. These establishments are less tied to other firms

and are therefore more capable of picking up and moving to take advantage of changes

in transportation costs due to changing spatial organization of the transportation

176

system.

Three sectors were selected to represent this broad category. The first two sectors

represent a variety of pursuits within the broader classification of chemical manufac-

turing. First, adhesive manufacturing (SIC 2891 and NAICS 325520), which refers

to those establishments primarily engaged in manufacturing adhesives, glues, and

caulking compounds, is a moderately unconcentrated industry with about 700 estab-

lishments and 20,000 employees in 1997 and an export market in 2000 of over $800

million. Second, gum and wood chemicals manufacturing (SIC 2860/2861 and NAICS

325191), which comprises “establishments primarily engaged in (1) distilling wood or

gum into products, such as tall oil and wood distillates, and (2) manufacturing wood

or gum chemicals, such as naval stores, natural tanning materials, charcoal briquettes,

and charcoal (except activated),” is a fairly concentrated industry with 63 establish-

ments and about 2,200 employees in 1997. It is also an exporter, but much smaller

by comparison. The third sector chosen is one that perhaps most exemplifies the

standardized product: turned product and screw, nut, and bolt manufacturing (SIC

3450 and NAICS 332720). This industry refers to “establishments primarily engaged

in (1) machining precision turned products or (2) manufacturing metal bolts, nuts,

screws, rivets, and other industrial fasteners. Included in this industry are estab-

lishments primarily engaged manufacturing parts for machinery and equipment on a

customized basis.” This large, deconcentrated sector has many establishments (3,785)

and employees (133,399) in 1997 has no export market, perhaps indicative of the ease,

low cost, and capital-intensity of producing these items.

Findings

It is not necessary to go into detail for each of these industries, as the patterns are

fairly consistent across them. The results can be found in Figures 7.4, 7.5, 7.6, 7.7,

177

7.8, 7.9, 7.10, and 7.11, and Tables C.10, C.11, C.13, C.14, C.15, C.8, C.9, and C.12.

Consistent with all the other regression analyses, the dominant locational determi-

nant is population density. This predictor is always significant in manufacturing and,

for those regressions with some degree of explanatory power (R-square greater than

0.2), impacts employment in the given sector by a quarter to a half a percent increase

for every percent density increases. So more urbanized counties have greater levels

of manufacturing employment. Income per capita does not have the same relation

found in other sectors. While it was often significant prior to the 1980s, it is now no

longer so. Only in two of the sectors did it matter: adhesives manufacturing, where it

was highly positive, and gum and wood chemicals manufacturing, where it was nega-

tive. Tax rates, contrary to many studies, seem relatively unimportant overall. In the

few cases where the rates showed significance, it was slight, less than a 0.02 percent

increase or decrease in employment per percent of tax. The other social indicators

behave generally as one would expect. Education was of some significance in pharma-

ceutical preparations and semiconductors, which are more knowledge-intensive sectors

than the others. In the rare occasions where it was significant, nonwhite population

led to a hair’s width increase in employment in the given industry. And the percent

of the foreign population had a moderately positive correlation with employment in

the given sector, ranging from 0.01 to 0.07 percent increase for every percent of the

population that was foreign.

Proximity to infrastructure unexpectedly demonstrates little significance in indus-

trial location. Of varying significance, the coefficients for ports range generally from

-0.02 to 0.02 percent increase in industry employment for every 100km the county

centroid is from the nearest container port. This implies a 0.2 percent increase as one

travels from New York City to Detroit (roughly 1,000km). If there is any trend with

regard to ports, it is a slight movement away, which would support Cooley’s view.

178

However, concluding thus would be stretching the data very close to its breaking point

and is thus unsupportable.

The relation to airports is more interesting. While almost always insignificant due

to very large standard errors, the point estimates fairly consistently demonstrate a

downward trend. This suggests that manufacturing activities are increasingly locating

closer to airports. That said, as the estimates are generally moving from positive

values toward zero, the data suggests that manufacturing is losing its relationship to

airports, i.e., it is deconcentrating. This would also suggest a decoupling from the

infrastructural nodes.

Finally, the relation of manufacturing employment to intermodal terminals is gen-

erally insignificant and unchanging. Several sectors (pharmaceutical preparations and

environmental controls) may be moving away from a previous attachment to these

nodes and gum and wood chemicals manufacturing appears to be deconcentrating

relative to them, but the remaining five sectors show an unchanging, insignificant

relation to intermodal terminals. Again, we are led to conclude that there is no

statistical correlation between intermodal terminals and manufacturing employment.

In sum, the analyses here suggest that manufacturing employment has not been

coupled to the infrastructure network during the containerization era. The possibility

that change occured between the first containerized shipment in 1956 and 1970 is ex-

tremely unlikely. It thus appears that Cooley’s view has been untenable for much of

the last century if not longer. It also calls the neo-Marshallian perspective into ques-

tion as well, however. It is generally believed that the shift away from manufacturing

toward services occured after 1970. If the regressions completed here accurately re-

flect industrial location over the past four decades,2 then the shift to services is not

2Future refinements of the data include accounting for spatial correlation through county ad-jacency matrices, multilevel modeling, incorporating additional sectors and perhaps aggregating asappropriate, and incorporating highway mileage or exits, though the latter would have problematic

179

1970 1979 1984 1989 1994 1999 2004

-0.0

50.

000.

050.

100.

15

port

1970 1979 1984 1989 1994 1999 2004

-0.0

50.

000.

050.

100.

15

airport

1970 1979 1984 1989 1994 1999 2004

-0.0

50.

000.

050.

100.

15

intermodal

Figure 7.4: Pharmaceutical preparations (SIC 2834 and NAICS 325412): Regressioncoefficients for distance (in 100km) from closest port, airport, and intermodal terminalby year against the logarithm of employment.Point represents estimate. Wide line indicates the 50 percent confidence interval, and thin line represents the95 percent confidence interval for the estimate. Dotted grey lines indicate industrial classification transitions.Red line indicates trend in estimators.

1970 1979 1984 1989 1994 1999 2004

-0.1

0-0

.05

0.00

0.05

0.10

0.15

port

1970 1979 1984 1989 1994 1999 2004

-0.1

0-0

.05

0.00

0.05

0.10

0.15

airport

1970 1979 1984 1989 1994 1999 2004

-0.1

0-0

.05

0.00

0.05

0.10

0.15

intermodal

Figure 7.5: Electroplating, plating, polishing, anodizing, and coloring (SIC 3471 andNAICS 332813): Regression coefficients for distance (in 100km) from closest port,airport, and intermodal terminal by year against the logarithm of employment.Point represents estimate. Wide line indicates the 50 percent confidence interval, and thin line represents the95 percent confidence interval for the estimate. Dotted grey lines indicate industrial classification transitions.Red line indicates trend in estimators.

responsible for the decoupling.

7.2.4 Retail

For the retail sector, the broad grouping of book, periodical, and music stores was

selected. The definitions and hence industrial coding for many other goods that are

more likely to be shipped by boat, like televisions, have changed over the period

considered here and thus cannot be evaluated. Book, periodical, and music store

establishments, which are “primarily engaged in retailing new books, newspapers,

magazines, and prerecorded audio and video media”(U.S. Dept. of Commerce, Bureau

correlations with population density.

180

1970 1979 1984 1989 1994 1999 2004

-0.0

50.

000.

050.

10

port

1970 1979 1984 1989 1994 1999 2004

-0.0

50.

000.

050.

10

airport

1970 1979 1984 1989 1994 1999 2004

-0.0

50.

000.

050.

10

intermodal

Figure 7.6: Machine tool (metal forming types) manufacturing (SIC 3542 and NAICS333513): Regression coefficients for distance (in 100km) from closest port, airport,and intermodal terminal by year against the logarithm of employment.Point represents estimate. Wide line indicates the 50 percent confidence interval, and thin line represents the95 percent confidence interval for the estimate. Dotted grey lines indicate industrial classification transitions.Red line indicates trend in estimators.

1970 1979 1984 1989 1994 1999 2004

-0.0

50.

000.

05

port

1970 1979 1984 1989 1994 1999 2004

-0.0

50.

000.

05

airport

1970 1979 1984 1989 1994 1999 2004

-0.0

50.

000.

05

intermodal

Figure 7.7: Automatic environmental control manufacturing for residential, commer-cial, and appliance use (SIC 3822 and NAICS 334512): Regression coefficients for dis-tance (in 100km) from closest port, airport, and intermodal terminal by year againstthe logarithm of employment.Point represents estimate. Wide line indicates the 50 percent confidence interval, and thin line represents the95 percent confidence interval for the estimate. Dotted grey lines indicate industrial classification transitions.Red line indicates trend in estimators.

1970 1979 1984 1989 1994 1999 2004

-0.1

0-0

.05

0.00

0.05

0.10

port

1970 1979 1984 1989 1994 1999 2004

-0.1

0-0

.05

0.00

0.05

0.10

airport

1970 1979 1984 1989 1994 1999 2004

-0.1

0-0

.05

0.00

0.05

0.10

intermodal

Figure 7.8: Semiconductors and related device manufacturing (SIC 3674 and NAICS334413): Regression coefficients for distance (in 100km) from closest port, airport,and intermodal terminal by year against the logarithm of employment.Point represents estimate. Wide line indicates the 50 percent confidence interval, and thin line represents the95 percent confidence interval for the estimate. Dotted grey lines indicate industrial classification transitions.Red line indicates trend in estimators.

181

1970 1979 1984 1989 1994 1999 2004

-0.0

50.

000.

050.

10

port

1970 1979 1984 1989 1994 1999 2004

-0.0

50.

000.

050.

10

airport

1970 1979 1984 1989 1994 1999 2004

-0.0

50.

000.

050.

10

intermodal

Figure 7.9: Adhesive manufacturing (SIC 2891 and NAICS 325520): Regression co-efficients for distance (in 100km) from closest port, airport, and intermodal terminalby year against the logarithm of employment.Point represents estimate. Wide line indicates the 50 percent confidence interval, and thin line represents the95 percent confidence interval for the estimate. Dotted grey lines indicate industrial classification transitions.Red line indicates trend in estimators.

1970 1979 1984 1989 1994 1999 2004

-0.0

4-0

.02

0.00

0.02

0.04

0.06

port

1970 1979 1984 1989 1994 1999 2004

-0.0

4-0

.02

0.00

0.02

0.04

0.06

airport

1970 1979 1984 1989 1994 1999 2004

-0.0

4-0

.02

0.00

0.02

0.04

0.06

intermodal

Figure 7.10: Gum and wood chemicals manufacturing (SIC 2860/2861 and NAICS325191): Regression coefficients for distance (in 100km) from closest port, airport,and intermodal terminal by year against the logarithm of employment.Point represents estimate. Wide line indicates the 50 percent confidence interval, and thin line represents the95 percent confidence interval for the estimate. Dotted grey lines indicate industrial classification transitions.Red line indicates trend in estimators.

1970 1979 1984 1989 1994 1999 2004

-0.1

0-0

.05

0.00

0.05

0.10

0.15

0.20

port

1970 1979 1984 1989 1994 1999 2004

-0.1

0-0

.05

0.00

0.05

0.10

0.15

0.20

airport

1970 1979 1984 1989 1994 1999 2004

-0.1

0-0

.05

0.00

0.05

0.10

0.15

0.20

intermodal

Figure 7.11: Turned products and screw, nut and bolt manufacturing (SIC 3450 andNAICS 33272/): Regression coefficients for distance (in 100km) from closest port,airport, and intermodal terminal by year against the logarithm of employment.Point represents estimate. Wide line indicates the 50 percent confidence interval, and thin line represents the95 percent confidence interval for the estimate. Dotted grey lines indicate industrial classification transitions.Red line indicates trend in estimators.

182

of the Census 2007), are highly deconcentrated with over 12,000 establishments and

almost 80,000 employees. The regression results (Table C.7) offer several interesting

findings. The overwhelming locational determinant is population density, with a

fairly consistent increase of eight percent in retail employment for every ten percent

increase in population. There is also an inclination to locate toward more highly

educated populations. And there is strong negative trend with regard to income.

That is, once we have accounted for the effect of education in increasing the presence

of such retail outlets, they have been declining in higher income counties since the

late 1980s. While the explanation for this seeming contradiction is not clear, one

might posit that the advent of internet services, like Amazon books, may be shifting

this commerce away from local retail outlets to facilities located in lower cost areas

or accessed electronically. With regard to transportation, these retail establishments

demonstrate no relation whatsoever to ports and a consistent bias to locate away

from airports and intermodal terminals. We can thus suggest that retailing is also

decoupled from the broader, interregional transportation infrastructure but not the

intraregional infrastructure associated with high populations. Thus, both Cooley

and the neo-Marshallian’s conceptions are borne our in this sector for the industry

selected.

7.2.5 Information work

The information industries comprise a broad category that comprises “establishments

engaged in the following processes: (a) producing and distributing information and

cultural products, (b) providing the means to transmit or distribute these products as

well as data or communications, and (c) processing data” (U.S. Dept. of Commerce,

Bureau of the Census 2007). This category has been dubbed the creative class and

183

1970 1979 1984 1989 1994 1999 2004

-0.0

50.

000.

050.

100.

150.

200.

25

port

1970 1979 1984 1989 1994 1999 2004

-0.0

50.

000.

050.

100.

150.

200.

25

airport

1970 1979 1984 1989 1994 1999 2004

-0.0

50.

000.

050.

100.

150.

200.

25

intermodal

Figure 7.12: Book, periodical, and music stores (SIC 5942, 5994, and 5733/5735 andNAICS 451211, 451212, and 451220): Regression coefficients for distance (in 100km)from closest port, airport, and intermodal terminal by year against the logarithm ofemployment.Point represents estimate. Wide line indicates the 50 percent confidence interval, and thin line represents the95 percent confidence interval for the estimate. Dotted grey lines indicate industrial classification transitions.Red line indicates trend in estimators.

much has been made of its role in driving the contemporary economy (Currid 2006;

Florida 2002). While these writers’ definition and classification of knowledge work has

been criticized as too broad (Peck 2005), some scholars have begun to concentrate on

more narrowly defined artistic pursuits (e.g., Markusen 2006; Markusen and Schrock

2006).

In this study I employ two sectors to represent information work. The first is peri-

odical publishing, which comprises “establishments known as magazine or periodical

publishers. These establishments carry out the operations necessary for producing

and distributing magazines and other periodicals, such as gathering, writing, and

editing articles, and selling and preparing advertisements. These establishments may

publish magazines and other periodicals in print or electronic form” (U.S. Dept. of

Commerce, Bureau of the Census 2007). In 1997, this large, deconcentrated industry

employed almost 140,000 people in over 6,000 firms. The second industry, movie and

video production, is similarly large and deconcentrated. In 1997 over 80,000 people

worked in almost 9,000 firms comprised of “establishments primarily engaged in pro-

ducing, or producing and distributing motion pictures, videos, television programs,

or television and video commercials” (U.S. Dept. of Commerce, Bureau of the Census

184

2007).

While periodical publishing is most closely tied to population density, it is also

correlated with foreign populations and more highly educated populations. Employ-

ment rises currently by half a percent for every percent increase in population density,

by 0.15 percent for each percentage increase in the proportion of the population with

bachelor’s degrees, and by 0.07 percent for every percentage increase in the foreign

born population. This sector has demonstrated some significant geographical shifts

over the past four decades. As the Table C.16 and Figure 7.13 show, periodical

publishers are moving away from intermodal terminals and toward ports, while they

remain indifferent to airports. Though motion picture and video production is simi-

lar (though somewhat weaker) with regard to the social indicators, i.e., attracted to

population centers (0.35 percent), foreign populations (0.06 percent), and educated

populations (0.11 percent), it differs with regard to transportation. Like periodi-

cal publishing, this sector is indifferent to airport location and is moving away from

intermodal terminals, but it has remained consistently distant from ports. This is

presumably an artifact of the historical evolution of the industry in New York and

Los Angeles rather than any particular need for port services.

Curiously, in the one sector that both Cooley and the neo-Marshallians would

predict no coupling to infrastructure at all, there appears to be some mild relationship.

However, the nature of the relationship remains unclear. There is certainty that these

industries are moving away from intermodal terminals, but while neither is moving

closer to airports, one is immobile with respect to ports and the other is moving

closer to ports. It is tempting to combine these conflicting impulses and argue that

they cancel each other out, indicating a decoupling from infrastructure. This would

lend support to both the neo-Marshallians and Cooley, but the question bears more

inquiry.

185

1970 1979 1984 1989 1994 1999 2004

-0.0

50.

000.

050.

10

port

1970 1979 1984 1989 1994 1999 2004

-0.0

50.

000.

050.

10

airport

1970 1979 1984 1989 1994 1999 2004

-0.0

50.

000.

050.

10

intermodal

Figure 7.13: Periodical publishers (SIC 2720 and NAICS 511120): Regression coeffi-cients for distance (in 100km) from closest port, airport, and intermodal terminal byyear against the logarithm of employment.Point represents estimate. Wide line indicates the 50 percent confidence interval, and thin line represents the95 percent confidence interval for the estimate. Dotted grey lines indicate industrial classification transitions.Red line indicates trend in estimators.

1970 1979 1984 1989 1994 1999 2004

-0.0

50.

000.

050.

10

port

1970 1979 1984 1989 1994 1999 2004

-0.0

50.

000.

050.

10airport

1970 1979 1984 1989 1994 1999 2004

-0.0

50.

000.

050.

10

intermodal

Figure 7.14: Motion picture and video production (SIC 7813 and 7814/7812 andNAICS 512110): Regression coefficients for distance (in 100km) from closest port,airport, and intermodal terminal by year against the logarithm of employment.Point represents estimate. Wide line indicates the 50 percent confidence interval, and thin line represents the95 percent confidence interval for the estimate. Dotted grey lines indicate industrial classification transitions.Red line indicates trend in estimators.

7.3 Conclusion

It is apparent that the relation between transportation nodes and economic activity

is a complex one that is dependent upon the nature of each type of activity and their

interaction. As Table 7.3 summarizes, however, the neo-Marshallian perspective is

more right than wrong, and Cooley’s ideas no longer seem to hold for the relatively

efficient transportation systems of the United States.

Both Cooley and Marshall’s ideas about natural resource extraction are accurate.

Industries dealing with raw materials have not located with regard to available infras-

tructure nodes so much as they have determined the extents of the network’s branches.

186

neo-Marshallians Cooley Findings

Natural resources Decoupled Decoupled DecoupledHeavy industry Decoupled De/coupled Decoupled

Networked manufacturing Decoupled Coupled DecoupledMarket manufacturing Decoupled Coupled Decoupled

Retail Decoupled Coupled De/coupledInformation industry Decoupled Decoupled Uncertain

Table 7.3: Relationship of industrial sectors to transportation following Marshall andCooley

Heavy industry, presumably as a product of its large sunk costs, behaves in a similar

manner, and has not relocated relative to infrastructure over the last four decades.

Cooley’s strong thoughts on manufacturing location turn out to be inaccurate for our

time. Both networked and market-oriented manufacturing demonstrate little change

relative to the infrastructural nodes in question, except perhaps for airports. This

suggests a decoupling. Retail demonstrates a very high correlation to population

density, which suggests that it is coupled to that intraregional infrastructure that

supports urban commuting rather than the interregional infrastructure that supports

freight movements. This, in essence, supports both the neo-Marshallians and Coo-

ley. Finally, the sector upon which the neo-Marshallian position is originally founded,

that of knowledge work, is curiously uncertain. It seems to be demonstrating some

attraction toward ports and a movement away from intermodal terminals.

The unexpected finding here is the magnetism of airports for some industries. The

role of air freight has likely become a much more important factor in many manufac-

turing industries’ logistical calculations, especially as the value of manufactured goods

in the US is increasing, making them more amenable to the costs of air shipping.

Overall, the regression analyses suggest that the location of economic activity is

decoupled from the freight transportation network, as the neo-Marshallian approach

187

suggests. That is, the economic terrain has not shifted to reflect altered freight flow

patterns. However, because the relation to infrastructural nodes has shifted very little

if at all, the conclusion must be drawn that economic activity has been decoupled

from the break points of the infrastructure network for more than half a century.

This implies that containerization has not impacted firm location on the broad scale.

Instead, it appears more likely that the location of economic activity is a product

of unique historical trajectory on an industry by industry basis, though these may

find some common foundation in the break points established by the transportation

system as it was first built.

Therefore, the analyses in this chapter and the preceding suggest that the main

public rationales for subsidizing ports do not hold up. Direct employment is de-

creasing as mechanization transforms operations, and formerly local activities, like

warehousing, move inland for better access to highways and affordable buildings.

Meanwhile, so-called port-dependent activities do not seem to have changed their lo-

cation despite a spatial restructuring of the logistics system. On the whole, economic

activity has decoupled from transportation infrastructure’s break points, rendering

the subsidization of intermodal facilities with the goal of attracting industry to one’s

territorial jurisdiction ineffective except insofar as the territory supports population

density. That is, the product market area for the economic activities examined here

is determined by consumption flows rather than freight flows. Subsidy competition,

then, results not in the creation of jobs and growth but rather in the transfer of wealth

from local users and taxpayers to private companies.

188

Chapter 8

Territorial Monopoly

The development of rational economic action from its origins

in the instinctively reactive search for food or in traditional

acceptance of inherited techniques and customary social

relationships has been to a large extent determined by

non-economic events and actions, including those outside

everyday routine and also by the pressure of necessity in cases

of increasing absolute or relative limitations on subsistence.

Weber (1997, 168)

Old methods have remained along with the new, when the

new were avowedly more effective. There is a conservatism

about longshore work which has ever met the spirit of progress

reluctantly; a kind of traditionalism which, from a sense of

habit, clings to things as they have been.

Barnes (1915, 30)

189

8.1 Introduction: Rationalizing congestion

Thus far, the dissertation has shown how containerization and intermodalism re-

shaped the logistics network and altered the economic terrain of the United States

and how this network’s newfound flexibility disabled port authorities’ and longshore-

men’s ability to territorially bind the flow of economic activity by subsidizing global

shipping companies, undermining the political rationale for subsidizing these activi-

ties. The present chapter looks to containerization’s pre-history to search for a way

forward for port authorities by exploring the successful territorial strategies employed

by longshoremen to improve their working conditions and the technological strategies

shipping companies adopted to reverse those gains.

One little considered alternative to market competition among ports is port con-

solidation and rationalization, which would coordinate port traffic at the national

level, balancing throughput across available capacity with attention paid to the needs

of local labor pools. Though dismissed by many port officials, rationalization is not

without positive precedent in the U.S. This chapter examines one such case: the ra-

tionalization of labor and freight handling through a fractious process of conflict of

negotiation between unions and employers during the first half of the twentieth cen-

tury. By scaling their geographic monopoly up from single piers to the regional and

national levels, longshoremen were able to slow the adoption of labor-saving technolo-

gies and smooth out the distribution of employment opportunities. As labor victories

increased handling costs on the docks, shipping companies sought to introduce tech-

nologies that would make them more competitive relative to other modes, like rail and

trucking. As longshore unions came to recognize that the economics of containeriza-

tion ensured the transformation of labor processes, they opted to reach a compromise

with employers in which the employers “bought out” restrictive work rules on the

190

introduction of new technologies in exchange for income guarantees. The outcome

was job security for the longshoremen and the introduction of containerization.

The language of rationalization has been employed consistently through the last

century in the discussion of labor-capital relations in ports. In finance and economics,

rationalization refers to the restructuring of an organization in pursuit of increased

efficiency or profits, and this, I will argue, has been the main motivation for employ-

ers’ efforts to mechanize the waterfront. For Weber, however, whose conception is

perhaps more consistent with that of his early twentieth century contemporaries on

the docks, rational economic action is a broader conception that includes, inter alia,

“the systematic distribution, as between present and future, of utilities, on the control

of which the actor for whatever reason feels able to count”(Weber 1997, 168), which

refers to the systematic temporal distribution of opportunities for the consumption

and production of goods and services for a given actor. This means, in essence, that an

actor’s production and consumption of goods and services is predictable and smooth

over time. Weber (1997, 168) then adds to his list of economically rational action:

“The systematic distribution of available utilities as between their various potential

uses in the order of their estimated relative urgency, according to the principle of

marginal utility.” If one accepts the idea that control over utilities allows for rational

redirection of those utilities to others, then the call for equitable social distribution

in a situation of relative or absolute deprivation also constitutes rational economic

action. For longshoremen, this systematic distribution took the form of replacing

irregular, uncertain, and poorly remunerated employment with regular employment

and predictable wages for all longshoremen. Attaining this goal required that the

unions establish a territorial monopoly over the “effective area of production” (Levin-

son 1967) that allowed them to control entry to the profession and actual work on

the docks.

191

Levinson (1967) makes a distinction between “area of effective production” and

“product market area” in his discussion of the relative power of labor unions and

oligopolies. The product market area is that spatial range throughout which a good

or service is sold, while the area of effective production refers to the geographical space

within which that good or service can be produced. For example, under European

law which protects the designation of origin for food and beverages, chees labeled

Parmigiano-Reggiano can only be manufactured in a few areas in Emilia-Romagna,

Italy. Its area of effective production is geographically delimited. However, its prod-

uct market area is global, as it is available in supermarkets and cheese shops around

the world. The principle is the same for any geographical constraint (physical or

social) that delimits the space in which a good or service can actually be produced.

Drawing on Herod’s (2001) work on labor geographies, in which he illustrates the

spatial strategies of unions, I propose to refer to the ability to control this area of

effective production as a “territorial monopoly.” While one might also posit a territo-

rial monopoly over the product market area if an actor is able to control distribution

of a good or service (like Disney themeparks, for instance), this is not relevant to our

discussion in this chapter, which focuses on struggles over production processes.

Though the area of effective production is typically diffuse for most manufacturing

efforts, this is not the case for water transportation. Physical characteristics, like

shallow waters, extreme tides, and ocean exposure, limit the number of points at

which cargo can be transferred between water and land. Thus, one facet of the struggle

between labor and capital in freight transportation manifests itself as a struggle by

unions to monopolize the labor supply in a port or port range and an effort by

employers to undermine or circumvent these monopolies through competition and

technological innovation. The goal of this chapter is to demonstrate the importance

of securing a territorial monopoly over a port range to ensure that workers receive

192

their fair share of productivity gains and enjoy secure, stable employment.

8.2 Before mechanization

Longshoring has always been back-breaking work. And it has always been collective

work. Amidst the tangle of ropes, swinging booms, and ceaseless movement, the

longshore gang’s members must develop a rhythm if they are to work as efficiently

and as safely as possible. Developing a rhythm requires a constant awareness of each

member’s interdependence with the others (Barnes 1915, 33–34). As a consequence

of this collective relationship, longshore work fosters a social consciousness that the

constant threat of unemployment and poverty have long undermined.

The expression of communal spirit is reflected in the work-preserving tradition-

alism described in this chapter’s epigraph by Barnes. Barnes was sponsored by the

Russell Sage Foundation just before World War I to conduct the first thorough study

of American longshore work. His comprehensive description of New York’s ports is

modeled on Charles Booth’s study of the dock laborer in London, which Barnes cites

as the first study of dock laborers. The oversupply of labor described in his report and

conveyed in more detail below generated resistance among the longshoremen to any

development that might put themselves or their mates out of work. Yet, perversely,

the possibility of organizing initially generated the same fears, as less skilled workers

accurately believed work opportunities would be monopolized by the more skilled.

In the early twentieth century, both mechanization and organization threatened to

reduce employment opportunities, and both were therefore resisted.

193

8.2.1 Meeting the hook

There are two basic types of cargo: bulk and general. At the turn of the twentieth

century, the former included goods like grain, sugar, oil, bananas, coal, and lum-

ber. These required no separate packaging and were generally shipped as complete

shiploads. The latter consisted of a miscellany of items of different sizes, densities,

and characteristics all destined for the same port, usually a foreign one. The handling

of commodities in these two categories consequently differed. For example, while coal

would be shoveled into large buckets and hoisted over the edge of the ship to the

dock, automobiles would be secured to pallets that were hoisted by a crane into the

ship’s hold, where they were moved into place and secured against shifting while at

sea. Although bulk cargoes, especially bananas, required a certain modicum of skill

to handle, general cargo demanded a high degree of skill that could only be learned

through on-the-job training. Barnes (1915) estimates that it took ten years to learn

how to properly estimate the amount of cargo that could be stowed in the hold and

how to stow the miscellaneous items that comprise general cargo so that the cargo

would not only make the journey undamaged but also not shift during sailing and

cause the ship to list and perhaps capsize. The subtleties that can be involved are

captured by Barnes’s (1915, 53) note of the importance of covering a load of cayenne

pepper with a tarp if it were in the same hold as horses, so that the latter would

not sneeze themselves into a panic and cause the ship to list. (For discussion of the

non-transferability of tacit knowledge see Dicken and Malmberg [2001], Malmberg

and Maskell [2006], Nelson and Winter [1982].)

When a ship arrived in port, a “hatch gang” averaging 18 to 23 men1 would be

1Barnes (1915) reports a wide range of gang sizes in the docks of New York Harbor, dependingon the size of the ship and the nature of the cargo. His study showed these ranging from a minimumof 12 to a maximum of 30.

194

assigned to one of the five to seven hatches on the larger vessels (roughly 400 to

500 feet at the time). Each hatch gang’s men—for they were invariably men—were

divided into three groups: pier men, deck men, and hold men. Four to six deck men

would start by “rigging the ship,” while the others waited nearby lest they miss the

start of work and be replaced before they could return. Rigging the ship consisted

of stringing the “up-and-down fall,” the rope with a hook at the end for lifting and

lowering cargo, through the derrick’s pulley “blocks” and the ship’s winches, and then

attaching the “burtons” for guiding the loads athwart the ship. The deck men would

then be responsible for all the movements of the rope and its cargo. Once the ship was

rigged, the hold men and pier men would jump into action. Five to seven hold men,

supervised by a “header,” or hold foreman, would arrange a load of cargo, a “draft,”

atop a “sling board,” a type of wooden board or pallet, then pass a “sling,” a rope

braided to itself in a circle, under both ends and feed one end of the sling through the

other to create a self-tightening cinch that theoretically pressed the cargo together

into a single load that would not damage itself nor topple out of the sling through

the movement of the derrick. After the hook passed through the sling and the call

to “Hoist away!” was made, the load was “broken out” or lifted by the deck men

using hand winches, or, in the occasional case of dock-mounted derricks, using horse-

drawn winches. Once it was transferred to the “square” of the hatch, the deck men

would employ a series of signals to safely and efficiently swing the “sling load” athwart

the ship and lower it to the dock, where one of the nine to twelve pier men would

disconnect the loaded sling from the hook and hang an empty one back on. As the

derrick swung back to the hatch, the pier men, under the direction of the pier foreman,

would truck the goods back and forth on the pier and stack them for storage until

the shipper came to pick them up. Barnes (1915, 129) reports that the average load

carried by a pier man or hold man on a single trip was roughly 200 to 250 pounds.

195

On smaller ships, like coastal freighters, they might be expected to carry these loads

down narrow gangplanks from the ship’s edge to the pier rather than using a derrick.

Loading a ship was effectively the same process in reverse.

If the gang was well coordinated and had established a“swing and rhythm”(Barnes

1915), then the derrick would operate in a smooth, graceful sweep as the pier men at

one end and the hold men at the other time their unloading and loading such that

they were able to “meet the hook” at the moment it arrived, removing one sling and

attaching another as the hook swung back into its return journey. This, according

to Finlay (1988, 40), was the clearly visible sign of a highly productive hatch gang.

Gangs and gang members who possessed the skills to work with such efficiency were

highly prized by employers and could find relatively regular work.

8.2.2 Congestion and casualization

Congestion on the docks

The shipping industry has an age-old adage that a ship only makes money at sea.

Since a ship generates income through the point-to-point transportation of cargo,

any time spent anchored at dock loading and unloading reduces the total number of

trips a ship can make over a given period (usually measured as trips per year) and

thus revenue. A ship, then, only generates profits through movement. Loading and

unloading, from the ship’s perspective, is a choke point that congests the ship’s move-

ments. If a dock is congested with goods that slow down the loading and unloading

of a ship, revenue is lost. And the more expensive the ship, the greater the losses that

accrue through dock congestion, as larger sunk costs must be amortized across fewer

transported goods. As inherent economies of scale in ship size and technological ad-

vances have continuously encouraged shippers to build ever larger ships, the pressure

196

to reduce turnaround time has increased over the decades.

In the early days of shipping, the relatively irregular arrival of ships would create

both periods of increased congestion, as goods from multiple ships piled up, and

periods of empty calm, as merchants removed their goods and longshoremen bided

their time waiting for the next ship. Generally, longshoremen would find out mere

hours before the arrival of a ship. For example, in New York a flag would be run up

a pole on Sandy Hook, which forms the southern lip of the entrance to New York

Harbor, when an approaching ship was spotted and identified. Men would then make

their way to the docks in hopes of being hired. Advanced notice and predictability

were increased with the construction of steamships, which were more powerful and

did not rely on unpredictable winds, and with the invention and dissemination of

the wireless telegraph after 1896, which allowed steamers to stay in fairly constant

contact with their home offices (Barnes 1915). The new steamships, however, though

involved in regular services from the turn of the century and more reliable than sailing

ships, were still subject to inclement weather and better at debarking on time than

arriving on time, since lost time could often be made up by working longshore gangs

more intensely.

Crowding on the docks would lead to mazes of narrow aisles with goods stacked

high along both sides. Such storage would regularly lead to the double and triple

handling of goods, increasing costs commensurately. According to Hobsbawm (1964,

212–13), these conditions were aggravated by merchants’ own strategies for avoiding

warehousing costs and transfer times by leaving their goods on the docks as long

as possible. So while there were incentives for longshoremen (in the form of more

work) and for merchants (in the form of lower storage costs) to let goods pile up,

merchants and stevedoring companies were also driven by the contradictory drive cut

down overall costs by rationalizing waterfront organization.

197

Congestion in the ranks

Not only was there congestion on the docks, there was also congestion in the ranks

of longshoremen. The “shape up,” most famously seen in contemporary times in the

movie On the Waterfront, has remained the primary means of hiring longshoremen,

though its outward manifestation has changed over time. When a ship came into

port, longshoremen would gather at the gates, shaping themselves into a large arc (see

Figure 8.2.2). From the center of the arc, the foreman, who was employed by local

stevedoring companies or shipping companies, could look the men over and choose

those whom he wanted on his gang. Generally, as the men were selected, they would

give their names to the time-keeper and pass through the gates to the pier to await

the start of their work. Because of this casual hiring system and the possibility that

any man otherwise unemployed might be able to pick up a spot of work, the number

of job seekers would swell far beyond the number needed, especially during difficult

economic times. Barnes (1915) suggests that in 1915 there were approximately three

men for each job. More often than not a single-file arc would be insufficient for the

number of men seeking work, and the shape up could become a half dozen or more

rows deep with men constantly jostling to get near the front. Barnes (1915, 64–65)

vividly captures the nature of this crowd in the following incident:

After the first six or seven hundred are called, the struggle of the remaining

men to get into conspicuous places increases. They are so thickly packed

near the doorway that often a man who is entitled to pass in has to pulled

through by his fellows. A few get through who either have no check or

have the wrong one. They are promptly stopped, jerked toward the rope

at the wagon entrance, and told to get under the rope and outside. This

is difficult, for the crowd of men against the rope is a solid mass. They

198

are forced to push themselves into this crowd. One man driven out lowers

his head till it is about on a level with the stomachs of the men in front,

makes a sort of catapult of himself, and shoots into the mass, striking the

men as he goes. His progress far out into the crowd can be traced by the

wriggle or wave motion he produces.

Figure 8.1: Shape-up in New York City, c.1932.Source: U.S. Bureau of Labor Statistics (1932)

While some docks may have literally thousands of men struggling for one of a

thousand places, others (generally with irregular and infrequent sailing schedules) may

have goods piling up for lack of workers (U.S. Bureau of Labor Statistics 1932). In

New York, the lack of a central location for workers to obtain information about where

workers were needed meant that longshoremen would have to walk from pier to pier

looking for work, allowing for the very real possibility that some piers would have too

few workers while others had too many. The consequence was an increase in congestion

199

in the ranks and on the docks, generating additional pressure for rationalization that

could even out the supply of workers and the movement of goods on the docks.

Congestion begets casualization

Congestion in the ranks and on the docks had at least three impacts. First, the

oversupply of longshoremen, under the classic law of supply and demand, lowered

hourly wages and pushed them toward subsistence levels. Barnes (1915) reports that

the average longshoremen’s wages in 1915 were abysmally low, though because of the

irregularity of work and the lack of record keeping for or by individual longshoremen,

who may work on any number of different piers for any number of different companies,

he is only able to offer ballpark figures. He estimates on the basis of interviews and

diaries from workers and foremen that a week’s wages in the foreign trade averaged

below $12 and that the wages of many fell below $10 per day. Over a year, he cites

experienced individuals as suggesting a range of $520 to $624, though he notes that

many others put the figure at less than $500. This is far below the minimum of

roughly $850 established by the Bureau of Personal Service of the Board of Estimate

and Apportionment for an unskilled worker’s family of five in 1915. Food and housing

alone amount to over $550, or the longshoremen’s annual average(Bureau of Applied

Economics 1920). Other studies, like that of the New York State Factory Investiga-

tion Commission (National Industrial Conference Board 1921, 26), found comparable

figures.

Second, because of the skills involved in longshoring and the need to clear ships

and docks as rapidly as possible, a two-tier system of regular gangs and “chenangoes,”

or spot workers, developed, irregularly distributing income across longshore workers

(Barnes 1915; Kahn 1980). Because of the high degree of nontransferrable knowledge

involved in longshoring, those individuals who had the opportunity to develop these

200

skills were in fairly high demand and were generally able to secure work on regular

gangs that would work for the same foreman and thus the same company in a situation

that approached regular employment. The shipping companies prized these men’s

skills for the efficiency and minimal damage to goods with which ships could be

unloaded. But while skilled longshoremen were able to secure employment regularly,

they were still required to shape up with the other men and face the possibility of

not being hired. The chenangoes, on the other hand, comprised the second tier of

workers. They were compelled to accept less desirable and less complex work or hope

to fill in for a regular man who was unable to work. As a result, their income was

even more irregularly distributed over time and generally had to be augmented with

other types of casual work.

Third, there was no security for the longshoreman. It should not be construed

from the preceeding paragraph that regular workers could rely on their favorable

situation. They retained their positions because they worked intensely in constant

fear that taking breaks or complaining about dangerously heavy loads could lead to

their dismissal and immediate replacement (Finlay 1988, 39). Not only was there the

worry and uncertainty of being hired, there was also worry about remaining hired.

“The longshoreman is on a more casual basis than the ordinary day laborer, who,

when he is hired, is at least assured of a day’s work. After work has begun, men are

knocked off and rehired at any hour according to the demands of the work” (Barnes

1915, 57). And should one lag too far behind at the end of scheduled lunch break or

work interruption, one was likely to be replaced.

Additionally, while wages and hours may average out over a year, in the short

term, they were highly unpredictable (Barnes 1915; U.S. Bureau of Labor Statistics

1932, 70). Longshoremen could never really be certain when a ship would arrive, and

they could never know how long they were going to have to work. Generally each ship

201

would generate three to five days of high intensity work. Though the official work day

was ten hours, it was not uncommon for men to work 12 to 14 hours, and there were

regular reports of men working 15, 24, or even 36 hours at a stretch if the foreman

demanded it. Again, fear of losing a particular job or the potential for future jobs

would lead men to work to the point of utter exhaustion, greatly increasing the risk

of fatal and nonfatal accidents. After such long stretches, men would require days,

sometimes up to a week, to recover, eliminating any possibility of working during

that period. Figures 8.2 and 8.3, reproduced from Table 32 of U.S. Bureau of Labor

Statistics (1932), capture the impact of such a schedule on individual income. The

two figures illustrate weekly swings of up to $50 and monthly variations up to $75

for the average earnings per gang member for two gangs in New York in 1928. These

wide swings supplied rather uncertain footing for the longshoreman, who had regular

recourse to payday lenders, often their own foremen.

8.2.3 Casualization and organization

Under such horrendous working conditions, conditions that threaten livelihoods and

unevenly distribute income temporally and socially, it seems reasonable to assume that

men would be driven to take rational economic action to establish predictability and

security. And there were periods in longshoring history when men did. The earliest

recorded longshoremen’s strike in the US was in 1836 and major strikes occurred in

1874, 1887, 1907, and 1919, but unions were particularly ephemeral on the waterfront

(Barnes 1915; Nelson 1988). Longhshoremen themselves were generally hostile to

labor organization at the beginning of the twentieth century.

Hobsbawm (1964, 208–11) suggests that the union organizer had to overcome two

obstacles to organize the docks. First, like all unions, the longshore organizer had to

202

Gang 1

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

025

50

75

Gang 8


025

50

75

Figure 8.2: Weekly earnings per man for two gangs in 1928 in New York. Horizontalline represents the annual mean.Source: U.S. Bureau of Labor Statistics (1932)


050

100

150

200

Gang 1


050

100

150

200

Gang 8

Figure 8.3: Monthly earnings per man for two gangs in 1928 in New York. Horizontalline represents the annual mean.Source: U.S. Bureau of Labor Statistics (1932)

restrict entry into the trade. It was only by this means that the oversupply of labor

could be tamed and wages driven up through scarcity and regular employment. In

this process, the organizer had to overcome the challenge of excluding the unskilled.

The second problem was that the union organizer had to find ways to prevent strong

sections of strategic workers from forming their own quasi-craft union. To this list,

however, we can add a third: organizers had to be sensitive to the appropriate scale

of organization.

203

Controlling supply

Restricting entry entailed addressing three obstacles: the two-tier system, the problem

of the unskilled, and the cycle of work. First, the two-tier system created a class of

worker that experienced some comparative level of income security and thus had little

incentive to organize. They feared that organizing would reduce their own hours and

lower their already diminished quality of life by more evenly distributing labor and

wages across all longshoremen. The skilled workers’ objections to organizing were

precisely contrary to that of the second class of worker, the unskilled. These men were

directly threatened with the loss of employment opportunities that organization would

entail. “[T]hough some trade union leaders might have appreciated the advantages

of decasualization from a bargaining point of view, the men opposed it. It was one

thing to stop new men entering the trade; quite another to throw Bill and Jack (and

perhaps oneself) out on the streets. . . . The poorer and more casual the docker, the

more he would cling to the rough justice of casualism, even if it was only the justice

of the lottery, in which anyone could draw the lucky number” (Hobsbawm 1964, 209).

For the poorest workers, the choice seemed to be between occasional work that would

at least put a morsel on the table and no work at all. The labor organizer had to

navigate between this Scylla and Charybdis to convince the skilled worker that his

wages and conditions would improve and the unskilled worker that a union would

look after him, too.

The third obstacle was longshoremen’s perception of the impact of a regular work

week. As described above, longshore work tended to come in bursts of intensive labor

over several days, followed by days of downtime. That downtime was most often

employed recuperating from the back-breaking work of the previous days. The work

could be so grueling that at the end of the work day, men would walk home looking

204

like ‘orangutans’, only to somehow straighten up in time for work the following day

(Levinson 2006, 18). In the longshoreman’s work cycle, periods of rest and recovery

were the necessary counterpart to periods of exertion. And, according to Barnes

(1915, 74), the men could not see past this pattern to a regular work week that did

not involve “killing stretches of overtime.” Consequently, the promises of regularity

by advocates of decasualization were conceived of as constant overexertion with no

opportunity for recovery and were thus unable to inspire and motivate the men to

organize.

Keeping strategic workers

Though in contemporary times, the crane operators comprise a small number of highly

skilled workers who might be able to break away to form a union based on their strate-

gic importance, in the early days of longshoring, as Hobsbawm (1964, 205) argues,

the great variety of interdependent work on the waterfront resulted in no obvious,

task-based core for organization. The concern that quasi-craft unions would form

was undermined in two ways. A major hindrance in the United Kingdom, according

to Hobsbawm, was that waterfront workers were not readily accepted by other unions.

There was no perception among other industries of the skills necessary to conduct

longshore work efficiently. This perspective also seems to have been prevalent in the

US (Nelson 1988, 114). The second factor, which will be discussed in the next section,

was that the nascent formation of such craft unions came under almost immediate

threat by employers’ efforts to mechanize the waterfront and aligned union members’

interests with those of the other longshoremen around them.

205

Localized control

To these two obstacles, we can add the requirement of scale. The organizational ba-

sis for waterfront work is inherently geographical. Because ports require particular

geographical features to be successful, such as sheltered water, deep channels, and

hinterland access, there are a limited number of locations for port activities and a

constrained area of effective production. Thus, though stevedoring or shipping com-

panies can move their businesses to escape labor difficulties, if labor can be organized,

it effectively obtains a territorial monopoly.

The issue then becomes one of the appropriate geographical scale of organization

(Herod 2001, cf.). Hobsbawm suggests that organization could have taken place at a

scale as large as the entire port or as specialized as individual piers. Over time, indi-

vidual piers had come to be operated by single stevedoring companies and to develop

their own traditions and work patterns (Barnes 1915, 28). The distinct patterns of

each dock favored a particular set of workers who developed tacit knowledge of these

highly localized work processes. This in turn made it possible for workers to organize

at the scale of the individual pier. However, while this might have produced some

modicum of bargaining success against an individual company, ultimately shippers

were able simply to redirect their shipments to nearby piers and undermine the or-

ganizing effort. To an extent limited by existing technology, this same principle of

redirection applied to entire ports. Shippers could move their goods through a com-

peting port and complete the delivery of goods over land. Thus, Hobsbawm, 206’s

suggestion that there was little strategic or tactical advantage to organizing UK long-

shoremen at the regional and national levels is misplaced. As Finlay (1988, 38) points

out, successful organizing efforts must incorporate workers from all competing ports,

a scale that made organizing that much more challenging.

206

8.3 Early mechanization

Though there were some turn of the century technological improvements, scientific

management was only actively introduced from the 1920s. The impetus was primarily

to relieve congestion on the docks, as goods would pile up while they awaited merchant

or ship pick up, adding greatly to the stevedoring companies’ costs of managing

freight. Though the technological interventions were relatively minor, it has been

suggested that this initial threat to longshoremen’s livelihoods provided the impetus

for them to overcome their collective action problems and organize into unions.

8.3.1 Fear of mechanization

Mechanization crept slowly onto US docks during the first two decades of the twentieth

century. Advancements were generally limited to the introduction of steam-powered

winches onboard the ships that docked, reducing the need for deck men by cutting

down the effort required to hoist a sling. Similarly, some winches were also installed

on the docks, but this was fairly rare. In fact, during this time, the most advanced

European shipyards were installing their first large cranes for moving cargo from ship

to shore (Barnes 1915). Additionally, small trucks came into use for hauling goods

away from the pier, and later fork-lift trucks were introduced to move goods around

the docks.

The literature suggests three reasons that dockworkers overcame their internal

division into two classes of workers and their opposition to organization. First, to the

outside world, including established unions, all longshoremen were simply laborers,

neither particularly skilled nor desirable as trade union members due to their “un-

ruly” character (Nelson 1988, 114; Hobsbawm 1964). Thus, narrow specializations

in longshoring failed to be recognized by the broader union movement. Second, de-

207

spite having developed specialized skills, these were not standardized across workers

responsible for a particular function, since each dock had a unique set of problems

and customs. As a consequence, workers’ bargaining power was generally limited to

the narrow geographic limits of their dock (Hobsbawm 1964, 210–211).

“Moreover—and this is no doubt the most important factor—hardly had the crafts-

men established their position than mechanization began to drive them out of it”

(Hobsbawm 1964, 211). Equipment in the bulk cargo sector like grain-chutes and

mechanical coaling cut down the number of specialized grain- and coal-porters and

replaced them with “semi-skilled” workers. Similarly, in the break bulk sector, equip-

ment like conveyor belts also began to reduce the number of deck men and to replace

specialized sling and winch handlers with semi-skilled workers. The short term impact

was a closing of ranks around specialties, but the long term effect was to clarify the

need for common action within the industry as a whole. Mechanization integrated

the problems of the skilled and the unskilled workers by threatening the security of

them all.

This homogenization of worker’s skills was augmented by the “man-killing pace”

that contemporary machinery was beginning to demand (Nelson 1988, 52–53). Fore-

men and shipowners refused to increase the size of gangs even as mechanical winches,

conveyor belts, and the like accelerated the movement of cargo between ship and

shore and “because it took more energy to handle continuously small packages than

to handle a few large containers at relatively infrequent intervals” (Kahn 1980, 374).

This aggravated the exploitative working conditions the dock workers were already

experiencing and produced a united and consistent demand for an increase in gang

size.

208

8.3.2 Decasualization and geographical expansion: The

1934 and 1936 West Coast strikes

West Coast dockers have traditionally been more progressive than their counterparts

elsewhere and set the model for employer relations. Though infighting between the

International Seamen’s Union (ISU), the Marine Workers Industrial Union (MWIU),

and the International Longshoremen’s Association (ILA) and active, anti-union em-

ployer aggression hindered comprehensive common action on the West Coast during

the 1920s and early 1930s (Nelson 1988, 68–73), the National Industrial Recovery Act

(NIRA) of 1933 led to a resolution. The NIRA granted workers “the right to organize

and bargain collectively through representatives of their own choosing,” and even if

the government was less than zealous in promoting this provision, millions of work-

ers took heart in it and began to implement it themselves (Nelson 1988, 118). This

brought the ILA back to San Francisco’s waterfront, where it attracted the bulk of

longshoremen from that port, effectively settling the debate over which union would

represent the longshoremen of the West Coast.

Control of the organization was quickly overtaken by a group that centered around

the publication of the Waterfront Worker, including Harry Bridges. The group advo-

cated direct control of the local by longshoremen rather than a few leaders, who they

characterized as self-interested and ineffective. Apparently the bulk of longshoremen

agreed with their assessment that organizing under the ILA has “meant a harvest for

a few officials and slave conditions for thousands of stevedores” (Waterfront Worker

1933b), as they actively rejected American Federation of Labor (AFL) norms and

officials during the so-called “Big Strike” of 1934, ushering in a new era of militant

activism on the West Coast (Nelson 1988, 127).

The Big Strike was a bloody, eighty-three day affair that emerged from the pent-

209

up anger of the longshoremen. According to Finlay (1988, 41–44), soon after the ILA

reestablished itself in San Francisco, it entered into contract negotiations with two

simple demands: a coastwide agreement and union control of dispatch halls. Though

dispatch halls already existed to select workers for each particular job, they were

controlled by the companies and used to exclude unionized workers. The strike was

initiated by the San Francisco local on 9 May 1934 and was supported by workers

in most other ports on the West Coast, with the notable exception of the important

port of Los Angeles. The coastwide strike ended 31 July 1934, after the rank and file

had rejected two agreements negotiated by high-level AFL officials and resigned itself

to federal arbitration. In October, the Roosevelt-appointed National Longshoremen’s

Board handed down its award, which included the following gains: 1. an increase in

the basic wage; 2. the location of dispatching in a central hiring hall in each port

that was to be “maintained and operated jointly by the International Longshoremen’s

Association, Pacific Coast District, and the respective employers’ associations” with

the dispatcher“selected by the International Longshoremen’s Association”; and 3. the

establishment of a Labor Relations Committee in each port (Finlay 1988, 43; Kahn

1980, 378).

This effectively gave control of hiring to the workers, inducing Fairley (1979, 9)

to declare the result “an overwhelming Union victory.” Finlay (1988, 43–44) claims

that the ILA’s selection of dispatchers allowed it to end permanent employment for a

single employer in the industry. Rather than companies hiring regular workers on an

effectively permanent basis, subsequent to the settlement, workers were dispatched

on single jobs only, after which, they were to return to the dipatching hall for their

next assignment. This further weakened the two-tier system that threatened effective

organization.

Most importantly for this dissertation, the union took a major step toward fully

210

rationalizing longshore labor by essentially inverting the employer-employee power

structure in which the employer selected the employee and used sanctions to keep

him in line. The union adopted a “low-man-out” system that gave the man with

the lowest hours priority for choosing which particular day to work and which job

to take. Thus, the union could more or less equally distribute the total number

of hours worked among the union members, providing some income stability for all

its members as Kahn (1980) demonstrates. This did not quite establish guaranteed

minimum hours, as the Waterfront Worker demanded as early as 1933 (Waterfront

Worker 1933a), but it began the rationalization process.

The union’s newfound power was put to great effect over the following years to

reinforce gains and expand its influence. When the 1934 contract expired in 1936,

a 92-day strike ended in an arbitrated settlement that maintained the longshore-

men’s earlier gains and established two more gains. First, preferential hiring was won

for union members, which deepened workers’ control of the hiring process. Second,

“a coastwide joint labor-management committee was established which negotiated

coastwide work load limits and a uniform set of safety rules” (Kahn 1980, 379). This

solidified the ILA’s territorial monopoly over the entire West Coast port range and

“meant that employers could not undercut the union in one West Coast port by going

to another one” (Kahn 1980, 380).

8.4 Mid-century: Spheres of dominance affirmed

Following these victories, the ILA employed direct action to win major gains. Gains

were made primarily in sling load size, gang size, and safety. All three of these foci

were oriented toward augmenting the basic rationalization established through the

low-man-out system. As these gains bit into employers’ profits, employers’ discontent

211

grew and they strove unsuccessfully to reassert their dominance over the waterfront.

According to Finlay (1988, 45–50), the union had taken a “substantial degree of

control over production” by 1937. They managed to limit the number of cases, boxes,

barrels, and sacks of particular goods that could be loaded on a sling and to cap the

general sling load at 2,100 pounds, which was a major decrease from the 3,000 plus

pounds they had often been handling. In an industry notoriously dangerous (as noted

above), they also gained the right to stop work if they felt conditions were unsafe and

endangered their health. The union was also able to gain an agreement on the number

of holdmen hired to load and unload ships, the most arduous job, and for other roles.

Finlay rightly associates the first and third of these gains with work pace, but all

three also contribute to the longshoremen’s rationalization of labor. By establishing

a basic work pace that obstructed employers’ ability to speed up work by increasing

sling loads, the number of workers was kept proportional to the amount of cargo, re-

sulting in more predictable job opportunities. Manning scales also clearly promoted

the same stability in employment by fixing the number of workers per job. The con-

nection of safety to rationalization is less direct, but serves to boost job security and

intertemporal distribution of wages by increasing the likelihood that any particular in-

dividual will be capable of ongoing, long term work, rather than being suddenly taken

out of the workforce through injury. The union’s struggles, thus, further rationalized

their work, stabilizing job opportunities for union members.

Employers did not take well to this loss of control over production on the docks.

Discontent rose rapidly, especially due to rising labor costs. The struggle between

the employers and the employed came to a peak in the 1948 contract negotiations,

when employers tried to use the anti-communist provisions of the Taft-Hartley Act

to demand that the union relinquish control of the dispatch halls (Finlay 1988, 48;

Schneider 1959, 554). After a lengthy strike, a more liberal group of employers led

212

by Joseph Paul St. Sure gained control of the employers’ bargaining unit and agreed

to union control of the dispatch halls and low-man-out dispatching in exchange for

the union’s agreement to stop job actions and wildcat strikes. The agreement was

a turning point in employer-labor relations. “The unions’s control of the hiring hall

was recognized and the employers’ control of production was reaffirmed”(Finlay 1988,

48). “Control of production”refers to the increased predictability of freight movements

through the reduction in job actions. This is another movement toward rationalization

for the employers, as was the affirmation of the union’s dominance over hiring.

As discussed previously (Section 3.2.2), efforts to containerize cargo stretched back

to the early twentieth century. They took on renewed vigor after dockworkers had

asserted their dominance over production on the docks. As Cargo Handling, an indus-

try journal targeted toward employers from the period, summarized in 1956, “Labour

costs in the United States have always been very high, and since the war have risen

to such an extent that mechanical handling of materials and finished products in fac-

tories and warehouses, at docks and railways terminals, has become almost universal”

(Cargo Handling 1956b). As labor gains increased labor costs for shipping compa-

nies, this quote indirectly conveys that mechanization was an attempt to subvert

union control of the docks. In the same journal, Tooth (1956) advocates for increased

use of forklifts dockside through an enthusiastic description of a company that cut

down railcar loading from 48 men to eight men through the use of forklifts.

8.5 Mechanization and modernization

While McLean was shaking up the East Coast, Matson Navigation Company was

moving steadily toward employing containerization for its routes from the West Coast

to the Hawaiian Islands. The motivation was labor costs. The company hired Foster

213

Weldon, a Johns Hopkins University geophysicist and pioneer in operations research,

who determined that almost half of Matson’s existing door-to-door shipping costs

were due to labor. “[T]his cost has increased steadily in the past and will continue to

do so indefinitely as long as the operation remains a manual one. There is certainly

no indication of a change in the current trend of spiraling longshore wages with no

corresponding increase in labor productivity” (Weldon 1958, 652–653). The only

solution deemed viable was automation. After more than two years of painstaking

research, cautious experimentation, and financial calculation, Matson finally intiated

its container service on 31 August 1958 from San Francisco. By this time the company

had begun developing fully dedicated ships for the trade, putting them into service by

1960. With Matson’s more conservative endorsement of a full embrace of containerized

intermodal transport and McLean’s flashy intervention into the East Coast market,

the industry’s shift toward containerization had been solidly established.

8.5.1 The Setting: West Coast weakness

The West Coast ports are now commonly associated with Asian trade. However, un-

til containerization and globalization, the volume was minimal and the bulk of West

Coast trade consisted of intracoastal lumber and goods shipping and the Hawaiian

trade. The military demands of the Pacific Theater in World War II and the Ko-

rean War refocused the orientation of transport, creating a huge boom in maritime

employment on the West Coast. After the second conflict subsided, however, port

throughput dropped precipitously. “By 1955 most Pacific ports were experiencing at

least a twenty-five percent decrease in tonnage handled by the port.” (Winter 1991).

Schneider (1959, 553) suggests that despite temporary boosts provided by wartime

activites, this was part of a long decline as transcontinental movements shifted to

214

truck, pipelines, and rail to escape water transport’s comparatively high cost.

“However,” Winter (1991) writes, “the employers equated union non-compliance

with the longshoring contract as the primary cause for the economic decline.” This

view is reinforced by Finlay’s (1988) statement that employers launched a series of

“Performance and Conformance” campaigns to eliminate practices that violated the

contract. But this is really one part of the larger problem of stagnating or declining

productivity on the waterfront. Figures drawn from Fairley (1979) and compiled by

Finlay (1988) bear out such claims, showing virtually no change in weighted tonnage

moved per manhour during the 1950s. This makes sense given that methods in

practice were not changing, but given rising labor costs, this would imply a decline

in capital’s productivity. And this became the central issue.

In 1957, William Roth, a member of the Matson Navigation Company’s operations

research team clearly identified the problem and the solution from the employers’

perspective.

The basic problem of American-flag shipping. . . is the question of produc-

tivity. You will appreciate that spiraling labor costs in American shipping

may be different in intensity but no different in principle from the prob-

lem faced by any other American industry, especially in the years since

the Second World War. Other businesses, however, such as the automo-

bile and steel industries, have been able to offset increased labor costs

to a great extent by a comparable increase in the productivity of each

worker.. . . Rather than aggressively pursuing new methods of doing their

proper work, shipping companies have met increased costs either by re-

lying on government assistance, raising their rates, being content with a

smaller margin of profit, or going out of business. In some companies all

215

four of these things have taken place in the order named. (Roth 1957,

105–106)

Meanwhile, according to Kossoris (1961, 3), director the Western Regional Office of

the Bureau of Labor Statistics and respected by both employers and unions, “union

leadership was not unmindful of the fact that high labor costs were driving a con-

siderable volume of coastal and intercoastal cargo to rails and trucks. They also

realized that changes in operating procedures were creeping up on the union, slowly

but surely, and that the union was losing ground.” Amidst the late 1950s atmosphere

of interunion cooperation in the maritime crafts (Schneider 1959, 556), Harry Bridges,

head of the ILWU, first tried to address the union’s concerns by working with the

California Teamsters and the ILA to establish a transportation union. However, this

made little progress and Bridges decided that the ILWU had to work with the employ-

ers to revive port activity on the West Coast if they were going to boost employment

for longshoremen and secure some share of the productivity gains, which he called

“the men’s share of the machine.”

Informal negotiations between the ILWU and the PMA that began in 1957 led to

an agreement in 1959 that in exchange for a payment from the PMA of $1.5m, the

union would “go along with any and all mechanization during the 1959–60 contract

year; but all restrictive rules were to remain in full effect” (Kossoris 1961, 4). The ulti-

mate objective of the agreement was to share savings generated through all increases

in productivity, primarily the elimination of restrictive rules and mechanization. The

union left the nature of the sharing vague, pending the results of calculations made

by Max Kossoris, who was lured to a one year hiatus from his position at the Bureau

of Labor Statistics to generate measures of man-hours.

Over the year covered by the agreement, the ILWU determined how it would use

216

the employers’ fund. Continuing the push toward rationalization, the union targeted

the funds for a guaranteed annual income and early retirement (Kossoris 1961, 5).

Union leaders were anticipating some loss of employment with automation based on

a 1957 report by the ILWU’s Coast Committee. The report incorrectly argued that

mechanization would progress slowly due to the industry’s disorganization, which

refers to the great variety of cargo moved and equipment employed (Finlay 1988,

64–65; cf. Picard 1967 on disorganization).

When negotiations resumed in April 1960, the union requested a one year con-

tinuation of the above Understanding in exchange for $3m, so that accurate man-

hour figures could be established. However, 17 May 1960 negotiations revealed that,

though sharing the gains had been widely accepted over the previous two years, em-

ployers were no longer interested in this approach. Early in 1960, some of the larger

steamship companies had reconsidered and decided that sharing gains was “an inva-

sion of management’s prerogatives and consequently was completely unacceptable”

(Kossoris 1961, 6). “Instead the employers’ position was: How much will it cost us

to get rid of the restrictive rules and get a free hand in the running of our business?”

(Kossoris 1961, 5). Management had decided that it would “buy out” the union to

end its restrictive practices and its opposition to mechanization.

The six-year agreement ultimately signed in 1960 represents a compromise that

allowed for the rationalization of dock operations and dock work. Kossoris (1961)

lists the primary terms of the agreement: 1. Employers were not required to hire

unnecessary men; 2. Slingload limits were to only apply to longshoremen-built loads

under same method of handling, which meant that methods introduced after 1937

were not restricted beyond safety and speedup concerns; 3. There was to be no

multiple handling of loads; 4. Minimum gang size was to be reduced below present

practice; 5. Manning for new methods operations was to be negotiated with the union

217

or follow the employer’s will with redress sought through grievance mechanisms; 6.

A fund was to be created, consisting of the previously paid $1.5m plus $5m per year

for 5.5 years for a total of $29m; 7. The fund was to be managed by the ILWU and

split into two parts, one for providing a guaranteed income if work opportunities were

to drop due to the new contract (though not to curtailed economic activity) and one

for retirement; and 8. Employer obligations were to be reduced during union-caused

work stoppages in violation of agreement.

Though the agreement meant closing the rolls of the union and allowing its mem-

bership to shrink by attrition (to match reduced labor demand as mechanization

progressed), the agreement secured the goals of rationalization sought after for half

a century. Men were guaranteed a secure annual income and retirement pay, which

more evenly distributed their labor and income across individuals and across time.

Data from Finlay (1988, 65) shows a tight, normally shaped distribution of income

for 217 men centering on a healthy, middle-class annual income of $25,000 in 1979,

as opposed to the daily, weekly, monthly, and annually uncertain proceeds of earlier

periods. The agreement simultaneously opened the door for the mechanization of

cargo handling, including containerization, which greatly facilitated throughput and

reduced both labor costs and dock congestion for the employers. Finlay (1988, 61)

shows a 550 percent increase in productivity (tonnage to man-hours) from 1960 to

1980. Though the agreement eventually worked out to the greater benefit of the em-

ployers, its importance for the workers cannot be dismissed and their benefits may

have been greater if their assessment of mechanization’s impacts had been more ac-

curate. Labor’s general level of satisfaction with the situation is exhibited by the

absence of any serious labor actions from the date of the agreement until 2002.

218

8.5.2 East Coast

The dominance of organized crime over the longshoremen of New York and New Jersey

and the ILA’s internal instability led to a different and delayed historical trajectory

for the rationalization of those workers’ jobs. Johnson (1950, 91–92), whose reporting

was turned into the film On the Waterfront, describes the local context:

The Port of New York, the greatest in the world, is an outlaw frontier.

Murder on the waterfront is commonplace, a logical product of widespread

gangsterism. Organized crime and racketeering add literally millions of

dollars annually to the cost of the port’s shipping. Pier facilities, rep-

resenting an investment of almost a billion dollars, are controlled by ex-

convicts and murderers who compete for the lucrative dock rackets. This

situation of the piers, which has existed for years, is made possible by a

powerful labor union, the International Longshoremen’s Association, an

American Federation of Labor affiliate. Gangsters have attained official

positions in the locals of this union. The union leaders condone water-

front crime and racketeering, protect the racketeers, and foster unhealthy

labor practices. Rank and file union members are the principal sufferers.

The longshoremen, who load and unload the ships in the world’s richest

port, are casual workers living in an atmosphere of fear and insecurity

and exploited by corrupt and indifferent labor leaders. Neither the union

nor the industry has ever shown any decent regard for the welfare of the

longshoremen.

The moderately successful solution to this corruption was the state implementation of

the Waterfront Commission in 1956, which took over the responsibility of dispatching

219

workers (Jensen 1974).2

The disorganization produced by the corruption inquiries and indictments, in-

cluding that of long term president Joseph Ryan in 1953, further fragmented an

already geographically splintered union. Unlike the West Coast longshoremen, who

only had to negotiate with one coastwide employers association, Atlantic and Gulf

Coast longshoremen dealt with a number of independent employers associations. As a

consequence, prior to 1956 each union local was compelled to negotiate its own labor

contracts (Levinson 2006, 118), though negotiations often took place concurrently.

Though Atlantic and Gulf ports generally followed the New York local’s lead with

regard to terms, they were not bound to do so.

Anthony Anastasia, emboldened by a general movement of freight handling away

from the Manhattan docks to Brooklyn and New Jersey, took advantage of the power

vacuum following the corruption proceedings to try to pull his Local 1814 in Brooklyn

out of the ILA. The American Federation of Labor also strove to take advantage

of the situation by establishing its own International Longshoremen’s Association

(AFL-ILA) to replace the old ILA, which had been expelled from the AFL in 1953

for “having engaged in a host of nefarious and undemocratic processes” and now

referred to itself as the ILA-Independent (ILA-IND) (Herod 2001, 109). The ILA-

IND fought this challenge by seeking to shore itself up and better represent its rank

and file by expanding its formal territorial monopoly beyond the Port of New York

by adopting the AFL-ILA’s popular call for national bargaining on the grounds of

a longstanding policy of equalizing wages across ports and the need to successfully

wage coastwide strikes rather than local efforts, which were easily avoided by shippers

(Herod 2001, 110). The ILA-IND’s efforts succeeded in establishing a regional Master

2It can only be considered moderately successful, as officials at the Commission eventually cameunder the influence of organized crime.

220

Contract system for North Atlantic ports in 1956 that covered a handful of core items,

including wages, hours of work, length of contract, and employer contributions to

welfare and pension funds. Other items were still to be negotiated on a port-by-port

basis. Facing the onslaught of mechanization, over the next fifteen years organization

around a guaranteed annual income finally enabled the union to expand its territorial

monopoly over all ports from Maine to Texas.

Guaranteed Annual Income Negotiations over mechanization began concurrently

with those on the West Coast, but union fragmentation resulted in a fractious, de-

layed, and fragile fifteen year movement toward a guaranteed annual income (GAI).

During the 1956 contract negotiations that established the North Atlantic master

contract, the New York Shipping Association, which controlled 85 percent of the East

Coast shipping business (Herod 2001, 110), proposed terms that would allow them to

hire only as many longshoremen as they required for any newly introduced handling

operation (Levinson 2006, 103). The ILA-IND absolutely refused this proposal as the

AFL-ILA threat and internal strife were absorbing all the union’s energies, and the

issue was shelved until the 1958 negotiations. By that time the ILA-IND had man-

aged to fend off the AFL-ILA and turned its attention to the rapid mechanization of

cargo-handling processes.

In November 1958, when Grace Lines anchored a ship and asked to hire only five

or six men per hatch for its new system that allowed cargo to be simply rolled in or

out of the side of the ship, tensions mounted to the point that the ILA called a strike.

Though the union agreed through arbitration to handle containers from shippers who

had used them before the strike, negotiations to settle terms for other container ship-

pers lumbered along for months as the ILA was sidetracked by internal concerns. A

strike as the contract expired in September 1959 led to a general agreement in Decem-

221

ber that paralleled the recent understanding reached on the West Coast. Employers

would be allowed to automate in exchange for protecting longshoremen’s incomes.

However, an arbitration team was unable to settle on details until autumn of 1960,

when they agreed in a two-to-one vote that employers in the Port of New York would

pay a per ton surcharge for goods carried by container in exchange for the right to

use container-handling equipment without restriction (Levinson 2006).

The compromise was followed, however, by a drop in port traffic as the economy

slowed down in 1960–61. As a result, the number of men whose work had declined

and who required the container royalty payouts increased, but there were no royalties

to pay out. As traffic on the Manhattan docks continued to slow down and talk of

replacing some piers with the World Trade Center began to circulate, job security

took over as the top priority. During the 1962 contract negotiations and the following

couple of years, negotiations were heated and marked by several strikes as the union

refused to“sell out jobs like Bridges did”(Gleason speech to World Trade Club in 1962,

quoted in Jensen 1974, 269). However, rising fear that automation would destroy the

union was discussed at an ILA conference in June 1964 and led to a conciliatory

admission by the union that it was time to negotiate a guaranteed annual wage.

After contract negotiations in 1964 led once again to strikes and arbitration, the

New York ILA local and the New York Shipping Association finally reached an lo-

calized agreement outside of the nationally negotiated Master Contract. The ILA

was to receive a rather sizable wage and benefit increase coastwide in exchange for

reducing the gang size for general cargo to seventeen men by 1967. Also, from 1966

employers would pay a royalty on every container that passed through the port into

a fund dedicated to guaranteeing a minimum annual income equivalent to a workload

of sixteen hundred hours for qualified longshoremen. As Levinson (2006, 123) notes,

“A union flyer summed up the huge changes that the new contract would bring: ‘This

222

agreement takes the industry from a completely casual workforce to a stable, secure

livelihood’.”

During the 1968 contract negotiations, the ILA sought to include a number of new

provisions in the Master Contract, including the GAI. Outport employers (i.e., those

located in ports other than New York-New Jersey) objected to including the GAI,

which they saw as serving the NYSA’s needs and not their own. Unable to overcome

their opposition, the union and the NYSA agreed to significant wage increases and

other benefits for the union, a move that served to unify labor and fracture capital.

Perceiving the NYSA’s concessions as an attempt to buy off the union’s opposition

to containerization, outport employers concluded that the NYSA was reaching agree-

ments that led to outport employers subsidizing containerization efforts in New York.

In the North Atlantic, employers believed that subsidizing New York’s containeriza-

tion efforts undermined their own ports’ viability by luring away business. In the

South, employer opposition was rooted in the fact that employers and dockers han-

dled many more bulk agricultural products than manufactured items and believed that

containerization was of little relevance to them. Thus, preceding the 1971 contract

negotiations, outport employers broke away to form the Council of North Atlantic

Steamship Associations (CONASA) and the South Atlantic Employers’ Negotiating

Committee (SAENC) to negotiate for employers in their respective regions and pro-

posed to negotiate independently of the NYSA. (Herod 2001)

The union accepted this offer with the proviso that the GAI and two other items be

included in the master contract. However, despite CONASA and SAENC’s resolute

opposition to the GAI’s inclusion, the union continued to negotiate with the new

organizations and decided that locals would negotiate for the GAI independently on

a port-by-port basis. Over a few weeks of negotiations, union locals succeeded in

forcing SEANC to agree to terms identical to those in New York and CONASA to

223

agree to roughly half the number of guaranteed hours. (Herod 2001)

Following this success, the ILA allowed the 1974 negotiations to pass quietly, but

it was determined to establish a national guaranteed annual income in 1977 as the

final step in rationalizing labor on the Atlantic and Gulf Coasts. The union was

concerned that individual ports might not be able to meet their GAI obligations if

port traffic were to decline or shift to another port, as employers sought economies of

scale by concentrating operations at large container ports like New York and closed

down smaller outport operations. South Atlantic, Gulf Coast, and North Atlantic

outport employers all opposed the ILA’s insistence on including guaranteed annual

income as a master contract item. The NYSA, which had historically opposed national

contracts for fear of having its own operations held up while outport negotiations were

completed, however, had lost this fear as the union made a new effort to negotiate all

contracts at the same time. The NYSA offered its support for a national agreement,

hoping that it could also shift some of the costs of containerization in New York to

other ports. (Herod 2001)

In October 1977, unable to overcome non-NYSA employers’ objections to including

the GAI in the master contract, the NYSA announced that New York containerized

carriers would create a separate fund to achieve the same end. Since the small number

of containerized carriers were only a small portion of all freight shippers serving the

port, they were unlikely to be recognized as a representative bargaining actor by the

NLRB. Thus, the carriers chose not to negotiate directly for the master contract,

opting instead to create a supplementary agreement that would be called the Job

Security Program (JSP). The agreement provided a common shortfall fund while

allowing employers to negotiate and fund benefits at the local level, allowing both

security and flexibility. The supplementary agreement was tied to the master contract

by adding a codocil to the master contract granting the union the right to refuse to

224

work ships owned by carriers who did not subscribe to the master contract and the

JSP. Despite the questionable legality of such a clause, the threat of a lengthy strike

was sufficient to persuade all employers’ assocations to sign on. A form of coastwide

bargaining by dockworkers had finally been achieved and had supplied the foundation

for a rationalization of income across workers and over time. (Herod 2001)

Unfortunately, this hard-won unity broke down ten years later as the industry

began to change and employers more actively utilized non-union labor in the right

to work states of the South and the Gulf. While the ILA leadership pushed contin-

uously for higher wages and greater benefits, South Atlantic dockers began to worry

that wages and benefits negotiated with regard to conditions in New York and New

Jersey were undermining their own competitiveness. Work was being shifted to non-

union employers and workers (cf. Erem and Durrenberger (2008) for a compelling

account one episode in this struggle). And there was fear that cargo was going to

shift to other ports. In early 1986, these concerns manifested themselves in a defi-

ance of ILA President Gleason by West Gulf and South Atlantic locals, who made

concessions on wages and the GAI. During master contract negotiations later that

year the JSP was abolished and wages were frozen for the first time since 1949. Over

the next decade, the ILA abandoned its policy of “one port down, all ports down”

and relinquished negotiations to locals. Meanwhile, as described in 3, the shipping

industry went through a period of consolidation that facilitated the consolidation of

carriers, stevedores, marine-terminal operators, and local port employer associations

into the USMaritime Alliance, Ltd. (USMX), which emerged as the sole bargaining

agent from Maine to Texas. (Herod 2001)

225

8.6 Conclusion

The evolution of labor-capital relations on the docks offers a positive precedent for ra-

tionalization in cargo handling, as it relieved congestion both in the ranks and on the

docks. At the beginning of the twentieth century, the stacking and storage of goods on

the docks raised costs and lowered efficiency as goods were double- and triple-handled

before leaving the docks; and an oversupply of labor reduced longshoremen to subsis-

tence incomes. Through relentless organizing, coastwide bargaining was established

first on the West Coast and then on the East and Gulf Coasts. Union efforts led first

to a more equitable distribution of work throughout the longshore community and

ultimately to a guaranteed annual income. In exchange, employers obtained the right

to introduce technology that greatly increased throughtput, reducing congestion on

the docks.

Rationalization of the waterfront entailed different gains for employers and em-

ployees. For employers, rationalization consisted of more efficient cargo handling,

which reduced costs and increased volume, thereby increasing profits. For employees,

facing job loss through mechanization, rationalization necessitated overcoming the

uncertainty and irregularity of employment by establishing organizational procedures

for controlling labor supply that would ensure regular employment and a minimum

annual income. The mutual gains obtained through rationalization, though purchased

at some cost by both parties, illustrate the positive potential of rationalization efforts

in resolving difficulties facing employers and employees. Indeed, the perceived suc-

cess of the arrangement led the International Labour Organisation (ILO) to adopt the

general framework as its formal recommendation to ports facing the same decision

(International Labour Office 1973).

One of the key lessons it offers is the importance of establishing a territorial

226

monopoly over a port range in strengthening an actor’s bargaining position. Ship-

ping companies and shippers have historically enjoyed this advantage, as the mobility

of their ships and their presence in multiple ports allowed them some ability to circum-

vent and undermine localized labor disputes. However, successful organizing created

a territorial monopoly over the “effective area of production” in cargo handling by

coordinating the longhsoremen along an entire coast. Once employers lost the ability

to evade labor’s control over the ports, it was compelled to make concessions that

improved the safety and livelihoods of longshoremen and their families.

As suggested by Offe (1985), such organization requires some basic unity of in-

terest. He suggests that the wide range of concerns that affect workers, like wages,

housing, and families, increases the difficulty of organizing successfully, while em-

ployers interests are more easily organized because their interests are simpler: profits.

The story presented here complicates this claim and suggests that spatially influenced

historical trajectories play a significant role. The East Coast case reflects the fragmen-

tation of interests and consequent weakness of labor. Corruption in the union’s lead

local in New York and geographically distinct cargos that lent themselves differently

to containerization divided the East and Gulf Coast union locals and delayed labor

gains by decades. Meanwhile, the long history of extraordinarily exploitative and

unsafe working conditions on the docks unified interests and facilitated organization

on the West Coast. Spatial differences also impact the presumably simple interests

of employers. West Coast employers remained fairly unified and, from the late 1940s

as their territorial monopoly over coastal transportation was being undermined by

railroads and trucking, their interest in peaceful relations that would facilitate reli-

able performance induced them to be more progressive in sharing productivity gains

with longshoremen. On the East Coast, employer unity weakened as employers broke

ranks because the North Atlantic interest in making concessions in exchange for con-

227

tainerization appeared to demand that South Atlantic and Gulf ports, which dealt

more in bulk commodities like grain, coal, and oil, subsidize North Atlantic ports.

This weakness, combined with a newfound unity in the ILA, led ultimately to the

establishment of a coastwide guaranteed annual wage. This suggests that employers’

bottom line is at times sensitive to the production processes that generate that profit.

In all these cases, the strength gained from establishing and maintaining a territorial

monopoly depended upon defining a common interest across the port range.

Common interest itself is not sufficient to rationalize. As the study reveals, those

interests must also be organized. And this requires some degree of centralized decision

making. Establishing a territorial monopoly in the first place required coordinating

labor actions across multiple ports. Then, having established that monopoly, the

first step in rationalizing labor required central dispatching halls to track jobs and

workers and to equitably distribute the one to the other. The degree of centralization

remains unclear, however. While most decisions on the East Coast were made by

the New York local leadership and simply echoed in outports, West Coast decisions

required approval of a majority of the rank and file in all ports. So strength in cargo

handling requires not only common interest across a port range but also some degree

of centralization in organization across that space.

The mechanization and modernization agreements effectively resolved the major

tensions between employers and longshoremen. As the labor movement generally

entered a period of quiesence, rationalization established a degree of efficiency and

certainty of income that effectively met the needs of both workers and owners. In

doing so, it freed shipping companies to turn their attention to extracting greater

profits from port authority concessions. Fortunately the earlier experience of labor’s

effort to gather the strength required to negotiate on a level with capital points

to two guiding principles for port authorities. First, their strength in dealing with

228

attempts to initiate interport competition will arise from establishing a territorial

monopoly over their respective port range. Second, decision making with regard to

port operations and investments will have to attain some level of centralization, likely

either through the federal government or some voluntary association like the American

Association of Port Authorities.

229

Chapter 9

Conclusion

This dissertation has explored the territorial and technological strategies employed by

companies, workers, and governments to alter the economic terrain in the pursuit of

accumulation within their respective circuits of capital. Specifically, it has traced the

evolution and impact of intermodalism over the twentieth century on the spatial and

organizational structure of freight transportation in the U.S. with respect to global

shipping companies, longshoremen, and port authorities in the hope of offering port

authorities strategies that will help them avoid the destructive impacts of incentive

competition. Here I will summarize those findings, make policy recommendations,

and offer some concluding remarks.

9.1 Summary of findings

Tensions in the shipping industry were generated originally between employers and

workers. Though workers often organized around highly localized sites, shipping

companies could readily avoid work actions by moving their freight through nearby

facilities. Organized labor’s ability to control the labor market at a single mechanical

230

break in the flow of freight proved insufficient for attaining longshoremen’s goal of

living wages and job security. To counter shipping companies’ power, longshoremen

had to scale up their territorial monopoly from single piers or ports to entire port

ranges. They were subsequently able to make significant gains toward rationalizing

the workplace and establishing job security.

To counter these advances and to reduce turnaround time in port, shipping com-

panies pursued the technological solution of a system for packing goods outside the

longshoremen-controlled ports in large, standardized containers that could be hoisted

by crane from trucks or railcars to ships’ holds and vice versa. The success of this

second-order system for moving freight had three unintentional but interrelated ef-

fects. First, the economic terrain shifted as stages of transportation moved inland

away from ports. Second, the large investment required to take advantage of con-

tainer shipping’s inherent economies of scale and intermodalism’s global reach en-

couraged the concentration of capital through mergers, acquisitions, and alliances.

Third, the resulting logistics firms developed the capacity to serve a given product

market area from a number of ports. Though ports had previously enjoyed a virtual

territorial monopoly of the product market area that consisted of their hinterlands,

logistics firms’ adoption of intermodalism expanded the area of effective production

for transportation services to a given port’s product market area. The consequent un-

dermining of port authority’s territorial monopolies over their product market areas

compelled them to compete through massive subsidies to attract the flows of capital

generated by the scarce large alliances or firms.

The competition in 1998–1999 between Halifax, Baltimore, and New York-New

Jersey illustrated the destructive effects of such competition. First, it showed that

Maersk and Sea-Land were still attracted to the product market area immediately

accessible from the terminals in New York and New Jersey and that subsidizing them

231

to stay was a waste of regional resources for the goal of job creation. Second, port

authorities with access to government funds and broader political mandates, particu-

larly Baltimore, were able to marshall significant resources. Third, fragmentation of

labor by local interests led to diminished working conditions, even for the losing port.

Regression analysis demonstrated that the supposed pay-off of job creation and

industrial support for success in subsidy competition is no longer defensible. First,

intermodal efficiencies have shifted the locus of job generation in freight transportation

away from ports to inland warehouse districts. Second, the mechanization of freight

movement has steadily eroded employment outside of warehousing, a trend that is

likely to continue as terminals become fully automated in the future. Third, the jobs

that are created tend to pay exploitative wages to the non-unionized and, in trucking,

can be positively dangerous to truck drivers and others on U.S. highways. Finally,

there is no evidence that increasing port traffic has any impact on the location of

employment in sectors outside of transportation either. That is, intermodalism has

so reduced the costs of transportation that firms no longer weight transportation costs

as highly in their location decisions.

9.2 Ongoing relevance

Though the dissertation focuses on an incident now more than a decade in the past, the

findings are becoming increasingly more relevant. Most immediately, the PANYNJ

is now moving forward to raise the Bayonne Bridge to allow larger ships to pass

under. Also, the PANYNJ’s lease with Maersk and Sea-Land is already a third done.

Renegotiation talks are already on a distant horizon. But contracts continue to expire

and to be renegotiated at ports throughout the world. Such opportunities allow global

shipping companies to continue to demand incentives that come at the cost of local

232

taxpayers and users.

Second, in the U.S., the continental dynamic in interport competition is growing.

A recent article in the Los Angeles Times (White 2010) reports that the once dominant

ports of Los Angeles and Long Beach can no longer take their territorial monopolies

for granted. “The days of Los Angeles and Long Beach being the one big river

for trade, with just trickles for everyone else, are over. Now you have as many as

seven contenders vying for the same business, and each one of them has big plans”

(White 2010, Asaf Ashar quoted in). Retailers’ “four corners” strategy of shipping

goods flexibly through ports on the West Coast, the Gulf Coast, the East Coast, and

Canada has continued to weaken port authorities relative to shippers and shipping

companies. And the scheduled widening of the Panama Canal in 2014 will increase

the range of post-Panamax ships, further increasing interport competition.

9.3 Policy recommendations

So what should port authorities do? Reflecting on the Maersk and Sea-Land nego-

tiations, which took place under her tenure as Director of Port Commerce for the

PANYNJ, Lillian Borrone 1999 laid out three basic options available to the port in-

dustry to respond to the growing power of global shipping companies. First, ports

can consolidate among themselves. Second, they can form alliances. And third,

they can share information. The evidence presented in this dissertation suggests that

national—if not continental—level consolidation would be the most effective option.

Borrone offers information sharing as a near term, interim policy, suggesting that

the American Association of Port Authorities or the International Association of

Ports and Harbors could develop a database on market conditions available to all

ports and could train port executives in negotiation skills. This would be in keeping

233

with recommendations made by LeRoy (2007) in Markusen (2007) for “sunshine” in

addressing interurban competition. By this, he means public disclosure of company-

specific subsidy amounts and terms. Enhancing transparency through information

sharing is intended to solve the problem of asymmetric information in which the

companies are aware of what localities are offering while the localities remain ignorant

of each others’ offers, which shifts bargaining advantage in favor of the companies

(cf. Klosterman 2003). True to her word, Borrone established such an association

for sharing rate and contract information during the Maersk and Sea-Land bidding

process and received anti-trust immunity from the Federal Maritime Commission to

do so (Brennan 1998e). It is not clear that such information-sharing arrangements

have been effective, however. The presence of the organization certainly did little

or nothing to forestall competitive bidding between the Port of Baltimore and the

PANYNJ. Similarly, archival research into the minutes of the longstanding North

Atlantic Ports Association (1999)(NAPA) revealed little of substance with regard to

information sharing activities beyond some broad discussion of trends. Despite stating

that one of the organization’s purposes was the “exchange of ideas and information on

all port activities,” NAPA does not appear to have exchanged much. These activities

may have remained informal so as to keep them off the record, they may not have been

deemed appropriate for the minutes, or perhaps ports are more niggardly with such

information than Borrone’s optimism would suggest. Therefore, this route appears

to be of limited value and unlikely to overcome the negative impacts of competition,

though it is probably an inevitable first step toward anything more substantive.

A stronger form of association would be the formation of alliances among ports.

These would mimic the carrier alliances in which two or more carriers rationalize

freight movements by sharing container slots on each others’ ships as well as termi-

nal operations. In the port context, it could entail ports collectively bargaining with

234

carriers for the provision of terminal operations in which cargo would be distributed

among ports through capacity-sharing arrangements. This is effectively a horizontal

integration across providers of similar services. This approach faces two challenges.

The first is certainly not insurmountable, but it adds to the complexity of such ar-

rangements. The potentially different administrative and institutional structures of

the ports that might participate in an alliance would greatly complicate allocation of

costs and benefits. Since some ports operate their own terminals (operating ports)

while others simply rent out their terminals (landlord ports), formulae would have to

be worked out for the allocation of costs and benefits between port authorities and

terminal operators. Additional complexity would be generated by the institutional

differences, as some port authorities are part of local or state governments while oth-

ers are independent. To this end, it may be suggested that a horizontal alliance of

terminal operators would be a more appropriate solution. The advantages to this are

evident in the recent growth of global terminal operators, like Dubai Ports World and

PSA International (formerly Port of Singapore Authority) (Heaver, Meersman, and

de Voorde 2005). However, alliances among terminal operators would simply shift

even more bargaining power away from port authorities, since the allied terminal op-

erators would more than likely be working with large shipping companies and thus

would be able to threaten to leave, much as Maersk and Sea-Land did.

The second problem is more serious. Alliances, unless established across a port

range or nationally, would fail to overcome the problems they are striving to address:

competition. As the global shipping alliances have demonstrated, the formation of one

alliance leads to the formation of competing alliances. Thus, the possibility is all too

real that one port alliance or port will seek to underbid another, much as Baltimore

underbid New York and New Jersey, and that the cycle of destructive competition

would begin anew. Lacking a territorial monopoly over the area of effective production

235

for port services, alliances would still be subject to competitive pressures. The lack of

a territorial monopoly distinguishes the present situation of ports from the PANYNJ’s

successful effort to break the airlines’ monopoly in the late 1940s. In many respects,

the position of airport operators in the 1940s resembles that of ports today. According

to Doig (2001), the airlines had negotiated highly favorable contracts with airports

by playing one off against the other when the two locations could serve the same

market, including Newark and New York. As contracts were being renegotiated in

the late 1940s, the airlines continued to demand contracts that charged them less

than cost to land. In response, the PANYNJ coupled a national effort to build a

coalition of airport authorities with a local effort to unify PANYNJ commissioners,

politicians, and the public in opposition to making concessions on the newly completed

Idlewild Airport (now JFK Airport). Though the coalition played an important role in

sharing information that led to the realization that some operators had negotiated self-

sustaining contracts and to the development of best practices, the airline monopolies

were not broken by the coalition. They were broken by the PANYNJ’s territorial

monopoly over the product market area for airline passengers. In the time since

the first concessionary leases were established by setting Newark and New York in

competition for the airlines, the PANYNJ had been given responsibility for both

airports. Therefore, if the airlines wanted to serve the New York metropolitan region,

they had to make a deal with the PANYNJ. By building a unified local front in favor

of self-sustaining airports, the PANYNJ was able to use its territorial monopoly to

overcome the airlines’ monopoly.

An alternative form of alliance that relies on vertical integration has been proposed

by Notteboom and Rodrigue (2005) and others. They suggest that ports should build

alliances with their inland freight transportation service providers, like railroads, dis-

tribution centers, and trucking firms. While horizontal alliances attempt to cut across

236

supply chains, vertical alliances strive to strengthen the competitiveness of specific

supply chains. At present, this “port cluster” strategy, which strives to mitigate issues

through consensus, may well be the easiest route forward and would surely generate

some competitive advantage. And such policies are being pursued by some ports

and argued as the best approach for others. The PANYNJ has signed a memoran-

dum of understanding with the Panama Canal Authority and has experimented with

a barge network stretching from Philadelphia to Albany to New London. However,

such alliances clearly fail to establish a sufficient territorial monopoly capable of coun-

tering shipping companies’ power. They are simply a new strategy in a destructive

competition.

Borrone’s third option is port consolidation, which would permit a proper ratio-

nalization of freight across ports. Port consolidation is effectively the merger of ports

into a single, larger entity. Borrone (1999) suggests that a number of ports have al-

ready done so, such as the Delaware River Ports, Houston and Galveston in the late

1990s, and the consolidation of ports of Hampton Roads, Virginia in 1982. The clas-

sic example of port consolidation is the Port Authority of New York and New Jersey

itself. As discussed in Section 5.2, the PANYNJ was originally formed—as were many

others—to rationalize regional transportation in the face of railroads’ intransigence

and power. This highly effective consolidation established a territorial monopoly over

regional freight transportation that permitted it to counter railroads’ opposition to

sharing track and terminal facilities and proved highly effective in reducing conges-

tion, particularly in Manhattan and on the Hudson. These are precisely the issues

facing ports today. Carriers are using their geographical mobility and size to shape

the structure of the port system, and the result is highway and rail congestion in

many of the nation’s most populated regions.

Port consolidation today would have the positive impact of guiding discretionary

237

cargo through less congested ports and regions, would have the potential of encourag-

ing economic development by increasing cargo through ports in struggling cities, and

would ensure that private freight transportation providers cover the expenses of their

own operations. In bundling the costs and benefits of port operations across all ports,

a consolidated port system would be able to limit cargo in highly congested regions

to locally destined goods and still enjoy profits from the shipment of discretionary

cargo through ports with excess capacity. This would benefit the congested region by

reducing traffic congestion and improving environmental conditions, and peripheral

ports with excess capacity would benefit from limited job creation. A consolidated

port system would also possess a territorial monopoly over the port services, like the

ILWU established, that would allow it to bargain effectively with global carriers and

make certain that local taxpayers and users are not bearing the cost of guaranteeing

private sector profits. Since any benefits to government investment in activities like

dredging, terminal construction, and intermodal connections now benefit the national

economy rather than the regional economy, a consolidated port system could subsidize

these activities if it believed they would help the nation compete with other countries.

The funding could then be derived from user fees distributed across all port users or

would give the consolidated port authority greater claim on funds from the national

treasury.

Consolidation would ideally take place at the continental level, including all ports

of the U.S., Canada, Mexico, and possibly the Caribbean, but it would still remain

highly effective at the national or port range scale. As this study has shown, Cana-

dian ports are in competition with U.S. ports. The same is true of Mexican and

Caribbean ports, though the Caribbean ports also often serve the additional function

of providing hemispheric transshipment points for North, Central, and South Amer-

ican freight. This is because global carriers are strategizing at the continental level

238

and freight can flow cost-effectively from most ports to any region of the continent

(Notteboom and Rodrigue 2005, 175). However, because the greatest population

concentrations function as the product market area for freight transportation and the

highest concentrations are coastal, an argument can be made that a coastwide, port

range consolidation would also have some potential to mute the power of the global

carriers.

Such a plan obviously faces serious obstacles. In addition to transnational legisla-

tive issues, there are more fundamental coordination challenges. If ports are unable

even to share information, how would they ever be able to consolidate? This challenge

suggests two routes forward. The first is patience and persistence in starting from

the ground up with information sharing, as Borrone suggested. Sustained efforts here

may help port authorities overcome their regional pride and develop trust and a com-

mon vision with their compatriots that could lead to increasing levels of cooperation.

After all, it did take the ILWU several decades to build its territorial monopoly. The

second option is a federally led effort, at least for the U.S. Outright nationalization

or federal mandate is probably constitutionally out of the question due to the lim-

itation of federal jurisdiction to navigable waters. However, the federal government

could develop a framework and enabling legislation to facilitate consolidation while

negotiating with and mediating between ports to encourage voluntary participation.

Even small steps toward port consolidation should encourage other ports to either

join or form their own alliances, which would still be a step toward countering global

shipping companies’ powers.

239

9.4 Planning implications

This dissertation’s findings and their suggestions for port authorities offer at least

three implications for planning. First, economic development planners in areas that

fall within the band of reconcentrated warehousing activity should engage warehous-

ing as a development opportunity. Not only does warehousing offer opportunities for

direct employment—albeit with concerns about job quality—but it could also lead to

more sophisticated jobs through the development of backward linkages. As discussed

in Chapter 6, some producers are relocating final production steps near their distri-

bution centers in their goods’ final markets. Thus, by capturing distribution centers,

an inland town or municipality also has the potential to draw these final production

steps and their accompanying jobs.

Second, planners in port regions need to rethink the role of their waterfronts and

ports. While there has been much discussion and reflection on the reuse of declin-

ing waterfronts (e.g., Brown 2008; Fainstein 2001), these considerations ought to be

redirected even to active waterfronts. I have shown that port region planners can

no longer depend on port activities to generate local employment either directly or

indirectly. If interport competition undermines ports’ ability to be self-sustaining

and leads to the consumption of local resources without generating sufficient employ-

ment to justify the costs, these municipalities should consider other potential uses for

waterfronts currently used for port activities. That is to say, port activities should

be weighed against other income- and job-generating activities in determining future

waterfront land uses.

Doing so would have two effects that suggest new activities for port region plan-

ners. The first effect would be to support port authorities in negotiations with ter-

minal operators and shipping companies by expanding the options available to port

240

authorities, thereby reducing their dependence on terminal operators for revenue.

This should pressure the terminal operators to offer better terms. The second effect

would be to draw additional funding and support from the federal government for

ports it deems vital to national economic growth. Since local residents and users now

subsidize ports that serve the nation as a whole without being compensated for their

costs, threatening to repurpose waterfront for non-port uses should provide impetus

for the federal government to take a more active role in funding critical transportation

infrastructure. Thus, port region planners should work to communicate this possibil-

ity to federal policy makers and work with those policy makers to establish a federal

framework of support for port infrastructure, which should follow the recommenda-

tions made above.

9.5 Concluding thought

Charles Horton Cooley demonstrated that transportation as a mechanical force is

in a dialectical relation with the social forces that develop it as a tool to achieve

their goals. This dissertation has shown how government organizations, labor or-

ganizations, and economic organizations have employed territorial and technological

strategies to shape the mechanical processes and physical location of transportation

to accumulate capital within their respective circuits and how these changes have in

turn influenced the strategies of the organizations themselves. Technological strate-

gies like containerization can be employed under the capitalistic logic of power to

overcome territorial strategies by transforming the economic terrain underneath es-

tablished territories and redirecting the circulation of capital. Such transformations

call for corresponding territorial responses. The primary focus of spatial strategies in

the face of increased global competition in a “regional world” has been on vertically

241

integrating translocal production networks into competing commodity chains. Here,

we have explored the possibility of horizontally integrating by building territorial mo-

nopolies over the area of effective production. As economic organizations grow to a

scale that overwhelms the power of existing territories, those territories, too, must

scale up to escape exploitative relations with capital.

242

Appendix A

Total employment estimation

Because County Business Patterns suppresses data to avoid unfair disclosure of indi-

vidual establishment’s employee count and annual payroll, it is necessary to estimate

employment totals. To accomplish this, for each year I regressed total employment

against the number of establishments in each size category where the figure for total

employment was available, employing a Poisson distribution in a generalized linear

model in order to capture the rapidly decaying count character of the distribution

and ensure that values were non-zero for potential log transformations. The result-

ing estimates, which demonstrate uniformly high significance (p ≤ 0.001), were used

to generate estimates only for those records where total employment figures were

suppressed.

The equations for these estimates follow:

1970 (based on 155,301 records out of 215,674)

tempmmest.poisson = 2.186 + 1.357 · t1to3 + 5.127 · t4to7 + 11.78 · t8to19 + 30.52 ·

t20to49 + 68.79 · t50to99 + 148 · t100to249 + 339 · t250to499 + 1388 · t500more



243

t20to49 + 72.26 · t50to99 + 155.65 · t100to249 + 354.92 · t250to499 + 698.04 · t500to999 +

1234.82 · 1000to1499 + 1907.95 · t1500to2499 + 3382.66 · 2500to4999 + 9259.06 · t5000more




1240.53 · 1000to1499 + 1881.21 · t1500to2499 + 3307.09 · 2500to4999 + 9701.90 · t5000more




1233.92 · 1000to1499 + 1902.75 · t1500to2499 + 3403.51 · 2500to4999 + 9874.01 · t5000more




1234.01 · 1000to1499 + 1902.19 · t1500to2499 + 3366.58 · 2500to4999 + 8710.92 · t5000more

1994 (based on 385,356 records out of 1,010,591)



1233.64 · 1000to1499 + 1899.64 · t1500to2499 + 3413.13 · 2500to4999 + 8531.05 · t5000more




1257.79 · 1000to1499 + 1924.30 · t1500to2499 + 3357.32 · 2500to4999 + 8513.24 · t5000more




1221.76 · 1000to1499 + 1891.65 · t1500to2499 + 3410.86 · 2500to4999 + 9617.25 · t5000more

244




1208.59 · 1000to1499 + 1924.77 · t1500to2499 + 3420.38 · 2500to4999 + 10108.15 · t5000more

245

Appendix B

NAICS and SIC code

correspondence for selected

industries

246

Table

B.1

:P

ort

-rel

ate

din

dust

ries

wit

hSIC

and

NA

ICS

codes

and

bri

dges

indic

ate

d

Indust

ry1970

Bri

dge?

1974–84

Bri

dge?

1989–94

Bri

dge?

1999

Bri

dge?

2004–07

Port-relate

dmanufactu

ring

Oil

and

gas

pip

elin

eand

rela

ted

stru

c-

ture

sco

nst

ruct

ion

1382,

1620*

Yes

1382,

1620*

No

1382,

1398,

1623,

1629,

8741

No

213112,

234910,

234930

No

23712

Fats

and

oils

refinin

gand

ble

ndin

g2091,

2092,

2093,

2094,

2099

Yes

2074,

2075,

2076,

2077,

2079

Yes

2074,

2075,

2076,

2077,

2079

No

311225

Yes

311225

Fro

zen

spec

ialt

yfo

od

manufa

cturi

ng

2036,

2037

No

2038

Yes

2038

Yes

311412

Yes

311412

Pet

role

um

refiner

ies

2910

Yes

2910

Yes

2910

Yes

324110

Yes

324110

Basi

cch

emic

al

manufa

cturi

ng

2813,

2815,

2816,

2818,

2819,

2895

Yes

2813,

2816,

2819,

2865,

2869,

2895

Yes

**

2813,

2816,

2819,

2865,

2869,

2895

Yes

**

325100

Yes

325100

247

Indust

ry1970

Bri

dge?

1974–84

Bri

dge?

1989–94

Bri

dge?

1999

Bri

dge?

2004–07

Rubb

erand

fibre

-manufa

cturi

ng

2821,

2822,

2823,

2824

Yes

2821,

2822,

2823,

2824

Yes

2821,

2822,

2823,

2824

Yes

325200

Yes

325200

Ship

buildin

gand

repair

ing

3731

Yes

3731

Yes

3731

Yes

336611

Yes

336611

Com

mer

cial

and

indust

rial

mach

in-

ery

and

equip

men

tre

pair

and

main

-

tenance

7623,

7694,

7699

Yes

7623,

7694,

7699

Yes

7623,

7694,

7699

No

811300,

811490

No

811300

Port-relate

dtrade

Met

aland

min

eral(e

xce

pt

pet

role

um

)

mer

chant

whole

sale

rs

5091

Yes

**

5050

Yes

5050

Yes

421500

No

423500

Tra

nsp

ort

ati

on

equip

men

tand

sup-

plies

(exce

pt

moto

rveh

icle

)m

erch

ant

whole

sale

rs

5088

Yes

5088

Yes

5088

No

421860

No

423860

Rec

ycl

able

mate

rial

mer

chant

whole

-

sale

rs

5093

Yes

5093

Yes

5093

No

421930

Yes�

423930

Chem

ical

and

allie

dpro

duct

sm

er-

chant

whole

sale

rs

5029

No

5160

Yes

5160

Yes

422600

Yes

424600

Pet

role

um

and

pet

role

um

pro

duct

s

mer

chant

whole

sale

rs

5092

Yes

**

5170

Yes

5170

No

422700

Yes

424700

Com

modit

yco

ntr

act

sbro

ker

age

6220

Yes

6220

Yes

6220

No

523140

Yes

523140

Com

modit

yco

ntr

act

sdea

ling

6050,

6220,

6799

Yes

6050,

6220,

6799

No

6090,

6220,

6799

No

523130

Yes

523130

248

Indust

ry1970

Bri

dge?

1974–84

Bri

dge?

1989–94

Bri

dge?

1999

Bri

dge?

2004–07

Tra

nsp

ort

Dee

p-s

eafr

eight

transp

ort

ati

on

4410

Yes

4410

Yes

4410

Yes

483111

Yes

483111

Coast

al

and

gre

at

lakes

frei

ght

trans-

port

ati

on

4420,

4430,

4450

Yes

4420,

4430,

4450

No

4424,

4432,

4492

No

483113

Yes

483113

Inla

nd

wate

rfr

eight

transp

ort

ati

on

4440,

4450

Yes

4440,

4450

No

4440,

4492

No

483211

Yes

483211

Gen

eral

frei

ght

truck

ing

4210

Yes

4210

Yes

4210

Yes

484100,

484200

Yes

484100,

484200

Gen

eral

frei

ght

truck

ing,

loca

l4210

Yes

4210

Yes

4210

No

484110

Yes

484110

Pip

line

transp

ort

ati

on

of

crude

oil

4600

No

4610

Yes

4610

Yes

486100

Yes

486100

Pip

elin

etr

ansp

ort

ati

on

of

natu

ral

gas

4920

Yes

4920

Yes

4920

No

486200

Yes

486200

Oth

erpip

elin

etr

ansp

ort

ati

on

4600

Yes

4600

Yes

4600

Yes

**

486900

Yes

486900

Couri

ers

and

mes

senger

s—

——

—4210

No

492000

Yes

492000

Port

and

harb

or

op

erati

on

4463

Yes

4463

Yes

4491

No�

488310

Yes

488310

Mari

ne

carg

ohandling

4463

Yes

4463

Yes

4491

No�

488320

Yes

488320

Port

and

harb

or

op

erati

on

and

mari

ne

carg

ohandling

4463

Yes

4463

Yes

4491

Yes

488310,

488320

Yes

488310,

488320

Carg

ohandling

Nav

igati

onal

serv

ices

tosh

ippin

g4450,

4469

Yes

4450,

4469

No

4492,

4499

No

488330

Yes

488330

Oth

ersu

pp

ort

act

ivit

ies

for

wate

r

transp

ort

ati

on

4469,

4780,

7699

Yes

4469,

4780,

7699

Yes

4499,

4780,

7699

No

488390

Yes

488390

249

Indust

ry1970

Bri

dge?

1974–84

Bri

dge?

1989–94

Bri

dge?

1999

Bri

dge?

2004–07

Supp

ort

act

ivit

ies

for

road

transp

ort

a-

tion

4170,

4230,

4780,

7549

Yes

4170,

4230,

4780,

7549

Yes

4170,

4230,

4780,

7549

No

488400

Yes

488400

Fre

ight

transp

ort

ati

on

arr

angem

ent

4710

No

4710,

4723

Yes

4730

Yes�

488510

Yes

488510

Oth

ersu

pp

ort

act

ivit

ies

for

trans-

port

ati

on

4720,

4780

No

4722,

4780

Yes

4729,

4780

No

488900

Yes

488900

Logistics

Gen

eral

ware

housi

ng

and

stora

ge

4225

Yes

4225

Yes

4225

No

493110,

531130

Yes

493110,

531130

Ref

riger

ate

dw

are

housi

ng

and

stora

ge

4222

No

4222

Yes

4222

Yes�

493120

Yes

493120

Oth

erw

are

housi

ng

and

stora

ge

4226

Yes

4226

Yes

4226

Yes�

493190

Yes

493190

Sourc

e:

Base

don

de

Langen

(2007)

and

Wate

rs(1

977)

and

augm

ente

dth

rough

US

Cen

sus

bri

dge

web

site

sand

ori

gin

al

CB

Pdocu

men

tati

on.

*C

BP

com

bin

es1622,

1623,

and

1629

into

1620.

Managem

ent

serv

ices

(8741)

not

acl

ass

ifica

tion

pri

or

to1987.

**

Confiden

tth

at

the

bri

dge

exis

ts.

�Wit

hin

3p

erce

nt

erro

r.

�Sev

ente

enp

erce

nt

of

Mari

ne

carg

ohandling

(4491)

isP

ort

and

harb

or

op

erati

ons

and

83

per

cent

isM

ari

ne

carg

ohandling.

250

Table

B.2

:Sel

ecte

din

dust

ries

wit

hSIC

and

NA

ICS

codes

and

bri

dges

indic

ate

d

Indust

ry1970

Bri

dge?

1974–84

Bri

dge?

1989–94

Bri

dge?

1999

Bri

dge?

2004–07

Natu

ralre

sourc

es

Iron

ore

min

ing

1010

Yes

1010

Yes

1010

Yes

212210

Yes

212210

Heavy

industry

Fla

tco

nta

iner

manufa

cturi

ng

3211

Yes

3211

Yes

3211

Yes

327211

Yes

327211

Gla

ssco

nta

iner

manufa

cturi

ng

3221

Yes

3221

Yes

3221

Yes

327213

Yes

327213

Cem

ent,

hydra

ulic

3241

Yes

3241

Yes

3241

Yes

327310

Yes

327310

Rea

dy-m

ixco

ncr

ete

manufa

cturi

ng

3272

Yes

3273

Yes

3273

Yes

327320

No

327320

Pet

role

um

refiner

ies

2910

Yes

2910

Yes

2910

Yes

324110

Yes

324110

Reta

il

Book,

per

iodic

al,

and

musi

cst

ore

s5942,

5994,

5733

Yes

5942,

5994,

5733

Yes

5942,

5994,

5733

Yes

451211,

451212,

451220

Yes

451211,

451212,

451220

Manufactu

ring

Adhes

ive

manufa

cturi

ng

2891

Yes

2891

Yes

2891

Yes

325520

Yes

325520

Gum

and

wood

chem

icals

2860

Yes

2861

Yes

2861

Yes

325191

Yes

325191

Pharm

ace

uti

cal

pre

para

tions

2834

Yes

2834

Yes

2834

Yes�

325412

Yes

325412

Ele

ctro

pla

ting,

pla

ting,

polish

ing,

an-

odiz

ing,

and

colo

ring

3741

Yes

3741

Yes

3741

Yes

332813

Yes

332813

Turn

edpro

duct

and

scre

w,

nut,

and

bolt

manufa

cturi

ng

3451

Yes

3451

Yes

3451

Yes

332720

Yes�

332720

251

Indust

ry1970

Bri

dge?

1974–84

Bri

dge?

1989–94

Bri

dge?

1999

Bri

dge?

2004–07

Mach

ine

tool

(met

al

form

ing

typ

es)

manufa

cturi

ng

3542

Yes

3542

Yes

3542

Yes

333513

Yes

333513

Auto

mati

cen

vir

onm

enta

lco

ntr

ol

manufa

cturi

ng

for

resi

den

tial,

com

-

mer

cial,

and

appliance

use

3822

Yes

3822

Yes

3822

Yes

334512

Yes

334512

Sem

iconduct

ors

and

rela

ted

dev

ice

manufa

cturi

ng

3674

Yes

3674

Yes

3674

334413

Yes

334413

Inform

ation

industries

Per

iodic

al

publish

ers

2720

Yes

2720

Yes

2720

Yes

511120

Yes�

511120

Data

pro

cess

ing

serv

ices

7399

No

7374

Yes

7374

Yes

514210

Yes

514210

Moti

on

pic

ture

and

vid

eopro

duct

ion

7813,

7814

Yes

7813,

7814

Yes

7812

Yes

512110

Yes

512110

Sourc

e:

US

Cen

sus

bri

dge

web

site

sand

ori

gin

al

CB

Pdocu

men

tati

on.

�Wit

hin

3p

erce

nt

erro

r.

252

Appendix C

Industrial sector regression

analyses

253

1970

1974

1979

1984

1989

1994

1999

2004

2007

(Intercep

t)-5.06***

-8.89***

-6.25***

-4.97***

-7.31***

-9.82***

-4.13**

-9.58***

-12.6

***

1(0.87)

(1.12)

(1.33)

(1.2)

(1.52)

(1.63)

(2.03)

(2.77)

(3)

popden

sity

log

0.24***

0.45***

0.46***

0.37***

0.51***

0.56***

0.62***

0.88***

0.92***

2(0.01)

(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

(0.03)

(0.03)

incper

caplog

0.56***

0.99***

0.46**

0.36**

0.48**

0.76***

0.07

0.67**

1.06**

3(0.12)

(0.15)

(0.16)

(0.15)

(0.18)

(0.19)

(0.23)

(0.31)

(0.33)

taxrate

-0.02**

-0.02**

-0.01

-0.01

-0.01

0-0.02*

-0.02**

-0.04**

4(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

ba

-0.04***

0.03**

0.02**

0.02**

0.05***

0.06***

0.08***

0.07***

0.05***

5(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

hs

00

0.03***

0.03***

0.03***

0.03***

0.04***

0.03***

0.02**

6(0)

(0.01)

(0)

(0)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

nonwhite

0.01***

0.02***

0.02***

0.01***

0.01***

0.01***

0.01***

0.01***

0.01***

7(0)

(0)

(0)

(0)

(0)

(0)

(0)

(0)

(0)

foreign

0.1

***

0.11***

0.12***

0.11***

0.1

***

0.1

***

0.07***

0.07***

0.06***

8(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

port

0.05***

0.07***

0.05***

0.04***

0.04***

0.04***

0.03***

0.03**

0.02*

9(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

airport

0.04

0.01

00.02

0.05

0.06

0.02

0.09

0.1

*10

(0.03)

(0.04)

(0.04)

(0.03)

(0.04)

(0.04)

(0.04)

(0.06)

(0.06)

interm

odal

-0.02

-0.02

0.01

00.07**

0.06*

0.02

0.03

0.03

11

(0.02)

(0.03)

(0.03)

(0.03)

(0.03)

(0.03)

(0.03)

(0.04)

(0.05)

Adj.

R-squared

0.2346

0.4124

0.4315

0.3941

0.5261

0.5498

0.5699

0.555

0.5263

Standard

dev

iationsare

inparenth

eses.Sig.***<

0.001<

**<

0.05<

*<

0.1

Tab

leC

.1:

Gen

eral

War

ehou

sing

and

Sto

rage

(SIC

4225

and

NA

ICS

4931

10an

d53

1130

)(L

ogof

emplo

ym

ent)

254

1970

1974

1979

1984

1989

1994

1999

2004

2007

(Intercep

t)-19.89***

-10.29***

-18.28***

-18.44***

-12.12***

-10.15***

-3.34

-3.47

-3.4

1(2.01)

(1.33)

(1.68)

(1.66)

(1.92)

(1.83)

(2.2)

(2.23)

(2.19)

popden

sity

log

1.07***

0.86***

0.8

***

0.75***

0.87***

0.87***

0.82***

0.85***

0.85***

2(0.03)

(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

incper

caplog

2.29***

1.39***

2.19***

2.26***

1.46***

1.28***

0.51**

0.51**

0.51**

3(0.28)

(0.18)

(0.21)

(0.2)

(0.22)

(0.21)

(0.25)

(0.25)

(0.24)

taxrate

0.06***

0.01

0.03**

0.03**

0.02

0.01

0.01

0.02**

0.01

4(0.02)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

ba

0.13***

0.04**

-0.04***

-0.03**

-0.02*

-0.02**

-0.02*

-0.02**

-0.02**

5(0.02)

(0.02)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

hs

0.02**

0.03***

0.02***

0.02**

00

0.01

0.01

0.01*

6(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

nonwhite

00

0.01**

0.01**

00

00

0**

7(0)

(0)

(0)

(0)

(0)

(0)

(0)

(0)

(0)

foreign

0.11***

0.08***

0.09***

0.08***

0.05***

0.04***

0.04***

0.04***

0.05***

8(0.02)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

port

-0.01

0.01

0.03**

0.03**

0.03**

0.03**

0.03**

0.02*

0.03**

9(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

airport

0.32***

0.18***

0.17***

0.08*

0.13**

0.17***

0.15***

0.19***

0.16***

10

(0.07)

(0.04)

(0.05)

(0.05)

(0.05)

(0.04)

(0.04)

(0.05)

(0.04)

interm

odal

0.05

-0.04

-0.03

-0.01

0.08**

0.03

-0.03

0-0.01

11

(0.05)

(0.03)

(0.04)

(0.04)

(0.04)

(0.03)

(0.03)

(0.04)

(0.03)

Adj.

R-squared

0.5558

0.6122

0.5587

0.5468

0.5749

0.586

0.5408

0.5471

0.555

Standard

dev

iationsare

inparenth

eses.Sig.***<

0.001<

**<

0.05<

*<

0.1

Tab

leC

.2:

Gen

eral

frei

ght

truck

ing

(SIC

4210

and

NA

ICS

4841

00an

d48

4200

)(L

ogof

emplo

ym

ent)

255

1970

1974

1979

1984

1989

1994

1999

2004

2007

(Intercep

t)-5.04***

-7.21***

-3.74**

-4.46***

-5.56***

-9.06***

-3.35

-3.35

-5.87**

1(0.94)

(0.96)

(1.21)

(1.29)

(1.65)

(1.65)

(2.13)

(2.2)

(2.29)

popden

sity

log

0.28***

0.33***

0.37***

0.42***

0.51***

0.54***

0.58***

0.65***

0.65***

2(0.01)

(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

incper

caplog

0.52***

0.82***

0.2

0.27*

0.31

0.72***

-0.01

0.01

0.28

3(0.13)

(0.13)

(0.15)

(0.16)

(0.19)

(0.19)

(0.24)

(0.25)

(0.26)

taxrate

-0.01**

-0.02**

-0.01

-0.02**

0.01

-0.01

-0.01

-0.02*

-0.02**

4(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

ba

0.01

00.01

0.03***

0.04***

0.04***

0.06***

0.07***

0.06***

5(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

hs

0-0.01

0.03***

0.03***

0.03***

0.02***

0.04***

0.04***

0.04***

6(0)

(0)

(0)

(0)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

nonwhite

0.01***

0.01***

0.01***

0.01***

0.01***

0.01***

0.01***

0.01***

0.01***

7(0)

(0)

(0)

(0)

(0)

(0)

(0)

(0)

(0)

foreign

0.17***

0.17***

0.18***

0.18***

0.12***

0.12***

0.1

***

0.1

***

0.1

***

8(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

port

0.03***

0.05***

0.06***

0.05***

0.05***

0.06***

0.04***

0.05***

0.05***

9(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

airport

0.06*

-0.03

-0.08**

-0.06*

-0.01

-0.05

-0.07

-0.03

-0.07

10

(0.03)

(0.03)

(0.03)

(0.04)

(0.04)

(0.04)

(0.04)

(0.04)

(0.05)

interm

odal

0.04

0.04*

0.05**

0.02

0.08**

0.1

**

0.04

0.03

0.04

11

(0.02)

(0.02)

(0.03)

(0.03)

(0.03)

(0.03)

(0.03)

(0.03)

(0.04)

Adj.

R-squared

0.3478

0.4015

0.4502

0.4752

0.4886

0.5306

0.525

0.5362

0.5321

Standard

dev

iationsare

inparenth

eses.Sig.***<

0.001<

**<

0.05<

*<

0.1

Tab

leC

.3:

Fre

ight

tran

spor

tati

onar

range

men

t(S

IC47

10an

d47

23an

dN

AIC

S48

8510

)(L

ogof

emplo

ym

ent)

256

1970

1974

1979

1984

1989

1994

1999

2004

2007

(Intercep

t)0.06

-0.25

-0.38

-0.12

0.63

0.76

0.41

0.42

0.65

1(0.55)

(0.58)

(0.63)

(0.52)

(0.52)

(0.5)

(0.56)

(0.53)

(0.52)

popden

sity

log

-0.01

-0.02*

00

00.01*

0-0.01

02

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

incper

caplog

-0.02

0.03

0.04

0.01

-0.08

-0.1

*-0.05

-0.05

-0.08

3(0.08)

(0.08)

(0.08)

(0.06)

(0.06)

(0.06)

(0.06)

(0.06)

(0.06)

taxrate

0.02***

0.01**

00

00

00

04

(0)

(0)

(0)

(0)

(0)

(0)

(0)

(0)

(0)

ba

00.01

00

00

00

0**

5(0.01)

(0.01)

(0)

(0)

(0)

(0)

(0)

(0)

(0)

hs

00

00

00**

00**

0*

6(0)

(0)

(0)

(0)

(0)

(0)

(0)

(0)

(0)

nonwhite

00

00

00

00

07

(0)

(0)

(0)

(0)

(0)

(0)

(0)

(0)

(0)

foreign

0.01*

0.01**

0.01*

0.01**

00

00

08

(0)

(0)

(0)

(0)

(0)

(0)

(0)

(0)

(0)

port

00

0.01

00

00

00

9(0)

(0)

(0)

(0)

(0)

(0)

(0)

(0)

(0)

airport

-0.04**

-0.02

-0.03*

-0.02

-0.03**

-0.03**

-0.03**

-0.03**

-0.02**

10

(0.02)

(0.02)

(0.02)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

interm

odal

0.02

00.01

0.01

0.02*

0.01

0.01

0.01

011

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

Adj.

R-squared

0.0066

0.005

0.0005

0.0023

0.001

0.0025

0.0006

0.0021

0.0005

Standard

dev

iationsare

inparenth

eses.Sig.***<

0.001<

**<

0.05<

*<

0.1

Tab

leC

.4:

Min

ing,

iron

ore

(SIC

1010

and

NA

ICS

2122

10)

(Log

ofem

plo

ym

ent)

257

1970

1974

1979

1984

1989

1994

1999

2004

2007

(Intercep

t)-2.73**

-4.47***

-11.35***

-11.38***

-2.09

-1.48

-1.52

-2.03

-3.59*

1(1.21)

(1.31)

(1.55)

(1.52)

(1.67)

(1.62)

(1.89)

(1.84)

(1.89)

popden

sity

log

0.19***

0.21***

0.17***

0.17***

0.19***

0.16***

0.16***

0.14***

0.14***

2(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

incper

caplog

0.28*

0.52**

1.26***

1.26***

0.09

0.04

0.03

0.1

0.27

3(0.17)

(0.18)

(0.19)

(0.19)

(0.19)

(0.19)

(0.21)

(0.21)

(0.21)

taxrate

-0.04***

-0.04***

-0.03**

-0.02*

-0.02*

-0.01

-0.01

-0.01

-0.01

4(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

ba

-0.01

-0.02

-0.05***

-0.05***

0-0.01

00

-0.01

5(0.01)

(0.02)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

hs

0.02**

0.01**

0.01**

0.01**

0.03***

0.02***

0.02***

0.02**

0.02***

6(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

nonwhite

0.01***

0.01***

0.01***

0.01***

0.01***

0.01***

0.01***

0.01***

0.01***

7(0)

(0)

(0)

(0)

(0)

(0)

(0)

(0)

(0)

foreign

0.04***

0.04***

0.04***

0.04***

0.03***

0.03***

0.02***

0.02***

0.02***

8(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0)

(0)

(0)

port

0.04***

0.05***

0.05***

0.04***

0.04***

0.04***

0.03***

0.03***

0.03***

9(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

airport

-0.09**

-0.14**

-0.11**

-0.1

**

-0.1

**

-0.1

**

-0.1

**

-0.1

**

-0.1

**

10

(0.04)

(0.04)

(0.04)

(0.04)

(0.04)

(0.04)

(0.04)

(0.04)

(0.04)

interm

odal

0.06*

0.09**

0.08**

0.07**

0.08**

0.08**

0.1

**

0.08**

0.07**

11

(0.03)

(0.03)

(0.03)

(0.03)

(0.03)

(0.03)

(0.03)

(0.03)

(0.03)

Adj.

R-squared

0.0773

0.0907

0.0951

0.1007

0.0798

0.0581

0.0689

0.0616

0.0605

Standard

dev

iationsare

inparenth

eses.Sig.***<

0.001<

**<

0.05<

*<

0.1

Tab

leC

.5:

Pet

role

um

Refi

ner

ies

(SIC

2911

and

NA

ICS

3241

10)

(Log

ofem

plo

ym

ent)

258

1970

1974

1979

1984

1989

1994

1999

2004

2007

(Intercep

t)-3.15**

-3.37**

-2.79**

-3.13**

-2.51*

-2.3

*-0.52

-0.52

-1.46

1(1.13)

(1.12)

(1.35)

(1.27)

(1.34)

(1.24)

(1.33)

(1.25)

(1.24)

popden

sity

log

0.15***

0.15***

0.14***

0.12***

0.1

***

0.09***

0.08***

0.08***

0.06***

2(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

(0.01)

(0.01)

(0.01)

(0.01)

incper

caplog

0.37**

0.4

**

0.24

0.3

*0.2

0.19

-0.01

-0.01

0.1

3(0.16)

(0.16)

(0.17)

(0.16)

(0.16)

(0.14)

(0.15)

(0.14)

(0.14)

taxrate

-0.01

-0.01

0.01

0.01

0.01

0.01*

0.01

0.01

0.01

4(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

ba

-0.02

-0.02

-0.02*

-0.01

0-0.01

00

-0.01

5(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0)

hs

0.01

0.01

0.01**

0.01**

0.01**

0.01**

0.01**

0.01**

0.01*

6(0.01)

(0.01)

(0)

(0)

(0)

(0)

(0)

(0)

(0)

nonwhite

0*

0*

0**

0**

0**

0**

0**

0**

0**

7(0)

(0)

(0)

(0)

(0)

(0)

(0)

(0)

(0)

foreign

0.02**

0.02**

0.03***

0.02**

0.02***

0.02***

0.01***

0.01***

0.01***

8(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0)

(0)

(0)

(0)

port

00

0.01

00.01

00

00

9(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

airport

0.01

0.01

0.01

0.01

0.02

0.03

0.02

0.01

0.02

10

(0.04)

(0.04)

(0.04)

(0.04)

(0.03)

(0.03)

(0.03)

(0.03)

(0.03)

interm

odal

0.02

0.02

00

0.01

0.01

00.01

011

(0.03)

(0.03)

(0.03)

(0.03)

(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

Adj.

R-squared

0.0601

0.0669

0.0617

0.0564

0.0473

0.0371

0.0296

0.0318

0.0279

Standard

dev

iationsare

inparenth

eses.Sig.***<

0.001<

**<

0.05<

*<

0.1

Tab

leC

.6:

Gla

ssco

nta

iner

man

ufa

cturi

ng

(SIC

3221

and

NA

ICS

3272

13)

(Log

ofem

plo

ym

ent)

259

1970

1974

1979

1984

1989

1994

1999

2004

2007

(Intercep

t)-6.05***

-9.77***

-9.08***

-8.42***

-2.37

-3.69**

9.4

***

7.75***

9.99***

1(0.98)

(1.06)

(1.41)

(1.41)

(1.61)

(1.7)

(2.16)

(2.15)

(2.08)

popden

sity

log

0.29***

0.69***

0.73***

0.72***

0.73***

0.77***

0.84***

0.84***

0.8

***

2(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

incper

caplog

0.63***

1.03***

0.73***

0.65***

-0.04

0.12

-1.36***

-1.17***

-1.4

***

3(0.14)

(0.15)

(0.17)

(0.17)

(0.19)

(0.2)

(0.24)

(0.24)

(0.23)

taxrate

00

00.01

0.01

0.01

-0.01

-0.01

-0.02**

4(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

ba

0.09***

0.25***

0.13***

0.15***

0.16***

0.18***

0.18***

0.18***

0.19***

5(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

hs

-0.01**

00.03***

0.03***

0.01*

00.02***

0.02**

0.02**

6(0)

(0.01)

(0)

(0)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

nonwhite

0.01***

00**

0.01***

00

00

07

(0)

(0)

(0)

(0)

(0)

(0)

(0)

(0)

(0)

foreign

0.12***

0.1

***

0.11***

0.11***

0.08***

0.07***

0.05***

0.06***

0.06***

8(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

port

0.04***

0.01

0.01

0.01

00

0.01

0.01

0.01

9(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

airport

0.03

0.12***

0.14***

0.12**

0.16***

0.16***

0.17***

0.15***

0.14**

10

(0.03)

(0.04)

(0.04)

(0.04)

(0.04)

(0.04)

(0.04)

(0.04)

(0.04)

interm

odal

0.01

0.11***

0.08**

0.09**

0.12***

0.15***

0.07*

0.09**

0.05

11

(0.03)

(0.03)

(0.03)

(0.03)

(0.03)

(0.03)

(0.03)

(0.03)

(0.03)

Adj.

R-squared

0.3518

0.6785

0.6488

0.6503

0.6822

0.6863

0.6602

0.6699

0.6784

Standard

dev

iationsare

inparenth

eses.Sig.***<

0.001<

**<

0.05<

*<

0.1

Tab

leC

.7:

Book

,p

erio

dic

al,

and

musi

cst

ores

(SIC

5942

,59

94,

and

5733

/573

5an

dN

AIC

S45

1211

,45

1212

,an

d45

1220

)(L

ogof

emplo

ym

ent)

260

1970

1974

1979

1984

1989

1994

1999

2004

2007

(Intercep

t)-1.61**

-5.71***

-2.75**

-3.52**

-6.62***

-6.21***

-5.89***

-6.96***

-7.49***

1(0.52)

(0.85)

(1.05)

(1.1)

(1.39)

(1.4)

(1.69)

(1.72)

(1.7)

popden

sity

log

0.07***

0.22***

0.24***

0.25***

0.28***

0.27***

0.25***

0.24***

0.24***

2(0.01)

(0.01)

(0.01)

(0.01)

(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

incper

caplog

0.18**

0.68***

0.17

0.27**

0.57***

0.52**

0.45**

0.59**

0.64***

3(0.07)

(0.12)

(0.13)

(0.13)

(0.16)

(0.16)

(0.19)

(0.19)

(0.19)

taxrate

0-0.01

0-0.01

00

0-0.01

04

(0)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

ba

-0.01*

-0.02*

00

0.01

00.01**

0.01

05

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

hs

0-0.01*

0.02***

0.02***

0.02**

0.02***

0.02***

0.01**

0.01**

6(0)

(0)

(0)

(0)

(0)

(0)

(0)

(0)

(0)

nonwhite

0**

0.01***

0.01***

0.01***

0.01***

0.01***

0.01***

0.01***

0.01***

7(0)

(0)

(0)

(0)

(0)

(0)

(0)

(0)

(0)

foreign

0.05***

0.08***

0.08***

0.08***

0.05***

0.04***

0.04***

0.03***

0.03***

8(0)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0)

(0)

(0)

port

0.01

0.02***

0.02***

0.02**

0.02**

0.03***

0.03***

0.02**

0.02**

9(0)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

airport

0.03

0.04

0.03

00.05

0.03

-0.01

-0.01

010

(0.02)

(0.03)

(0.03)

(0.03)

(0.03)

(0.03)

(0.03)

(0.04)

(0.03)

interm

odal

0.01

0.01

00.01

0.02

0.01

00

0.02

11

(0.01)

(0.02)

(0.02)

(0.02)

(0.03)

(0.03)

(0.03)

(0.03)

(0.03)

Adj.

R-squared

0.0928

0.2348

0.2539

0.2452

0.2618

0.2321

0.2426

0.2204

0.2171

Standard

dev

iationsare

inparenth

eses.Sig.***<

0.001<

**<

0.05<

*<

0.1

Tab

leC

.8:

Adhes

ive

man

ufa

cturi

ng

(SIC

2891

and

NA

ICS

3255

20)

(Log

ofem

plo

ym

ent)

261

1970

1974

1979

1984

1989

1994

1999

2004

2007

(Intercep

t)0.16

0.34

1.69**

1.3

*0.6

1.39*

1.31*

1.85**

1.5

**

1(0.41)

(0.62)

(0.72)

(0.68)

(0.68)

(0.72)

(0.74)

(0.69)

(0.74)

popden

sity

log

0.01

0.02*

0.03**

0.03***

0.03***

0.03***

0.03***

0.03***

0.03***

2(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

incper

caplog

-0.02

-0.02

-0.21**

-0.16*

-0.08

-0.17**

-0.15*

-0.2

**

-0.17**

3(0.06)

(0.09)

(0.09)

(0.08)

(0.08)

(0.08)

(0.08)

(0.08)

(0.08)

taxrate

00

00

00

00

04

(0)

(0)

(0)

(0)

(0)

(0)

(0)

(0)

(0)

ba

0.01**

0.01

0.01

00

00

00

5(0.01)

(0.01)

(0)

(0)

(0)

(0)

(0)

(0)

(0)

hs

00*

00

00

00

06

(0)

(0)

(0)

(0)

(0)

(0)

(0)

(0)

(0)

nonwhite

00

00

00

00**

0*

7(0)

(0)

(0)

(0)

(0)

(0)

(0)

(0)

(0)

foreign

-0.01**

00

00.01**

0.01*

0**

0.01**

0.01***

8(0)

(0.01)

(0)

(0)

(0)

(0)

(0)

(0)

(0)

port

-0.01**

-0.01**

-0.01**

00

00

00

9(0)

(0)

(0)

(0)

(0)

(0)

(0)

(0)

(0)

airport

-0.02

00

00.02

0.01

0.01

00

10

(0.01)

(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

(0.01)

(0.01)

(0.02)

interm

odal

0.04***

0.03*

0.04**

0.03**

0.01

0.01

0.01

0.02*

0.01

11

(0.01)

(0.02)

(0.02)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

Adj.

R-squared

0.0078

0.0058

0.0073

0.0055

0.0079

0.0048

0.0044

0.007

0.0087

Standard

dev

iationsare

inparenth

eses.Sig.***<

0.001<

**<

0.05<

*<

0.1

Tab

leC

.9:

Gum

and

wood

chem

ical

sm

anufa

cturi

ng

(SIC

2860

/286

1an

dN

AIC

S32

5191

)(L

ogof

emplo

ym

ent)

262

1970

1974

1979

1984

1989

1994

1999

2004

2007

(Intercep

t)-6.19***

-8.16***

-3.98**

-2.62*

-7.93***

-7.12***

-2.62

-1.28

-1.95

1(1.05)

(1.18)

(1.49)

(1.49)

(1.77)

(1.82)

(2.19)

(2.31)

(2.35)

popden

sity

log

0.22***

0.3

***

0.3

***

0.31***

0.31***

0.31***

0.33***

0.33***

0.33***

2(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

incper

caplog

0.72***

0.95***

0.26

0.1

0.68**

0.58**

0-0.14

-0.06

3(0.14)

(0.16)

(0.18)

(0.18)

(0.21)

(0.21)

(0.24)

(0.26)

(0.26)

taxrate

-0.01

00.01

0.01*

0.03***

0.03***

0.03**

0.02

0.02**

4(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

ba

0.01

0.03**

0.03***

0.05***

0.04***

0.04***

0.06***

0.07***

0.06***

5(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

hs

-0.01*

-0.01*

0.02***

0.02**

00.01

0.02**

0.02***

0.02**

6(0)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

nonwhite

0.01***

0.01***

0.01***

0.01***

0.01***

0.01***

0.01***

0.01***

0.01***

7(0)

(0)

(0)

(0)

(0)

(0)

(0)

(0)

(0)

foreign

0.1

***

0.11***

0.1

***

0.1

***

0.07***

0.07***

0.06***

0.07***

0.06***

8(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

port

0.03***

0.02**

0.02**

0.03**

0.04***

0.04***

0.02**

0.01

0.01

9(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

airport

0.1

**

0.1

**

0.05

0.05

0.08**

0.08*

0.05

0.04

0.04

10

(0.04)

(0.04)

(0.04)

(0.04)

(0.04)

(0.04)

(0.04)

(0.05)

(0.05)

interm

odal

-0.02

00

-0.01

0.01

0.03

0.04

0.03

0.05

11

(0.03)

(0.03)

(0.03)

(0.03)

(0.03)

(0.03)

(0.03)

(0.04)

(0.04)

Adj.

R-squared

0.1903

0.2738

0.2468

0.2527

0.2901

0.272

0.2997

0.2821

0.2762

Standard

dev

iationsare

inparenth

eses.Sig.***<

0.001<

**<

0.05<

*<

0.1

Tab

leC

.10:

Phar

mac

euti

cal

pre

par

atio

ns

(SIC

2834

and

NA

ICS

3254

12)

(Log

ofem

plo

ym

ent)

263

1970

1974

1979

1984

1989

1994

1999

2004

2007

(Intercep

t)-7.48***

-12.6

***

-6.08***

-6.92***

-9.44***

-9.73***

-7.71***

-5**

-3.83*

1(0.99)

(1.16)

(1.5)

(1.48)

(1.83)

(1.82)

(2.22)

(2.14)

(2.15)

popden

sity

log

0.25***

0.45***

0.46***

0.46***

0.51***

0.5

***

0.48***

0.46***

0.47***

2(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

incper

caplog

0.88***

1.51***

0.46**

0.57**

0.81***

0.85***

0.53**

0.24

0.12

3(0.14)

(0.16)

(0.18)

(0.18)

(0.21)

(0.21)

(0.25)

(0.24)

(0.24)

taxrate

00.01

-0.01

-0.01

-0.02*

-0.01

-0.01

-0.02**

-0.02*

4(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

ba

-0.03**

-0.02

0-0.01

00.01

0.02**

0.03**

0.03**

5(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

hs

-0.01*

00.04***

0.04***

0.03***

0.02***

0.03***

0.04***

0.04***

6(0)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

nonwhite

0.01***

0.01***

0.01***

0.01***

0**

0*

0*

0**

07

(0)

(0)

(0)

(0)

(0)

(0)

(0)

(0)

(0)

foreign

0.12***

0.13***

0.12***

0.13***

0.07***

0.07***

0.05***

0.06***

0.06***

8(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

port

0.03***

0.04***

0.03***

0.03***

0.05***

0.05***

0.04***

0.04***

0.04***

9(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

airport

0.09**

0.06*

0.02

0.02

0.03

0.07

0.01

0.03

010

(0.03)

(0.04)

(0.04)

(0.04)

(0.04)

(0.04)

(0.05)

(0.04)

(0.04)

interm

odal

-0.02

-0.02

-0.04

-0.04

0-0.02

-0.03

-0.03

-0.02

11

(0.03)

(0.03)

(0.03)

(0.03)

(0.03)

(0.03)

(0.04)

(0.03)

(0.03)

Adj.

R-squared

0.2496

0.4261

0.3781

0.3839

0.3724

0.371

0.3517

0.34

0.3342

Standard

dev

iationsare

inparenth

eses.Sig.***<

0.001<

**<

0.05<

*<

0.1

Tab

leC

.11:

Ele

ctro

pla

ting,

pla

ting,

pol

ishin

g,an

odiz

ing,

and

colo

ring

(SIC

3471

and

NA

ICS

3328

13)

(Log

ofem

plo

y-

men

t)

264

1970

1974

1979

1984

1989

1994

1999

2004

2007

(Intercep

t)-9.07***

-13.55***

-6.83***

-6.5

***

-9.97***

-11.2

***

-9.89***

-10.38***

-10.8

***

1(1.18)

(1.34)

(1.71)

(1.67)

(2.02)

(2.05)

(2.74)

(2.66)

(2.69)

popden

sity

log

0.27***

0.38***

0.42***

0.41***

0.43***

0.41***

0.53***

0.54***

0.53***

2(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

(0.03)

(0.03)

(0.03)

incper

caplog

1.09***

1.67***

0.59**

0.56**

0.92***

1.09***

0.83**

0.88**

0.92**

3(0.16)

(0.19)

(0.21)

(0.2)

(0.24)

(0.24)

(0.31)

(0.3)

(0.3)

taxrate

0.01

0.03**

00

0.01

00

-0.01

-0.02

4(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

ba

-0.05***

-0.07***

-0.02

-0.01

00

0.02**

0.02**

0.02*

5(0.01)

(0.02)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

hs

-0.01

00.04***

0.04***

0.02**

0.01*

0.02**

0.03***

0.03***

6(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

nonwhite

0**

00

00

0-0.01**

-0.01**

-0.01**

7(0)

(0)

(0)

(0)

(0)

(0)

(0)

(0)

(0)

foreign

0.12***

0.12***

0.11***

0.11***

0.06***

0.06***

0.04***

0.04***

0.04***

8(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

port

0.01

0.01

00

0.01

0.02*

0.01

00.01

9(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

airport

0.12**

0.09**

0.05

0.05

0.1

**

0.05

0.04

0.04

0.04

10

(0.04)

(0.05)

(0.05)

(0.05)

(0.05)

(0.05)

(0.06)

(0.05)

(0.05)

interm

odal

-0.01

-0.01

-0.03

-0.03

0.01

-0.01

-0.01

0-0.03

11

(0.03)

(0.03)

(0.04)

(0.04)

(0.04)

(0.04)

(0.04)

(0.04)

(0.04)

Adj.

R-squared

0.2381

0.3252

0.2864

0.2872

0.2839

0.2818

0.3165

0.3409

0.3317

Standard

dev

iationsare

inparenth

eses.Sig.***<

0.001<

**<

0.05<

*<

0.1

Tab

leC

.12:

Turn

edpro

duct

and

scre

w,nut,

and

bol

tm

anufa

cturi

ng

(SIC

3450

and

NA

ICS

3327

20)

(Log

ofem

plo

ym

ent)

265

1970

1974

1979

1984

1989

1994

1999

2004

2007

(Intercep

t)-3.48***

-6.01***

-3.84***

-3.85***

-4.09***

-4.92***

-3.96**

-1.95

-2.7

**

1(0.76)

(0.93)

(1.13)

(1.04)

(1.17)

(1.14)

(1.49)

(1.29)

(1.3)

popden

sity

log

0.09***

0.18***

0.19***

0.19***

0.13***

0.12***

0.13***

0.13***

0.12***

2(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.02)

(0.01)

(0.01)

incper

caplog

0.42***

0.71***

0.34**

0.35**

0.4

**

0.49***

0.35**

0.11

0.2

3(0.1)

(0.13)

(0.14)

(0.13)

(0.14)

(0.13)

(0.17)

(0.14)

(0.14)

taxrate

0.01

0.01

0-0.01

-0.01*

0-0.01

-0.01

04

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

ba

-0.03**

00

-0.01**

0-0.01

00.01

0.01

5(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

hs

00

0.01***

0.02***

00

0.01*

0.01***

0.01**

6(0)

(0)

(0)

(0)

(0)

(0)

(0)

(0)

(0)

nonwhite

0**

0**

0**

0**

00*

00

07

(0)

(0)

(0)

(0)

(0)

(0)

(0)

(0)

(0)

foreign

0.05***

0.07***

0.06***

0.06***

0.02***

0.02***

0.01**

0.02***

0.02***

8(0.01)

(0.01)

(0.01)

(0.01)

(0)

(0)

(0)

(0)

(0)

port

00.01*

0.01*

0.01

0.01

0.01**

0.01**

0.02**

0.01**

9(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

airport

0.05**

0.07**

0.03

0.02

0.01

0.01

-0.01

00.02

10

(0.03)

(0.03)

(0.03)

(0.03)

(0.03)

(0.03)

(0.03)

(0.03)

(0.03)

interm

odal

-0.01

0-0.01

0.01

0.02

0.01

0-0.01

-0.01

11

(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

Adj.

R-squared

0.084

0.1823

0.1674

0.1748

0.091

0.0976

0.0834

0.1034

0.0999

Standard

dev

iationsare

inparenth

eses.Sig.***<

0.001<

**<

0.05<

*<

0.1

Tab

leC

.13:

Mac

hin

eto

ol(m

etal

form

ing

typ

es)

man

ufa

cturi

ng

(SIC

3542

and

NA

ICS

3335

13)

(Log

ofem

plo

ym

ent)

266

1970

1974

1979

1984

1989

1994

1999

2004

2007

(Intercep

t)-2.6

***

-3.34***

-3.77***

-2.05**

-0.8

-3.68**

-2.04

-2.59*

-1.68

1(0.72)

(0.84)

(1.06)

(1.03)

(1.27)

(1.29)

(1.49)

(1.38)

(1.36)

popden

sity

log

0.05***

0.1

***

0.12***

0.13***

0.16***

0.17***

0.15***

0.14***

0.14***

2(0.01)

(0.01)

(0.01)

(0.01)

(0.02)

(0.02)

(0.02)

(0.01)

(0.01)

incper

caplog

0.33***

0.4

***

0.36**

0.17

0.01

0.31**

0.1

0.17

0.07

3(0.1)

(0.12)

(0.13)

(0.13)

(0.15)

(0.15)

(0.17)

(0.15)

(0.15)

taxrate

00.01

0.01*

00

0.01

0.01

0.01

0.01*

4(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

ba

-0.01

0.01

0.01*

0.01

0.01**

0.02**

0.02**

0.01**

0.02***

5(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

hs

00

0.01

0.01**

0.01

0.01

0.01**

0.01*

0.01**

6(0)

(0)

(0)

(0)

(0)

(0)

(0)

(0)

(0)

nonwhite

00

0*

00

00

00*

7(0)

(0)

(0)

(0)

(0)

(0)

(0)

(0)

(0)

foreign

0.02***

0.04***

0.05***

0.04***

0.04***

0.04***

0.03***

0.03***

0.02***

8(0.01)

(0.01)

(0.01)

(0.01)

(0)

(0)

(0)

(0)

(0)

port

00.01

0.01*

0.01*

0.02**

0.02**

0.01

0.01**

0.01

9(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

airport

0.02

0.02

0.03

00.02

0.01

0.01

00.01

10

(0.02)

(0.03)

(0.03)

(0.03)

(0.03)

(0.03)

(0.03)

(0.03)

(0.03)

interm

odal

-0.01

-0.03

-0.03

-0.02

-0.02

-0.01

-0.01

0.01

0.01

11

(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

Adj.

R-squared

0.0366

0.0979

0.1265

0.1083

0.131

0.1664

0.1491

0.1556

0.1575

Standard

dev

iationsare

inparenth

eses.Sig.***<

0.001<

**<

0.05<

*<

0.1

Tab

leC

.14:

Auto

mat

icen

vir

onm

enta

lco

ntr

olm

anufa

cturi

ng

for

resi

den

tial

,co

mm

erci

al,

and

applian

ceuse

(SIC

3822

and

NA

ICS

3345

12)

(Log

ofem

plo

ym

ent)

267

1970

1974

1979

1984

1989

1994

1999

2004

2007

(Intercep

t)-2.47**

-2.57**

-2.78**

-1.78

-4.37**

-4.96**

-0.55

-0.86

-1.71

1(0.88)

(1)

(1.29)

(1.41)

(1.73)

(1.66)

(2.04)

(1.91)

(1.91)

popden

sity

log

0.09***

0.16***

0.2

***

0.21***

0.2

***

0.17***

0.22***

0.17***

0.17***

2(0.01)

(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

incper

caplog

0.27**

0.24*

0.16

0.04

0.33

0.41**

-0.15

-0.07

0.02

3(0.12)

(0.14)

(0.16)

(0.17)

(0.2)

(0.19)

(0.23)

(0.21)

(0.21)

taxrate

0.01*

0.02**

0.02**

0.02**

0.02**

0.03***

0.02**

0.01

0.01

4(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

ba

0.03**

0.07***

0.06***

0.07***

0.07***

0.06***

0.07***

0.07***

0.06***

5(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

hs

00

0.01**

0.01*

0.01

00.01**

0.01*

0.01*

6(0)

(0)

(0)

(0)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

nonwhite

00

0*

00

00*

00

7(0)

(0)

(0)

(0)

(0)

(0)

(0)

(0)

(0)

foreign

0.07***

0.1

***

0.1

***

0.12***

0.08***

0.08***

0.06***

0.06***

0.05***

8(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0)

(0)

port

00

0.02**

0.02*

0.01

0.01

0.01

00.01

9(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

airport

0.04

0.03

0.04

-0.01

-0.01

0.01

0-0.02

-0.03

10

(0.03)

(0.03)

(0.04)

(0.04)

(0.04)

(0.04)

(0.04)

(0.04)

(0.04)

interm

odal

00.04

00.01

0.05

0.05

0.03

0.03

0.01

11

(0.02)

(0.03)

(0.03)

(0.03)

(0.03)

(0.03)

(0.03)

(0.03)

(0.03)

Adj.

R-squared

0.1004

0.189

0.2372

0.2581

0.2617

0.2593

0.2794

0.2605

0.2427

Standard

dev

iationsare

inparenth

eses.Sig.***<

0.001<

**<

0.05<

*<

0.1

Tab

leC

.15:

Sem

icon

duct

ors

and

rela

ted

dev

ice

man

ufa

cturi

ng

(SIC

3674

and

NA

ICS

3344

13)

(Log

ofem

plo

ym

ent)

268

1970

1974

1979

1984

1989

1994

1999

2004

2007

(Intercep

t)-4.89***

-7.1

***

-3.96**

-2.97**

-6.73***

-5.97***

1.1

0.57

-1.08

1(0.88)

(1)

(1.27)

(1.3)

(1.56)

(1.56)

(1.93)

(2.11)

(2.1)

popden

sity

log

0.2

***

0.34***

0.36***

0.36***

0.42***

0.41***

0.44***

0.5

***

0.5

***

2(0.01)

(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

incper

caplog

0.52***

0.75***

0.19

0.06

0.45**

0.37**

-0.48**

-0.45*

-0.25

3(0.12)

(0.14)

(0.16)

(0.16)

(0.18)

(0.18)

(0.22)

(0.24)

(0.23)

taxrate

00

0.01

00.01*

0.01

0.01

0.01

0.01

4(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

ba

0.02*

0.11***

0.08***

0.11***

0.11***

0.13***

0.14***

0.15***

0.14***

5(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

hs

00

0.03***

0.03***

0.02***

0.02**

0.03***

0.03***

0.03***

6(0)

(0)

(0)

(0)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

nonwhite

0.01***

0.01***

0.01***

0.01***

0.01***

0.01***

0.01***

0.01***

0**

7(0)

(0)

(0)

(0)

(0)

(0)

(0)

(0)

(0)

foreign

0.08***

0.11***

0.11***

0.11***

0.07***

0.08***

0.06***

0.07***

0.07***

8(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0)

(0.01)

(0.01)

port

0.05***

0.04***

0.04***

0.03***

0.03**

0.02**

0.01

0.01

0.01

9(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

airport

0.04

0.02

00

0.07*

0.05

0.03

0.04

0.04

10

(0.03)

(0.03)

(0.04)

(0.04)

(0.04)

(0.04)

(0.04)

(0.04)

(0.04)

interm

odal

-0.03

0.01

-0.01

-0.02

0.02

0.06**

0.05*

0.04

0.06*

11

(0.02)

(0.03)

(0.03)

(0.03)

(0.03)

(0.03)

(0.03)

(0.03)

(0.03)

Adj.

R-squared

0.213

0.4044

0.4189

0.455

0.515

0.5406

0.544

0.538

0.5433

Standard

dev

iationsare

inparenth

eses.Sig.***<

0.001<

**<

0.05<

*<

0.1

Tab

leC

.16:

Per

iodic

alpublish

ers

(SIC

2720

and

NA

ICS

5111

20)

(Log

ofem

plo

ym

ent)

269

1970

1974

1979

1984

1989

1994

1999

2004

2007

(Intercep

t)-1.23**

-4.05***

-2.06**

-2.78**

-4.2

***

-5.1

***

-1.28

-1.06

-1.19

1(0.38)

(0.61)

(0.84)

(0.89)

(1.13)

(1.19)

(1.45)

(1.45)

(1.45)

popden

sity

log

0.04***

0.17***

0.2

***

0.23***

0.28***

0.34***

0.34***

0.37***

0.37***

2(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

incper

caplog

0.14**

0.44***

0.07

0.15

0.27**

0.33**

-0.16

-0.2

-0.18

3(0.05)

(0.09)

(0.1)

(0.11)

(0.13)

(0.14)

(0.16)

(0.16)

(0.16)

taxrate

0-0.01

00

00

-0.01

00

4(0)

(0)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

ba

00.06***

0.04***

0.07***

0.08***

0.1

***

0.1

***

0.11***

0.11***

5(0)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

hs

0-0.01**

0.01***

0.01***

0.01**

0.01**

0.02***

0.02***

0.02***

6(0)

(0)

(0)

(0)

(0)

(0)

(0)

(0)

(0)

nonwhite

0***

0.01***

0.01***

0.01***

0.01***

0.01***

0.01***

0.01***

0.01***

7(0)

(0)

(0)

(0)

(0)

(0)

(0)

(0)

(0)

foreign

0.03***

0.08***

0.08***

0.09***

0.06***

0.07***

0.06***

0.07***

0.06***

8(0)

(0.01)

(0.01)

(0.01)

(0)

(0)

(0)

(0)

(0)

port

0.01**

0.03***

0.04***

0.03***

0.03***

0.04***

0.03***

0.03***

0.02**

9(0)

(0)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

airport

0.02

0.02

0.02

0.01

0.04

0.01

-0.01

0-0.02

10

(0.01)

(0.02)

(0.02)

(0.03)

(0.03)

(0.03)

(0.03)

(0.03)

(0.03)

interm

odal

-0.01

0-0.02

-0.01

0.03

0.04*

0.04*

0.04*

0.05**

11

(0.01)

(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

Adj.

R-squared

0.0677

0.3425

0.3528

0.4313

0.4964

0.5689

0.5696

0.617

0.622

Standard

dev

iationsare

inparenth

eses.Sig.***<

0.001<

**<

0.05<

*<

0.1

Tab

leC

.17:

Mot

ion

pic

ture

and

vid

eopro

duct

ion

(SIC

7813

and

7814

/781

2an

dN

AIC

S51

2110

)(L

ogof

emplo

ym

ent)

270

Appendix D

Additional industrial sector

regression analyses

271

1970 1979 1984 1989 1994 1999 2004

-0.0

50.

000.

050.

100.

15

port

1970 1979 1984 1989 1994 1999 2004

-0.0

50.

000.

050.

100.

15

airport

1970 1979 1984 1989 1994 1999 2004

-0.0

50.

000.

050.

100.

15

intermodal

Figure D.1: Deep sea freight transportation (SIC 4410 and NAICS 483111): Regres-sion coefficients for distance (in 100km) from closest port, airport, and intermodalterminal by year against the logarithm of employment.Point represents estimate. Wide line indicates the 50 percent confidence interval, and thin line represents the95 percent confidence interval for the estimate. Dotted grey lines indicate industrial classification transitions.Red line indicates trend in estimators.

1970 1979 1984 1989 1994 1999 2004

-0.2

-0.1

0.0

0.1

0.2

port

1970 1979 1984 1989 1994 1999 2004

-0.2

-0.1

0.0

0.1

0.2

airport

1970 1979 1984 1989 1994 1999 2004

-0.2

-0.1

0.0

0.1

0.2

intermodal

Figure D.2: Marine cargo handling (SIC 4463 and 4491 and NAICS 488310 and488320): Regression coefficients for distance (in 100km) from closest port, airport,and intermodal terminal by year against the logarithm of employment. NAICS andSIC comparable to within three percent.Point represents estimate. Wide line indicates the 50 percent confidence interval, and thin line represents the95 percent confidence interval for the estimate. Dotted grey lines indicate industrial classification transitions.Red line indicates trend in estimators.

1970 1979 1984 1989 1994 1999 2004

-0.1

0.0

0.1

0.2

port

1970 1979 1984 1989 1994 1999 2004

-0.1

0.0

0.1

0.2

airport

1970 1979 1984 1989 1994 1999 2004

-0.1

0.0

0.1

0.2

intermodal

Figure D.3: Refrigerated warehousing and storage (SIC 4222 and NAICS 493120):Regression coefficients for distance (in 100km) from closest port, airport, and inter-modal terminal by year against the logarithm of employment.Point represents estimate. Wide line indicates the 50 percent confidence interval, and thin line represents the95 percent confidence interval for the estimate. Dotted grey lines indicate industrial classification transitions.Red line indicates trend in estimators.

272

1970

1974

1979

1984

1989

1994

1999

2004

2007

(Intercep

t)-1.03**

-1.61**

-1.09

-1.23*

-0.78

-0.63

-0.72

-0.41

-0.81

1(0.47)

(0.5)

(0.71)

(0.72)

(0.81)

(0.81)

(1.1)

(1.08)

(1.03)

popden

sity

log

0.06***

0.08***

0.13***

0.13***

0.12***

0.12***

0.14***

0.15***

0.13***

2(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

incper

caplog

0.11

0.17**

0.03

0.05

-0.03

-0.04

-0.09

-0.13

-0.06

3(0.07)

(0.07)

(0.09)

(0.09)

(0.09)

(0.09)

(0.12)

(0.12)

(0.11)

taxrate

-0.01**

-0.01**

00

00

00.01

04

(0)

(0)

(0)

(0)

(0)

(0)

(0)

(0)

(0)

ba

-0.01

-0.01**

-0.01

00.01**

0.01***

0.02***

0.02***

0.02***

5(0.01)

(0.01)

(0)

(0)

(0)

(0)

(0)

(0)

(0)

hs

00

0.01***

0.01**

0.01***

0.01***

0.02***

0.02***

0.02***

6(0)

(0)

(0)

(0)

(0)

(0)

(0)

(0)

(0)

nonwhite

0***

0***

0.01***

0.01***

0.01***

0***

0.01***

0.01***

0***

7(0)

(0)

(0)

(0)

(0)

(0)

(0)

(0)

(0)

foreign

0.04***

0.06***

0.07***

0.07***

0.04***

0.04***

0.04***

0.04***

0.04***

8(0)

(0)

(0)

(0)

(0)

(0)

(0)

(0)

(0)

port

00.01**

0.01**

0.01**

0.01*

0.01*

00

-0.01*

9(0)

(0)

(0)

(0)

(0)

(0)

(0)

(0)

(0)

airport

0.01

0.02

0.03

0.02

0.01

0.01

-0.01

-0.02

-0.02

10

(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

interm

odal

0.03**

0.03**

0.05**

0.06***

0.07***

0.07***

0.1

***

0.12***

0.13***

11

(0.01)

(0.01)

(0.02)

(0.02)

(0.01)

(0.01)

(0.02)

(0.02)

(0.02)

Adj.

R-squared

0.088

0.1438

0.1992

0.2105

0.1942

0.2112

0.2514

0.2844

0.2634

Standard

dev

iationsare

inparenth

eses.Sig.***<

0.001<

**<

0.05<

*<

0.1

Tab

leD

.1:

Dee

pSea

Fre

ight

Tra

nsp

orta

tion

(SIC

4410

and

NA

ICS

4831

11)

(Log

ofem

plo

ym

ent)

273

1970

1974

1979

1984

1989

1994

1999

2004

2007

(Intercep

t)-2.48**

-3.86***

-3.11**

-4.05**

0.39

0.82

2.67

2.12

1.37

1(0.91)

(0.99)

(1.26)

(1.24)

(1.42)

(1.46)

(1.7)

(1.7)

(1.71)

popden

sity

log

0.18***

0.26***

0.25***

0.25***

0.26***

0.27***

0.25***

0.26***

0.25***

2(0.01)

(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

incper

caplog

0.27**

0.43**

0.21

0.33**

-0.25

-0.31*

-0.52**

-0.47**

-0.39**

3(0.13)

(0.14)

(0.15)

(0.15)

(0.17)

(0.17)

(0.19)

(0.19)

(0.19)

taxrate

-0.04***

-0.04***

00

00

00.01

0.01

4(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

ba

-0.04***

-0.05***

-0.03***

-0.03***

0.01

0.01

0.02**

0.02**

0.02**

5(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

hs

0.01*

0.01

0.02***

0.02***

0.03***

0.04***

0.03***

0.03***

0.03***

6(0)

(0)

(0)

(0)

(0)

(0.01)

(0)

(0)

(0)

nonwhite

0.01***

0.01***

0.01***

0.01***

0.01***

0.01***

0.01***

0.01***

0.01***

7(0)

(0)

(0)

(0)

(0)

(0)

(0)

(0)

(0)

foreign

0.1

***

0.1

***

0.1

***

0.09***

0.05***

0.05***

0.03***

0.04***

0.04***

8(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0)

(0)

(0)

port

-0.01

-0.01

00

-0.01

-0.01

-0.01

-0.01

-0.01

9(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

airport

-0.04

-0.07**

-0.08**

-0.06*

-0.09**

-0.11**

-0.12***

-0.12***

-0.12***

10

(0.03)

(0.03)

(0.04)

(0.03)

(0.03)

(0.04)

(0.03)

(0.03)

(0.03)

interm

odal

0.15***

0.19***

0.19***

0.16***

0.22***

0.22***

0.19***

0.22***

0.22***

11

(0.02)

(0.03)

(0.03)

(0.03)

(0.03)

(0.03)

(0.03)

(0.03)

(0.03)

Adj.

R-squared

0.1719

0.2283

0.2194

0.2254

0.2017

0.2113

0.1899

0.2028

0.2028

Standard

dev

iationsare

inparenth

eses.Sig.***<

0.001<

**<

0.05<

*<

0.1

Tab

leD

.2:

Mar

ine

carg

ohan

dling

(SIC

4463

and

4491

and

NA

ICS

4883

10an

d48

8320

)(L

ogof

emplo

ym

ent)

274

1970

1974

1979

1984

1989

1994

1999

2004

2007

(Intercep

t)-1.49**

-3.37**

-3.77**

-2.82**

-1.55

-2.03

-0.43

-2.06

-2.75

1(0.52)

(1.06)

(1.31)

(1.2)

(1.45)

(1.52)

(1.85)

(2.11)

(2.14)

popden

sity

log

0.07***

0.27***

0.24***

0.25***

0.3

***

0.3

***

0.28***

0.33***

0.34***

2(0.01)

(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

incper

caplog

0.17**

0.3

**

0.25

0.15

-0.04

0.04

-0.17

-0.02

0.05

3(0.07)

(0.15)

(0.16)

(0.15)

(0.17)

(0.18)

(0.21)

(0.24)

(0.24)

taxrate

-0.01***

0.01

0.02**

0.01

0-0.01

-0.01

0.01

0.02*

4(0)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

ba

-0.01**

-0.01

-0.02*

-0.01

0.01**

0.01

0.02**

0.02**

0.01

5(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

hs

00.03***

0.04***

0.03***

0.03***

0.02***

0.03***

0.02***

0.02***

6(0)

(0.01)

(0)

(0)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

nonwhite

0***

0.01***

0.01***

0.01***

0.01***

0.01***

0.01***

0.01***

0.01***

7(0)

(0)

(0)

(0)

(0)

(0)

(0)

(0)

(0)

foreign

0.04***

0.09***

0.1

***

0.11***

0.08***

0.09***

0.07***

0.08***

0.09***

8(0)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0)

(0.01)

(0.01)

port

0.01*

0.04***

0.04***

0.03***

0.03***

0.03***

0.03***

0.03***

0.03***

9(0)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

airport

0.03

0.15***

0.16***

0.16***

0.15***

0.14***

0.09**

0.13**

0.11**

10

(0.02)

(0.04)

(0.04)

(0.03)

(0.04)

(0.04)

(0.04)

(0.04)

(0.04)

interm

odal

0-0.1

***

-0.11***

-0.09***

-0.04

-0.03

-0.05

-0.04

-0.02

11

(0.01)

(0.03)

(0.03)

(0.03)

(0.03)

(0.03)

(0.03)

(0.03)

(0.03)

Adj.

R-squared

0.0881

0.2099

0.1983

0.2442

0.257

0.2418

0.2449

0.2765

0.2778

Standard

dev

iationsare

inparenth

eses.Sig.***<

0.001<

**<

0.05<

*<

0.1

Tab

leD

.3:

Ref

rige

rate

dw

areh

ousi

ng

and

stor

age

(SIC

4222

and

NA

ICS

4931

20)

(Log

ofem

plo

ym

ent)

275

1970 1979 1984 1989 1994 1999 2004

-0.1

0-0

.05

0.00

0.05

0.10

0.15

port

1970 1979 1984 1989 1994 1999 2004

-0.1

0-0

.05

0.00

0.05

0.10

0.15

airport

1970 1979 1984 1989 1994 1999 2004

-0.1

0-0

.05

0.00

0.05

0.10

0.15

intermodal

Figure D.4: Ready-mix concrete manufacturing (SIC 3273 and NAICS 327320): Re-gression coefficients for distance (in 100km) from closest port, airport, and intermodalterminal by year against the logarithm of employment.Point represents estimate. Wide line indicates the 50 percent confidence interval, and thin line represents the95 percent confidence interval for the estimate. Dotted grey lines indicate industrial classification transitions.Red line indicates trend in estimators.

1970 1979 1984 1989 1994 1999 2004

-0.1

0-0

.05

0.00

0.05

port

1970 1979 1984 1989 1994 1999 2004

-0.1

0-0

.05

0.00

0.05

airport

1970 1979 1984 1989 1994 1999 2004

-0.1

0-0

.05

0.00

0.05

intermodal

Figure D.5: Cement, hydraulic (SIC 3241 and NAICS 327310): Regression coefficientsfor distance (in 100km) from closest port, airport, and intermodal terminal by yearagainst the logarithm of employment.Point represents estimate. Wide line indicates the 50 percent confidence interval, and thin line represents the95 percent confidence interval for the estimate. Dotted grey lines indicate industrial classification transitions.Red line indicates trend in estimators.

Falling under the nonmetallic mineral products category of NAICS (327), ready-

mix concrete manufacturing (SIC 3273 and NAICS 327320) is defined as the manufac-

ture of concrete delivered to a purchaser in a plastic and unhardened state, and such

establishments may mine, quarry, or purchase sand and gravel (U.S. Dept. of Com-

merce, Bureau of the Census 2007). 61 $4m 20,000+ employees Ready-mix concrete

manufacturings shows a similar pattern to natural resources. However, population

density and per capital income are highly correlated, implying an urban locational

bias.

276

1970 1979 1984 1989 1994 1999 2004

-0.0

50.

000.

050.

10

port

1970 1979 1984 1989 1994 1999 2004

-0.0

50.

000.

050.

10

airport

1970 1979 1984 1989 1994 1999 2004

-0.0

50.

000.

050.

10

intermodal

Figure D.6: Flat glass manufacturing (SIC 3211 and NAICS 327211): Regressioncoefficients for distance (in 100km) from closest port, airport, and intermodal terminalby year against the logarithm of employment.Point represents estimate. Wide line indicates the 50 percent confidence interval, and thin line represents the95 percent confidence interval for the estimate. Dotted grey lines indicate industrial classification transitions.Red line indicates trend in estimators.

277

1970

1974

1979

1984

1989

1994

1999

2004

2007

(Intercep

t)-7.61***

-10.32***

-13.94***

-11.99***

-5.83**

-9.08***

-1.4

-4.12*

-6.72**

1(1.1)

(1.41)

(1.72)

(1.67)

(1.94)

(1.98)

(2.42)

(2.35)

(2.38)

popden

sity

log

0.3

***

0.58***

0.56***

0.52***

0.61***

0.61***

0.65***

0.65***

0.63***

2(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

incper

caplog

0.86***

1.3

***

1.52***

1.27***

0.51**

0.88***

-0.03

0.33

0.63**

3(0.15)

(0.19)

(0.21)

(0.2)

(0.23)

(0.23)

(0.27)

(0.26)

(0.27)

taxrate

-0.03**

-0.03**

-0.02

-0.02**

-0.02**

-0.02*

-0.04***

-0.03**

-0.04***

4(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

ba

0.05***

0.12***

0.02

0.03**

0.04***

0.04***

0.05***

0.04***

0.03**

5(0.01)

(0.02)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

hs

00.01

0.03***

0.03***

0.02***

0.02**

0.03***

0.01**

0.01*

6(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

nonwhite

0.01***

0**

0.01***

0.01***

00

00

07

(0)

(0)

(0)

(0)

(0)

(0)

(0)

(0)

(0)

foreign

0.11***

0.05***

0.06***

0.08***

0.07***

0.05***

0.05***

0.04***

0.04***

8(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

port

0.05***

0.02*

0.03**

0.04***

0.03**

0.04***

0.03**

0.02**

0.03**

9(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

airport

-0.03

-0.02

0.05

0.01

-0.02

0.06

0.04

0.01

0.04

10

(0.04)

(0.05)

(0.05)

(0.05)

(0.05)

(0.05)

(0.05)

(0.05)

(0.05)

interm

odal

0.02

0.01

0.04

0.06

0.1

**

0.04

0.01

0.05

0.04

11

(0.03)

(0.04)

(0.04)

(0.04)

(0.04)

(0.04)

(0.04)

(0.04)

(0.04)

Adj.

R-squared

0.2978

0.4447

0.3935

0.3921

0.4354

0.4256

0.423

0.4474

0.4299

Standard

dev

iationsare

inparenth

eses.Sig.***<

0.001<

**<

0.05<

*<

0.1

Tab

leD

.4:

Rea

dy-m

ixco

ncr

ete

man

ufa

cturi

ng

(SIC

3273

and

NA

ICS

3273

20)

(Log

ofem

plo

ym

ent)

278

1970

1974

1979

1984

1989

1994

1999

2004

2007

(Intercep

t)-2.54**

-2.91**

-2.12

-0.99

-1.13

-1.18

-1.23

0.69

11

(1.02)

(1.16)

(1.35)

(1.22)

(1.41)

(1.35)

(1.66)

(1.67)

(1.7)

popden

sity

log

0.08***

0.13***

0.13***

0.1

***

0.1

***

0.1

***

0.1

***

0.11***

0.12***

2(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

incper

caplog

0.3

**

0.34**

0.16

0.05

0.06

0.08

0.07

-0.16

-0.19

3(0.14)

(0.16)

(0.16)

(0.15)

(0.16)

(0.16)

(0.19)

(0.19)

(0.19)

taxrate

0-0.01

-0.01*

-0.02**

-0.02**

-0.01**

-0.01

-0.01*

-0.01

4(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

ba

00

00

0.01

0.01

0.01*

0.01**

0.02**

5(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

hs

00.01

0.02***

0.02***

0.02**

0.01**

0.01*

0.01**

0.01**

6(0)

(0.01)

(0)

(0)

(0)

(0)

(0)

(0)

(0)

nonwhite

00**

0.01***

0**

0*

00

00

7(0)

(0)

(0)

(0)

(0)

(0)

(0)

(0)

(0)

foreign

0.02**

0.02**

0.03***

0.03***

0.01**

0.01**

0.01**

0.02***

0.02***

8(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0)

(0)

(0)

port

0.01*

0.02**

0.03**

0.02**

0.01*

0.01*

0.01

0.01

09

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

airport

-0.04

-0.06

-0.06

-0.05

-0.03

-0.02

0-0.02

010

(0.03)

(0.04)

(0.04)

(0.03)

(0.03)

(0.03)

(0.03)

(0.03)

(0.03)

interm

odal

-0.01

0-0.04

-0.04

-0.02

-0.01

-0.01

0.01

0.02

11

(0.03)

(0.03)

(0.03)

(0.03)

(0.03)

(0.02)

(0.03)

(0.03)

(0.03)

Adj.

R-squared

0.0389

0.0559

0.0654

0.0532

0.042

0.0454

0.0414

0.0409

0.0435

Standard

dev

iationsare

inparenth

eses.Sig.***<

0.001<

**<

0.05<

*<

0.1

Tab

leD

.5:

Cem

ent,

hydra

ulic

(SIC

3241

and

NA

ICS

3273

10)

(Log

ofem

plo

ym

ent)

279

1970

1974

1979

1984

1989

1994

1999

2004

2007

(Intercep

t)-0.61

-1.5

**

-2.21**

-1.81**

-1.33

-1.11

0.27

0.26

0.28

1(0.66)

(0.74)

(0.84)

(0.75)

(0.97)

(0.89)

(1.06)

(1.05)

(1.1)

popden

sity

log

0.05***

0.08***

0.08***

0.07***

0.09***

0.06***

0.05***

0.05***

0.05***

2(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

incper

caplog

0.06

0.16

0.22**

0.18*

0.09

0.09

-0.07

-0.07

-0.07

3(0.09)

(0.1)

(0.1)

(0.09)

(0.11)

(0.1)

(0.12)

(0.12)

(0.12)

taxrate

00

-0.01

0-0.01

00

00

4(0.01)

(0.01)

(0.01)

(0)

(0.01)

(0)

(0)

(0)

(0)

ba

-0.01

-0.01

-0.01*

-0.01**

00

00

05

(0.01)

(0.01)

(0.01)

(0)

(0)

(0)

(0)

(0)

(0)

hs

00

00.01*

0.01**

00.01*

00.01*

6(0)

(0)

(0)

(0)

(0)

(0)

(0)

(0)

(0)

nonwhite

00

00**

0**

0**

00

07

(0)

(0)

(0)

(0)

(0)

(0)

(0)

(0)

(0)

foreign

00.02**

0.03***

0.02***

0.02***

0.01***

0.01**

0.01**

0.01***

8(0.01)

(0.01)

(0.01)

(0)

(0)

(0)

(0)

(0)

(0)

port

0.01

0.01**

00.01

0.01

0.01*

00

09

(0)

(0.01)

(0.01)

(0)

(0.01)

(0)

(0)

(0)

(0)

airport

0.03

0.02

0.06**

0.04**

0.04*

0.05**

0.04*

0.03

0.02

10

(0.02)

(0.03)

(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

interm

odal

-0.02

-0.02

-0.02

-0.02

-0.01

-0.03

-0.03*

-0.02

-0.02

11

(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

Adj.

R-squared

0.0163

0.0397

0.0522

0.0432

0.0668

0.0294

0.0171

0.0201

0.0211

Standard

dev

iationsare

inparenth

eses.Sig.***<

0.001<

**<

0.05<

*<

0.1

Tab

leD

.6:

Fla

tgl

ass

man

ufa

cturi

ng

(SIC

3211

and

NA

ICS

3272

11)

(Log

ofem

plo

ym

ent)

280

Bibliography

Alix, Y., B. Slack, and C. Comtois (1999). Alliance or acquisition? Strategies for

growth in the container shipping industry, the case of CP ships. Journal of Trans-

port Geography 7, 203–208.

Anderson, P. L. (2002). Flash estimate: Impact of West Coast shutdown. Technical

report, Anderson Economic Group LLC, Lansing.

Armbruster, W. (2000a, January 10). Carrier committed to NY-NJ despite dispute.

Journal of Commerce, 1.

Armbruster, W. (2000b, June 2). Industry welcomes settlement of ny-nj dispute.

Journal of Commerce, WP.

Armbruster, W. (2000c, July 31). ‘Perfect storm’ at NY-NJ. Journal of Commerce,

16.

Bagli, C. V. (1998, December 11). Study questions plan for Hudson Rail Tunnel. The

New York Times, B5.

Bagli, C. V. (1999a, April 13). Pataki rejects Whitman’s $120 million deal to keep

shipping lines. The New York Times, B1.

Bagli, C. V. (1999b, February 14). Port Authority struggles to keep 2 shippers. The

New York Times, 37.

281

Bagli, C. V. (2001, April 27). Deal is signed to take over Trade Center. The New

York Times, B1.

Barnes, C. B. (1915). The Longshoremen. New York: The Russel Sage Foundation.

Bensman, D. (2008). Globalization and the labor markets of the logistics industry.

In Sloan Industries Studies Conference, Boston. Massachussetts Institute of Tech-

nology.

Bensman, D. (2009). Port trucking down the low road: A sad story of deregulation.

Technical report, Demos, New Brunswick.

Berry, B. J. L., F. E. Horton, and J. O. Abiodun (1970). Geographic perspectives on

urban systems; with integrated readings. Englewood Cliffs: Prentice-Hall.

Blomme, J. (2005). Northern range port strategy. In H. Leggate, J. McConville,

and A. Morvillo (Eds.), International Maritime Transport: Perspectives, Routledge

Advances in Maritime Studies, pp. 160–172. New York: Routledge.

Borrone, L. C. (1999, September 7). Carrier consolidation challenges ports. Journal

of Commerce, 7.

Bowen, D. J. (1998, May 18). Whitman to propose Army terminal as container port.

Journal of Commerce, 1A.

Bowen, Jr., J. T. (2008). Moving places: The geography of warehousing in the US.

Journal of Transport Geography 16, 379–387.

Branch, A. E. (1996). Elements of shipping (7th ed.). New York: Chapman & Hall.

Braudel, F. (1992). The perspective of the world. Berkeley: University of California

Press.

282

Braun, I. and B. Joerges (1994). How to recombine large technical systems: The

case of european organ transplantation. In J. Summerton (Ed.), Changing Large

Technical Systems, pp. 25–51. Boulder: Westview Press.

Brennan, T. (1998a, December 14). Carriers handicap ports on short-list. Journal of

Commerce, 12A.

Brennan, T. (1998b, June 22). Eastern ports’ fate hangs on dredging: Deep channels

needed for big vessels. Journal of Commerce, 1C.

Brennan, T. (1998c, December 11). It’s official: New York, Baltimore, Halifax on

short-list for Atlantic hub. Journal of Commerce, 1A.

Brennan, T. (1998d, July 1). Port: If we build it, will you stay? New York proposes

huge terminal in effort to keep Sea-Land, Maersk. Journal of Commerce, 1A.

Brennan, T. (1998e, May 20). Six East Coast ports to share info: Deal made to

disclose rates, service terms. Journal of Commerce, 1A.

Brennan, T. (1999a, April 15). Gov. Whitman’s offer accepted as final bid; Gov.

Pataki holds out. Journal of Commerce, 1A.

Brennan, T. (1999b, January 21). Infighting puts New York off course. Journal of

Commerce, 1A.

Brennan, T. (1999c, April 9). Labor urges governors to sign port proposal. Journal

of Commerce, 11A.

Brennan, T. (1999d, August 3). Merger was logical step, executives say. Journal of

Commerce, 1.

283

Brennan, T. (1999e, 8). New York hinders effort to win mega-hub. Journal of

Commerce, 1A.

Brennan, T. (1999f, March 11). New York to unveil final hub proposal. Journal of

Commerce, 1A.

Brennan, T. (1999g, April 26). Pataki makes new demands on New Jersey. Journal

of Commerce, 1A.

Brennan, T. (1999h, March 30). Politics once more stalls port hub plan. Journal of

Commerce, 1A.

Brennan, T. (1999i, May 11). Rail costs determined megaport selection. Journal of

Commerce, 1A.

Brennan, T. (1999j, May 6). Sea-Land, Maersk eye big East Coast terminal: Pressure

rises on New York as 5 ports asked for proposals. Journal of Commerce, 1A.

Brennan, T. (1999k, June 21). Upheaval avoided as 2 carriers stay. Journal of

Commerce, 13.

Brennan, T. (1999l, June 29). Waterfront employers will start hiring for first time in

30 years. Journal of Commerce, 21.

Brenner, N. and S. Elden (2009). Henri Lefebvre on state, space, territory. Interna-

tional Political Sociology 3, 353–377.

Brooks, M. (2000). Sea Change in Liner Shipping: Regulation and Managerial

Decision-Making in a Global Industry. Pergamon Press.

Brown, P. H. (2008). America’s Waterfront Revival: Port Authorities and Urban

Redevelopment. Philadelphia: University of Pennsylvania Press.

284

Brusco, S. (1982). The Emilian model: Productive decentralization and social inte-

gration. Cambridge Journal of Economics 6, 167–184.

Bucholz, A. (1994). Armies, railroads, and information: The birth of industrial war-

fare. In J. Summerton (Ed.), Changing Large Technical Systems, pp. 53–70. Boul-

der: Westview Press.

Bureau of Applied Economics (1920). Standards of living: A compilation of budgetary

studies. Washington, D.C.: Bureau of Applied Economics.

Callon, M. (1986). Some elements of a sociology of translation: Domestication of the

scallops and the fishermen of St. Brieuc Bay. In J. Law (Ed.), Power, action, and

belief: A new sociology of knowledge?, Sociological review monograph, pp. viii, 280.

Boston: Routledge & Kegan Paul.

Cargo Handling (1956a). Trend towards mechanisation at the smaller overseas ports.

Cargo Handling 3 (10 and 11), 254–255.

Cargo Handling (1956b). Variety of attitudes among U.S. ports. Cargo Handling 3 (10

and 11), 255–256.

Caro, R. A. (1975). The Power Broker: Robert Moses and the Fall of New York. New

York: Vintage Books.

Ceres (2010, April 7). Ceres. http://www.ceresglobal.com.

Chlomoudis, C. I., A. V. Karalis, and A. A. Pallis (2003). Port reorganisations and

the worlds of production theory. European Journal of Transport and Infrastructure

Research 3 (1), 77–94.

Clarke, R. C. (1997). An analysis of the international ocean shipping conference

system. Transportation Journal 36 (4), 17–24.

http://www.ceresglobal.com

285

CN (2010, April 7). CN: North America’s Railroad. http://www.cn.ca.

Cohen, R. B. (1981). The new international division of labor, multinational corpora-

tions and urban hierarchy. In M. J. Dear and A. J. Scott (Eds.), Urbanization and

urban planning in capitalist society, pp. 287–315. London ; New York: Methuen.

Columbia Law Review (1966). “Piggyback” service under the Interstate Commerce

Act. Columbia Law Review 66 (2), 318–337.

Containerisation International (2006). Containerisation International Yearbook.

Technical report, Containerisation International.

Cooley, C. H. (1894). The Theory of Transportation. Baltimore: American Economic

Association.

Cooper, M. (2003, October 16). City offers longer airport leases for $700 million and

more rent. The New York Times, A1.

Council of State Governments (1978). The Book of the States. Lexington, KY: Council

of State Governments.









http://www.cn.ca

286



Crake, R. T. (1962). Long distance economics. Cargo Handling 10 (7), 20–24.

Cullinane, K. and M. Khanna (2000). Economies of scale in large containerships:

optimal size and geographical implications. Journal of Transport Geography 8,

181–195.

Currid, E. (2006). New York as a global creative hub: A competitive analysis of four

theories on world cities. Economic Development Quarterly 20 (4), 330–350.

David, G. (1957, April). The economic case for trailerships. Cargo Handling 4 (12),

290–292.

Davis, H. C. (1983). Regional port impact studies: A critique and suggested method-

ology. Transportation Journal 23 (1), 61–71.

de Langen, P. W. (2007). The economic performance of seaport regions. In J. Wang,

D. Olivier, T. Notteboom, and B. Slack (Eds.), Ports, Cities, and Global Supply

Chains, pp. 187–202. Burlington: Ashgate Publishing Limited.

Dicken, P. (2003). Global Shift: Reshaping the Global Economic Map in the 21st

Century (Fourth ed.). The Guilford Press.

Dicken, P. and A. Malmberg (2001). Firms in territories: A relational perspective.

Economic Geography 77 (4), 315–328.

Doig, J. W. (2001). Empire on the Hudson: Entrepreneurial vision and political power

at the Port of New York Authority. New York: Columbia University Press.

287

Egyedi, T. M. (1996, August 23–24). The standardized container: Gateway tech-

nologies in cargo transportation. In Helsinki Workshop on Standardisation and

Transportation, Helsinki.

Erem, S. and E. P. Durrenberger (2008). On the Global Waterfront: The Fight to

Free the Charleston 5. New York: Monthy Review Press.

Fainstein, S. S. (2001). The City Builders: Property Development in New York

and London, 1980-2000 (Studies in Government and Public Policy) (2nd ed.).

Lawrence, KS: Univeristy of Kansas Press.

Fainstein, S. S. (2005). Ground Zero’s landlord: The role of the Port Authority of

New York and New Jersey in the reconstruction of the World Trade Center site. In

J. Mollenkopf (Ed.), Contentious City: The Politics of Recovery in New York City,

pp. 73–94. New York: Russell Sage Foundation.

Fainstein, S. S. and N. I. Fainstein (1989). Technology, the new international division

of labor, and location: Continuities and disjunctures. In R. A. Beauregard (Ed.),

Economic restructuring and political response, Volume 34 of Urban Affairs Annual

Reviews, pp. 17–39. Newbury Park, Calif.: Sage Publications.

Fairley, L. (1979). Facing Mechanization: The West Coast Longshore Plan. Los

Angeles: Institute of Industrial Relations, University of California.

Federal Highway Administration (FHWA) (2007). Freight analysis framework. Tech-

nical report, Federal Highway Administration (FHWA), Washington D.C. Accessed

25 Aug 2009.

Federal Maritime Commission (2001). The impact of the Ocean Shipping Reform Act

of 1998. Technical report, Federal Maritime Commission.

288

Finlay, W. (1988). Work on the Waterfront. Philadelphia: Temple University Press.

Fleming, D. K. and A. J. Baird (1999). Some reflections on port competition in

the United States and western Europe. Maritime Policy & Management 26 (4),

383–394.

Florida, R. (2002). The Rise of the Creative Class: And How It’s Transforming Work,

Leisure, Community and Everyday Life. New York, NY: Basic Books.

Flyvbjerg, B., N. Bruzelius, and W. Rothengatter (2003). Megaprojects and Risk: An

Anatomy of Ambition. New York: Cambridge University Press.

Frankel, E. G. (1987). Port Planning and Development. New York: John Wiley &

Sons.

Friedmann, J. and G. Wolff (1982). World city formation: An agenda for research

and action. International Journal of Urban and Regional Research 6 (3), 309–343.

Frobel, F., J. Heinrichs, and O. Kreye (1980). The new international division of

labour: Structural unemployment in industrialised countries and industrialisation

in developing countries. New York: Cambridge University Press.

Fujita, M., P. R. Krugman, and A. Venables (1999). The Spatial Economy: Cities,

Regions and International Trade. Cambridge, Mass.: MIT Press.

Garrison, W. L. and D. M. Levinson (2006). The Transportation Experience. New

York: Oxford University Press.

Gereffi, G., J. Humphrey, and T. Sturgeon (2005). The governance of global value

chains. Review of International Political Economy 12 (1), 78–104.

Gersten, A. (1998, January 9). Expanding cargo. Journal of Commerce, 1B.

289

Getzfred, M. (1998, December 18). Hanjin already feeling squeeze at Long Beach.

Journal of Commerce, 2B.

Giuliani, R. W. (1999, January 14). An agenda to prepare for the next century.

http://nyc.gov/html/rwg/html/99a/stcitytext.html.

Glaeser, E. L. and J. E. Kohlhase (2003). Cities, regions and the decline of transport

costs. Technical Report 9886, National Bureau of Economic Research, Cambridge.

Gokalp, I. (1992). On the analysis of large technical systems. Science, Technology, &

Human Values 17 (1), 57–78.

Gordon, D. M. (1978). Capitalist development and the history of american cities. In

W. K. Tabb and L. Sawers (Eds.), Marxism and the metropolis: New perspectives

in urban political economy, pp. 25–63. New York: Oxford University Press.

Graham, S. and P. Healey (1999). Relational concepts of space and place: Issues for

planning theory and practice. European Planning Studies 7 (5), 623–646.

Grubler, A. (1990). The rise and fall of infrastructures: Dynamics of evolution and

technological change in transport. Heidelberg: Physica-Verlag.

Haezendonck, E. (2001). Essays on strategy analysis for seaports. Leuven: Garant

Publishers.

Halifax Port Authority (2010, April 5). Port of Halifax. http://www.portofhalifax.

ca.

Hall, P. V. (2003). Regional institutional convergence? Reflections from the Baltimore

waterfront. Economic Geography 79 (4), 347–363.

http://nyc.gov/html/rwg/html/99a/stcitytext.html

http://www.portofhalifax.ca

http://www.portofhalifax.ca

290

Hall, P. V. (2004). “We’d have to sink the ships”: Impact studies and the 2002 West

Coast port lockout. Economic Development Quarterly 18 (4), 354–367.

Halterm Limited (2010, April 7). Halterm Limited. http://www.halterm.com.

Hammond, R. (1957). Railway containers. Cargo Handling 4 (9), 207–8.

Hartman, R. T. and R. J. L. Fish (1957). A modern system for handling miscellaneous

cargoes at ports and airports. Cargo Handling 4 (11), 250–256.

Harvey, D. (1982). The Limits to Capital. Chicago: The University of Chicago Press.

Harvey, D. (1989). From managerialism to entrepreneurialism: The transformation

in urban governance in late capitalism. Geografiska Annaler B 71 (1), 3–17.

Harvey, D. (2006). Spaces of global capitalism: Towards a theory of uneven geograph-

ical development. New York: Verso.

Hayuth, Y. (1981). Containerization and the load center concept. Economic Geogra-

phy 57 (2), 160–176.

Heaver, T., H. Meersman, F. Moglia, and E. V. D. Voorde (2000). Do mergers and

alliances influence european shipping and port competition? Maritime Policy &

Management 27 (4), 363–373.

Heaver, T. D., H. Meersman, and E. V. de Voorde (2005). Co-operation and com-

petition in international container transport: Strategies for ports. In H. Leggate,

J. McConville, and A. Morvillo (Eds.), International Maritime Transport: Perspec-

tives, Routledge Advances in Maritime Studies, pp. 145–159. New York: Routledge.

http://www.halterm.com

291

Heins, M. (2009, October 24). The shipping container in the era of postmodern

globalization. In Infrastructure’s Domain: Architectural Manifestations of Techno-

Bureaucratic Systems, Princeton University. CONFANN.

Henderson, J. V. (1983). Industrial bases and city sizes. The American Economic

Review 73 (2), 164–168.

Herod, A. (2001). Labor Geographies: Work and the Landscapes of Capitalism. New

York: The Guilford Press.

Hevesi, D. (1994, November 25). Plan for toxic dump off Coney I. is attacked. The

New York Times, 1B.

Hirschman, A. O. (1958). The strategy of economic development. New Haven,: Yale

University Press.

Hobsbawm, E. (1964). Labouring Men: Studies in the History of Labour. New York:

Basic Books, Inc.

Hughes, T. P. (1983). Networks of Power: Electrification in Western Society, 1880-

1930. Baltimore: The Johns Hopkins University Press.

International Labour Office (1973). Social repercussions of new methods of cargo

handling (docks). Technical Report V(2), International Labour Office.

International Trade Association (2010, April 22). Table 41: U.S. total exports, 1998-

03 by six-digit NAICS product groups. Technical report, U.S. Dept. of Commerce.

Ircha, M. C. (2001). Port strategic planning: Canadian port reform. Maritime Policy

& Management 28 (2), 125–140.

292

Jensen, V. H. (1974). Strife on the Waterfront: The Port of New York since 1945.

Ithace: Cornell University Press.

Jessop, B., N. Brenner, and M. Jones (2008). Theorizing sociospatial relations. En-

vironment & Planning D 26, 389–401.

Joerges, B. (1988). Large technical systems: Concepts and issues. In R. Mayntz

and T. P. Hughes (Eds.), The Development of Large Technical Systems, pp. 9–36.

Boulder: Westview Press.

Johnson, M. (1950). Crime on the Labor Front. New York: McGraw-Hill Book

Company, Inc.

Journal of Commerce (1998a, July 16). Port competition. Journal of Commerce, 6A.

Journal of Commerce (1998b, December 14). A risky course. Journal of Commerce,

6A.

Journal of Commerce (1999a, January 19). A curious course. Journal of Commerce,

6A.

Journal of Commerce (1999b, April 16). Gov. Pataki’s pique. Journal of Commerce,

4A.

Journal of Commerce (1999c, April 14). Governors continue snit over port bid. Journal

of Commerce, 6A.

Journal of Commerce (2000a, June 1). NY/NJ terminal gets a tenant. Journal of

Commerce, 18.

Journal of Commerce (2000b, June 1). Text of the New York-New Jersey Port Au-

thority agreement. Journal of Commerce, WP.

293

Kahn, L. M. (1980). Unions and internal labor markets: The case of the San Francisco

longshoremen. Labor History 21 (3), 369–391.

Kannapell, A. (1997, December 17). NEW JERSEY & CO.; If only one East Coast

port can be no. 1, which one will it be? The New York Times , Section 14NJ, page

8.

Keith, R. C. (2005). Baltimore Harbor: A Pictorial History (3rd ed.). Baltimore:

The Johns Hopkins University Press.

Kendall, L. C. and J. J. Buckley (2001). The business of shipping (7th ed.). Centre-

ville, Md.: Cornell Maritime Press.

Klosterman, R. E. (2003). Arguments for and against planning. In S. Campbell and

S. S. Fainstein (Eds.), Readings in Planning Theory (2nd ed.)., pp. 86–101. Oxford:

Blackwell.

Knox, P. L. (1995). World cities in a world-system. In P. L. Knox and P. J. Taylor

(Eds.), World cities in a world-system, pp. 3–20. New York: Cambridge University

Press.

Knuth, D. E. (1984). Literate programming. The Computer Journal 27 (2), 97–111.

Accessed on 27 May 2009.

Kossoris, M. D. (1961). Working rules in west coast longshoring. Monthly Labor

Review 84 (1), 1–10. Max D. Kossoris, director of the Western Regional Office of

the Bureau of Labor Statistics.

Krugman, P. R. (1991). Geography and trade. Gaston Eyskens lecture series. Cam-

bridge: MIT Press.

294

Lahr, M. L. and A. Strauss-Weider Inc. (2004). Economic impacts of the New

York/New Jersey port industry. Technical report, Rutgers University, Center for

Urban Policy Research, New Brunswick.

Latour, B. (1996). Aramis, or the Love of Technology. Cambridge: Harvard University

Press.

Latour, B. (1997). On actor-network theory: A few clarifications.

Latour, B. (1999). Pandora’s hope: Essays on the reality of science studies. Cam-

bridge: Harvard University Press.

LeRoy, G. (2007). Nine concrete ways to curtail the economic war among the states.

In A. Markusen (Ed.), Reining in the Competition for Capital, pp. 183–197. Kala-

mazoo: W.E. Upjohn Institute for Employment Research.

Levinson, H. M. (1967). Unionism, concentration, and wage changes: Toward a unified

theory. Industrial and Labor Relations Review 20 (2), 198–205.

Levinson, M. (2006). The Box: How the Shipping Container Made the World Smaller

and the World Economy Bigger. Princeton: Princeton University Press.

Lewis, I. and D. B. Vellenga (2000). The Ocean Shipping Reform Act of 1998. Trans-

portation Journal 39 (4), 27–34.

Losch, A. (1954). The economics of location. New Haven: Yale University Press.

Luberoff, D. and J. Walder (2000). U.S. ports and the funding of intermodal facilities:

An overview of key issues. Technical report, Massachussetts Institute of Technology,

New England Transportation Center, Cambridge.

295

Luhmann, N. (1995). Social systems. Writing science. Stanford: Stanford University

Press.

Magden, R. and A. D. Martinson (1982). The Working Waterfront: The Story of

Tacoma’s Ship and Men. Tacoma: I.L.W.U. Local 23 and Port of Tacoma.

Maher Terminals, LLC (2008, June 11). Maher Terminals, LLC v. The Port Authority

of New York and New Jersey.

Malmberg, A. and P. Maskell (2006). Localized learning revisited. Growth and

Change 27 (1), 1–18.

Maloney, Jr., J. C. (1996). History and significance of military packaging. Technical

Report 1-96, Defense Systems Management College, Fort Belvoir, VA.

Marcuse, P. (2005). ‘The city’ as perverse metaphor. City 9 (2), 247–254.

Maritime Cargo Transportation Conference (1957). Cargo Ship Loading: An Analysis

of General Cargo Loading in Selected U.S. Ports. Number 474. Washington, D.C.:

National Academy of Sciences-National Research Council.

Markusen, A. (2006). Urban development and the politics of a creative class: Evidence

from a study of artists. Environment & Planning A 38, 1921–1940.

Markusen, A. (Ed.) (2007). Reining in the Competition for Capital. Kalamazoo:

W.E. Upjohn Institute for Employment Research.

Markusen, A. and K. Nesse (2007). Institutional and political determinants of incen-

tive competition. In A. Markusen (Ed.), Reining in the Competition for Capital,

pp. 1–42. Kalamazoo: W.E. Upjohn Institute for Employment Research.

296

Markusen, A. and G. Schrock (2006). The artistic dividend: Urban artistic speciali-

sation and economic development implications. Urban Studies 43 (10), 1661–1686.

Markussen, C. M. (1962). Transportation of general cargo. Cargo Handling 10 (5),

8–10.

Marshall, A. (1890). Principles of economics. New York: Macmillan and Co.

Marx, K. (1977). Capital: A critique of political economy, Vol. 1. New York: Vintage

Books.

Marx, K. and D. Fernbach (1981). Capital: A critique of political economy, Vol. 3.

New York: Penguin Books in association with New Left Review.

McCann, P. and D. Shefer (2004). Location, agglomeration and infrastructure. Papers

in Regional Science 83, 177–196.

Meersman, H., E. V. de Voorde, and T. Vanelslander (2005). Ports as hubs in the

logistics chain. In H. Leggate, J. McConville, and A. Morvillo (Eds.), International

Maritime Transport: Perspectives, Routledge Advances in Maritime Studies, pp.

137–144. New York: Routledge.

Miller, M. (2009, March 30). Personal communication.

Minnesota Population Center (2004). National historical geographic information sys-

tem: Pre-release version 0.1. Technical report, University of Minnesota, Minneapo-

lis, MN.

Mongelluzzo, B. (1999a, October 29). Maersk chooses LA as site for new 484-acre

box terminal. Journal of Commerce, 1.

297

Mongelluzzo, B. (1999b, March 19). Moody’s sees tough financial conditions for ship

lines and ports. Journal of Commerce, 1B.

Mumford, L. (1961). The City in History: Its Origins, Its Transformations and Its

Prospects. New York, NY: Harcourt Brace & World.

Murray, T. (2004). Canadian National Railway. St. Paul: MBI Publishing Company.

National Industrial Conference Board (1921). Family Budgets of American Wage-

Earners: A Critical Analysis. Number 41. New York: Century Co.

Nelson, B. (1988). Workers on the Waterfront: Seamen, Longshoremen, and Union-

ism in the 1930s. Chicago: University of Illinois Press.

Nelson, R. R. and S. G. Winter (1982). An Evolutionary Theory of Economic Change.

Cambridge: Harvard University Press.

Newman, A. (1997, December 25). Big year allows port to cut fees and challenge

rivals. The New York Times, B4.

North, D. C. (1981). Structure and change in economic history (1st ed.). New York:

Norton.

North Atlantic Ports Association (1949–1999). Proceedings of meetings. Typed min-

utes.

Notteboom, T. and J.-P. Rodrigue (2005). Port regionalization: Towards a new phase

in port development. Maritime Policy & Management 32 (3), 297–313.

Notteboom, T. and J.-P. Rodrigue (2007). Re-assessing port-hinterland relations in

the context of global commodity chains. In J. Wang, D. Olivier, T. Notteboom, and

298

B. Slack (Eds.), Ports, Cities, and Global Supply Chains, pp. 51–66. Burlington:

Ashgate Publishing Limited.

Notteboom, T. E. (2009). Path dependency and contingency in the development

of multi-port gateway regions and multi-port hub regions. In Ports in Proximity:

Competition and Coordination among Adjacent Seaports, pp. 55–72. Burlington:

Ashgate Publishing Company.

Offe, C. (1985). Disorganized Capitalism, Chapter Two Logics of Collective Action,

pp. 170–220. Cambridge: MIT Press.

Olson, M. (1971). The Logic of Collective Action; Public Goods and the Theory of

Groups. Cambridge: Harvard University Press.

Palmer, R. W. and F. P. DeGiulio (1989). Terminal operations and multimodal

carriage: History and prognosis. Tulane Law Review 64 (2–3), 281–360.

Panama Canal Authority (2009). http://www.pancanal.com/eng/expansion/

index.html.

Peck, J. (2005). Struggling with the creative class. International Journal of Urban

and Regional Research 29 (4), 740–770.

Picard, L. A. (1967). Report of the inquiry commission on the st. lawrence ports.

Technical report, l’Ecole des Hautes Etudes Commerciales. Obtained at SUNY

Maritime Library.

Popper, K. (1959). The Logic of Scientific Discovery. New York: Routledge.

Port Authority of New York & New Jersey (2001–2006). Comprehensive annual

financial report. Technical report, The Port Authority of New York & New Jersey.

http://www.pancanal.com/eng/expansion/index.html

http://www.pancanal.com/eng/expansion/index.html

299

Port Authority of New York and New Jersey (2010, April 5). Port Authority of New

York and New Jersey. http://www.panynj.gov.

Port of New York Authority, Development and Operations Department (1938). The

Port of New York primer: A story of cooperation between two states. Technical

report, The Port of New York Authority.

Porter, M. E. (1998). Clusters and the new economics of competition. Harvard

Business Review 76 (6), 77–90.

Powell, W. W. (1990). Neither market nor hierarchy: Network forms of organization.

In B. M. Staw and L. L. Cummings (Eds.), Research in organizational behavior,

Volume 12, pp. 295–336. Greenwich: JAI Press.

Prigogine, I. (1997). The End of Certainty: Time, Chaos, and the New Laws of

Nature. New York: Free Press.

Prince, T. (1999, May 20). The great port chase. Journal of Commerce, 5A.

Rath, E. (1975). Containers: Their definitions and implications. The Transportation

Law Journal 7, 53–81.

Revkin, A. C. (1997, August 2). $1.2 million for a plan on the future of a harbor.

The New York Times, 23.

Rietveld, P. and R. Vickerman (2004). Transport in regional science: The “death of

distance” is premature. Papers in Regional Science 83, 229–248.

Rodrigue, J.-P. (2004). Appropriate models of port governance: Lessons from the Port

Authority of New York and New Jersey. In D. Pinder and B. Slack (Eds.), Shipping

and Ports in the Twenty-first Century: Globalization, Technological Change and

http://www.panynj.gov

300

the Environment, Ocean Management and Policy Series, pp. 63–81. New York:

Routledge.

Rodrigue, J.-P. (2006, January 10). Transport geography on the web. http://

people.hofstra.edu/geotrans/.

Rodrigue, J.-P., C. Comtois, and B. Slack (2006). The Geography of Transport Sys-

tems. New York: Routledge.

Rodrigue, J.-P. and T. Notteboom (2009). The geography of containerization: Half

a century of revolution, adaptation and diffusion. GeoJournal 74, 1–5.

Rose, M. H., B. E. Seely, and P. F. Barrett (2006). The Best Transportation System

in the World. Columbus: Ohio State University.

Roth, W. M. (1957). The problem of american shipping—new questions and old

answers. Operations Research 5 (1), 104–110.

Sassen, S. (2002). Globalization and its discontents. In G. Bridge and S. Watson

(Eds.), The Blackwell City Reader, pp. 161–170. Malden: Wiley-Blackwell.

Saxenian, A. (1996). Regional Advantage: Culture and competition in Silicon Valley

and Route 128 (1st Harvard University Press pbk. ed.). Cambridge: Harvard

University Press.

Sayer, R. A. (1992). Method in Social Science: A realist approach (second ed.). New

York: Routledge.

Schneider, B. V. H. (1959). The maritime industry. Monthly Labor Review 82 (5),

552–557.

http://people.hofstra.edu/geotrans/

http://people.hofstra.edu/geotrans/

301

Schneider, V. (1994). Multinationals in transition: Global technical integration and

the role of corporate telecommunication networks. In J. Summerton (Ed.), Chang-

ing Large Technical Systems, pp. 71–92. Boulder: Westview Press.

Schumpeter, J. A. (1983). The Theory of Economic Development: An Inquiry into

Profits, Capital, Credit, Interest, and the Business Cycle. New Brunswick: Trans-

action Books.

Scott, A. J. (2001). Global City-Regions: Trends, Theory, Policy. New York: Oxford

University Press.

Sevunts, L. (2005, June 12). Northwest Passage redux. http://www.

washingtontimes.com/news/2005/jun/12/20050612-123835-3711r/.

Shashikumar, N. and G. Schatz (2000). The impact of U.S. regulatory changes on

intermodal movements. Transportation Journal 40 (1), 5–14.

Sherman, R. B. (2002). Seaport governance in the united states and canada. Technical

report, American Association of Port Authorities, Alexandria.

Sidey, J. M. N. (1962). Containers: Their commercial operation. Cargo Han-

dling 10 (6), 16–19.

Slack, B. (2004). Corporate realignment and the imperatives of container shipping.

In D. Pinder and B. Slack (Eds.), Shipping and Ports in the Twenty-first Century:

Globalization, Technological Change and the Environment, Ocean Management and

Policy Series, pp. 25–39. New York: Routledge.

Slack, B., C. Comtois, and R. McCalla (2002). Strategic alliances in the container

shipping industry: A global perspective. Maritime Policy & Management 29 (1),

65–76.

http://www.washingtontimes.com/news/2005/jun/12/20050612-123835-3711r/

http://www.washingtontimes.com/news/2005/jun/12/20050612-123835-3711r/

302

Smothers, R. (2000, April 28). Hopes are raised, then dashed, over stalemate at Port

Authority. The New York Times , B1.

Storey, G. and W. Armbruster (2000, June 12). After the NY-NJ deal. Journal of

Commerce, 21.

Storper, M. (1997). The Regional World: Territorial Development in a Global Econ-

omy. New York: Guilford.

Sturgeon, T. (2002). Modular production networks: A new american model of indus-

trial organization. Industrial and Corporate Change 11 (3), 451–496.

Sturgeon, T. J. (2003). What really goes on in Silicon Valley? Spatial clustering

and dispersal in modular production networks. Journal of Economic Geography 3,

199–225.

Summerton, J. (1994a). Changing Large Technical Systems, Chapter The Systems

Approach to Technological Change, pp. 1–21. Boulder: Westview Press.

Summerton, J. (Ed.) (1994b). Changing Large Technical Systems. Boulder: Westview

Press.

Taylor, P. J. (2004). World City Network: A Global Urban Analysis. New York:

Routledge.

Tirschwell, P. (1998, February 17). Now Maersk considers alternatives to New York.

Journal of Commerce, 1A.

Tirschwell, P. (1999, January 29). Packing more buzz than usual: Sea-Land sale

rumor returns. Journal of Commerce, 1A.

Tooth, E. S. (1956). Palletisation and shipping. Cargo Handling 3 (9), 224–228.

303

Tower, C. (1998a, September 28). Halifax bullish on super-terminal. Journal of

Commerce, 1B.

Tower, C. (1998b, September 28). Halifax secures funds for terminal. Journal of

Commerce, 3B.

United Nations Economic Commission for Europe (2007). UN/LOCODE (code for

trade and transport locations). Technical report, United Nations Economic Com-

mission for Europe, Geneva.

United States Army Corps of Engineers New York District (2009, September). Bay-

onne Bridge air draft analysis. Technical report, The Port Commerce Department,

the Port Authority of New York and New Jersey.

United States Court of Appeals, E. C. (1984). Tri-State Motor Transit Co. v. Inter-

state Commerce Commission. 739 F.2d 1373.

U.S. Bureau of Labor Statistics (1932, February). Cargo handling and longshore labor

conditions. Number Bulletin No. 550. U.S. Bureau of Labor Statistics.

U.S. Dept. of Commerce, Bureau of the Census (1986, April). County Business Pat-

terns, 1979 [United States]: U.S. summary, state, and county data. Technical Re-

port 8442, Inter-University Consortium for Political and Social Research (ICPSR),

Ann Arbor.

U.S. Dept. of Commerce, Bureau of the Census (1987, May). County Business Pat-



Ann Arbor.

304




Ann Arbor.

U.S. Dept. of Commerce, Bureau of the Census (1998, July). County Business Pat-



Ann Arbor.




Ann Arbor.

U.S. Dept. of Commerce, Bureau of the Census (2004). County Business Patterns,

2004 [United States]: U.S. summary, state, and county data. Technical report,

Inter-University Consortium for Political and Social Research (ICPSR), Washing-

ton D.C.

U.S. Dept. of Commerce, Bureau of the Census (2007). County Business Patterns,

2007 [United States]: U.S. summary, state, and county data. Technical report, U.S.

Dept. of Commerce, Bureau of the Census, Washington D.C.

U.S. Dept. of Commerce, Bureau of the Census (2009, February). County Business

Patterns, 1970-1976 [United States]: U.S. summary, state, and county data. Tech-

nical Report 27422, Inter-University Consortium for Political and Social Research

(ICPSR), Ann Arbor.

305

Usselman, S. W. (1994). Changing embedded systems: The economics and politics

of innovation in american railroad signaling, 1876–1914. In J. Summerton (Ed.),

Changing Large Technical Systems, pp. 71–92. Boulder: Westview Press.

Van der Lugt, L. M. and P. W. De Langen (2007). Port authority stategy: Beyond

the landlord; a conceptual approach.

Virilio, P. (2006). Speed and politics. Los Angeles: Semiotext(e).

Walsh, A. H. (1978). The Public’s Business. Cambridge: The MIT Press.

Waterfront Worker (1933a, August 15). Extracts from the code for longshoremen

drawn up by the Marine Workers Industrial Union. Waterfront Worker 1 (9), 2.

Waterfront Worker (1933b, August 19). Out of the history of the ILA. Waterfront

Worker 1 (8), 4.

Waters, R. C. (1977). Port economic impact studies: Practice and assessment. Trans-

portation Journal 16 (3), 14–18.

Watson, R. (1999a, May 11). Baltimore port sees gain in loss of Maersk deal. Journal

of Commerce, 1A.

Watson, R. (1999b, March 15). Baltimore seeks direct rail service to lure ship hub

deal. Journal of Commerce, 1A.

Watson, R. (1999c, April 5). CSX chairman says contending ports have time for

better offers. Journal of Commerce, 12A.

Watson, R. (1999d, April 13). CSX puts price on rail revamp for Baltimore. Journal

of Commerce, 12A.

306

Watson, R. (1999e, April 15). Showing unified front, Baltimore enhances bid for

Atlantic megaport. Journal of Commerce, 1A.

Weber, A. and C. J. Friedrich (1965). Theory of the location of industries. Chicago,:

University of Chicago Press.

Weber, M. (1997). The theory of social and economic organization. New York: Free

Press.

Weldon, F. L. (1958). Cargo containerization in the West Coast-Hawaiian trade.

Operations Research 6 (5), 649–670.

White, H. C. (1992). Identity and Control: A Structural Theory of Social Action.

Princeton: Princeton University Press.

White, R. D. (2010, January 4). Local ports fight to stay dominant. Los Angeles

Times .

Whitford, J. and C. Potter (2007). The state of the art: Regional economies, open net-

works and the spatial fragmentation of production. Socio-Economic Review 5 (3),

497–526.

Williamson, O. E. (1975). Markets and Hierarchies, Analysis and Antitrust Implica-

tions: A Study in the Economics of Internal Organization. New York: Free Press.

Winter, J. M. (1991). Joseph Paul St. Sure: Management labor negotiator, 1902–1966.

Master’s thesis, California State University, Sacramento.

World Bank (2007). Port Reform Toolkit. The International Bank for Reconstruction

and Development / The World Bank.

World Port Source (2009). World Port Source. Technical report, World Port Source.

307

World Port Source (2010, April 7). World Port Source. http://www.

worldportsource.com.

Yochum, G. R. and V. B. Agarwal (1987). Economic impact of a port on a regional

economy. Growth and Change 18 (3), 74–87.

Zeitlin, J. (2005). Industrial districts and flexible production: Past, present, and

future. In M. Lescure (Ed.), La mobilization du territoire: Les districts industriels

en Europe occidentale, du XVIIe au XXe siecle [The Mobilization of Territory:

Industrial Districts in Western Europe from the 17th through the 20th Centuries].

Paris: Comite pour l’Histoire Economique et Social de la France.

http://www.worldportsource.com

http://www.worldportsource.com

308

List of Interviews

William Dudley, President ILA Local 1233. March 12, 2009.

Andrew Genn, New York City Economic Development Corporation. September 18,

2008.

Richard Larrabee, Director of Port Commerce for the Port Authority of New York

and New Jersey. March 25, 2008.

Lombardi, Dennis, Deputy Director of Port Commerce for the Port Authority of New

York and New Jersey. March 25, 2008.

Documents

COLUMBIA UNIVERSITY 2010